Immunocytochemistry. Run 106/35. Also In This Issue IN THIS ISSUE

Transcription

1 Run 16/35 IN THIS ISSUE Articles / Reports Scheme Updates July Review of Neuropathology EQA Year: The Management of Uncertainty: Increasing Stringency for Validation and Verification of Immunohistochemistry in Tissue Diagnostics Immunocytochemistry Improving Immunocytochemistry for Over 25 Years Results - Summary Graphs - Pass Rates Best Methods - Selected Images Assessment Dates: 23rd June - 8th July 214 Immunocytochemistry Modules General Pathology: CK7 & Calretinin 8-15 Breast Pathology: ER Breast Pathology: HER2 IHC Gastric: HER2 IHC 33-4 Lymphoid Pathology: CD5 & CD Neuropathology: MIB-1 & CK Cytology: CD45 & Ki Alimentary Tract: GIST: CD117 & DOG Alimentary Tract: Lynch syndrome: MLH1 & PMS In-situ Hybridisation Modules Breast: HER2 ISH Interpretive Breast HER2 ISH Technical Click on sponsor logos below to go straight to the sponsor webpage Cover Photo: Taken from the General Pathology Module: Top Left: In house section of breast carcinoma stained with CK7 Top Right: Optimal and weak CK7 staining of the NEQAS normal breast section Bottom Left: In house section of mesothelioma stained with calretinin Bottom Right: NEQAS appendix stained with calretinin Also In This Issue UK NEQAS ICC & ISH Updates and important news A short review of the Neuropathology EQA year: The Management of Uncertainty: Increasing Stringency for Validation and Verification of Immunohistochemistry in Tissue Diagnostic

2 General Information Data shown in this article is collated from UK NEQAS ICC & ISH assessments and is presented and described as collected, and does not ether endorse nor denounce any particular product or method and is provided as a guide to highlight optimal and sub-optimal staining methodologies. UK NEQAS ICC & ISH does not endorse any of the products featured by the commercial sponsors and are placed at the discretion of UK NEQAS ICC & ISH. Furthermore, commercial companies featured do not have any input or influence over the content, including results that are shown. For further information of the UK NEQAS ICC & ISH scheme, general EQA enquiries, slide returns and advertising opportunities please contact: Dr Merdol Ibrahim, Scheme Manager UK NEQAS ICC & ISH Suite 4/12 Hamilton House Mabledon Place London WC1H 9BB, UK Tel: +44 (2) merdol.ibrahim@ucl.ac.uk For enquiries concerning training issues, meetings, or courses, please contact: Mr Keith Miller, Scheme Director UK NEQAS ICC & ISH Suite 4/12 Hamilton House Mabledon Place London WC1H 9BB, UK Tel: +44 (2) k.miller@ucl.ac.uk Director Mr Keith Miller (k.miller@ucl.ac.uk) Manager Dr Merdol Ibrahim (merdol.ibrahim@ucl.ac.uk) Deputy Director Mr Andrew Dodson (Andrew.Dodson@icr.ac.uk) Assistant Manager Ms Suzanne Parry (s.parry@ucl.ac.uk) Scientists Mrs Dawn Steele/Mr Neil Bilbe (dawn.wilkinson@ucl.ac.uk/ n.bilbe@ucl.ac.uk) Office Manager Mrs Ailin Rhodes (a.rhodes@ucl.ac.uk) Clerical Assistant Mrs Clara Lynch (clara.lynch@ucl.ac.uk) Quality Manager Mrs Elizabeth Wilkinson- Smith (e.wilkinsonsmith@ucl.ac.uk) ASSESSORS UK NEQAS ICC has over 1 Assessors (scientists and pathologists) from the UK, Australia, Canada, Denmark, Germany, Hungary, Ireland, Portugal, Slovenia, South Africa, Sweden and Switzerland. The list below shows assessors who took part in the current assessment. United Kingdom Mr D Allen, London Dr N Atkey, Sheffield Mr N Bilbe, London Mr J Brown, Cambridge Mr D Blythe, Leeds Dr L Carson, Aberdeen Ms A Clayton, Preston Ms A Cramer, Manchester Mr A Dodson, London Mr R Fincham, Cambridge Mr D Fish, Reading Mrs S Forrest, Liverpool Mr S Forrest, Liverpool Dr I Frayling, Cardiff Ms J Freeman, London Mr M Gandy, London Ms C Gillette, London Mr J Gregory, Birmingham Dr R Hunt, Stockport Dr N Hand, Nottingham Ms L Happerfield, Cambridge Dr M Ibrahim, London Prof B Jasani, Cardiff Ms S Jordan, London Ms K Kennedy, Belfast Dr G King, Aberdeen Mr C Marsh, Newcastle Dr P Maxwell, Belfast Mr K McAllister, Cork Mrs H McBride, Belfast Dr B Mahler Araujo, Cambridge Dr J Moorhead, London Dr M Morgan, Cardiff Mr K Miller, London Dr G Orchard, London Ms A Patterson, Belfast Ms S Parry, London Dr M Pitt, Cambridge Mrs F Rae, Edinburgh Mr G Rock, Birmingham Mr J Ronan, Nottingham Mr A Shore, Hertfordshire Dr J Starczynski, Birmingham Ms D Wilkinson, London Mrs C Thomas, Preston Dr P Wencyk, Nottingham Denmark Dr B Rasmussen, Copenhagen Germany Dr I Nagelmeier, Kassel Ireland Dr T O Grady, Dublin Portugal Mr R Roque, Lisbon Mr J Matos, Lisbon Ms A Tavares, Lisbon Dr A Ferro, Sobreda Slovenia Ms S Gabric, Golnik South Africa Ms R Van Wijk, Fish Hoek Switzerland Dr L Tornillo, Basel Journal layout and design prepared by UK NEQAS ICC & ISH UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent

3 1

4 Updates July 214 Update to Breast Modules, Cytopathology and Forthcoming NSCLC EML4-ALK Module Breast HER2 IHC Module The HER2 IHC module has been updated as of Run 16 (July 214) to not only include membrane staining interpretation on each sample (NEQAS and in-house), but reports now also include an overall numeric score for the overall quality of the samples. Updated Graphical Reports New graphs are provided to show more clearly the overall distribution of scores Assessment Procedure (see also Participants Manual 214) Assessors provide their interpretation on the membrane staining of each of the samples (NEQAS distributed and inhouse). This is also a very good exercise to compare inhouse self-assessment, and interpretation scores, with that of the UK NEQAS assessment team. Once the team of assessors have assessed the membrane interpretation for each of the NEQAS samples, they will award a potential score out of 5 marks; based on the interpretability of the membrane staining and technical staining quality. An overall pass mark is then awarded by combining the four assessors scores to give scores out of 2. Score and Interpretation 16-2/2: Excellent 13-15/2: Acceptable 1-12/2: Borderline Acceptable 4-9/2: Unacceptable Interpretation Overall the staining is at the expected level for each of the samples. Some slight technical issues noted by some of the assessors, but overall the staining is suitable for interpretation Overall the samples are borderline interpretable (still clinically relevant) indicating that technical improvements need to be made. Marks may have been deducted due to: Weaker/stronger staining than expected, cytoplasmic staining, morphological damage etc Overall the samples are of unacceptable quality for clinical interpretation and technical improvements need to be made. Marks may have been deducted due to: False positive/ negative membrane staining, excessive cytoplasmic staining, excessive morphological damage, staining of normal glands In-house controls In-house control tissue must include 3+, 2+ and 1+/ invasive breast carcinomas. Laboratories who have problems in sourcing appropriate invasive control material are permitted to use DCIS tissue, showing differing levels of membrane staining as an acceptable alternative. This must be indicated on the online methodology comments section. Cell lines are an acceptable alternative BUT commercial kit controls are not acceptable. And are not assessed if submitted. Participants who do not submit a composite in-house control (3+, 2+ & 1+/), will be scored a maximum of 12/2 (borderline: acceptable) Breast HER2 ISH Module Updated Participant Reports Individual reports have been modified to allow assessors to enter more pre-selected comments, there is also a larger free -text area for assessors to provide more in-depth feedback. Benchmarking graphs included on all individual reports, allow participants to quickly compare how they are scoring against the overall mean score and to monitor change in performance over time. ISH Interpretive Assessment Procedure and Report The ISH interpretive module assessment procedure is being updated to allow for more informative feedback on interobserver variability in counting. As of Run 36 (July 214) Assessment scoring data will be generated by using all participant submitted data to calculate mean scores for HER2 copy & Ch17 counts as well as overall ratio calculations. We will no longer be using reference centre ranges. Marks will be awarded for each criteria (HER2 copy, Ch17 & ratio) by comparing participant scores to standard deviation (1, 2 and 3 stdev) from the mean score. This will allow participants to better gauge where counts may be too high or too low. New HER2 ISH interpretive scoring guide UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 2

5 Updates July 214 Update to Breast Modules, Cytopathology and Forthcoming NSCLC EML4-ALK Module New individual report format Breast Hormonal Receptor (ER/PR) Module We have been validating Same Slide NEQAS Material + Participant In-house control. UK NEQAS distributed Participant cuts & places in-house control here & stains using ISH Technical Assessments Please always return your NEQAS FISH/CISH slides for technical assessment, Once we have the new interpretive module established we plan to combine the interpretive and technical modules so that participants have much clearer feedback as to where and why they may be having performance issues. FISH slides are more stable than we think! UK NEQAS receives FISH slides from all over the world and if treated carefully, the fluorescence will not deteriorate too quickly. a. Sections should be mounted using a fluorescence antifade mounting media, which prevents loss/quenching of fluorescence e.g. Vectashield (Vector labs), Dako Fluorescence Mounting Medium (Dako), Fluoromount (SIGMA/Aldrich, VWR). Note: The Abbott Vysis kit already contains DAPI in phenylenediamine dihydrochloride, which is an anti-fading reagent. b. Seal the edges of the coverslip onto the slide using clear varnish or another sealing agent, which will stop the slide from drying out during transportation. ISH in-house control At present there is no strict rule as to the number or types of in-house control to submit. Samples can be amplified, equivocal or non amplified. ISH inherently has internal stromal cells which act as internal markers to assess whether your hybridisation is optimal. FISH technical assessment procedure Fluorescent ISH slides are now technically assessed by 4 assessors at the same time. Achieved by incorporating a live-feed video onto the fluorescence microscope with the image viewed on a large high definition monitor. This allows all 4 assessors to view and score the FISH slides at the same time. Samples will be unbaked allowing labs to cut their in-house control and carry out their normal baking procedure. An initial feasibility study is now under way for breast hormonal receptor (ER/PR) markers. Initial assessment using the above method is planned for September 214 Methodology will: a. Reduce laboratory EQA reagent costs b. Ensure same protocol is applied to all samples c. Decrease assessments and report TATs d. Allow easier comparison of NEQAS & in-house samples If successful, this new EQA procedure will be expanded to other modules over time. Cytology Module UK NEQAS ICC & ISH has been distributing cytospins for the past 15 years. A recent survey (May 214) of participants indicated that 66% routinely used cell blocks As of Run 17 (August 214) participating labs will receive their preferred material for the cytology assessment Same samples will be used for both preparations Cell block slides will have CB printed on them. If you wish to change your requirement for subsequent runs please contact the scheme: n.bilbe@ucl.ac.uk EML4-ALK Pre-pilot Module We have had a good response and received a high level of interest for this module; we would like to thank those labs that have registered to take part in the initial pre-pilot and/or the pilot module. At present the most likely launch date for this will be Autumn 214. Further news will be sent in due course. UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 3

6 Neuropathology Module : A Short Review. Neil Bilbe and Robert Fincham Introduction Average scores by run and antigen (UK labs): The EQA year covers the Runs 12 to 15. This short article will look at the registration numbers, antigens, overall scores and any significant trends seen during this period, concentrating on UK labs, and the NEQAS tissues. Registration numbers The maximum number of active laboratories for the year was 71. The number of registered labs varies, as participants withdraw or register for the module, particularly at the start of the EQA year. The average number of registrants was 69. This was an increase of four from 212/213. Three European labs withdrew from the module, but we received 7 new registrations: 4 European, 1 UK, 1 Asian and 1 USA. The distribution of all Neuropathology registered labs: Disappointingly, the Gold (G) NEQAS sections, had the lowest average score (15). Generally, labs performed better on the 2nd antigens (J), than with GFAP (G), with EMA being the only exception (run 15) with an average of 15. In-house slides were fairly consistent with averages of 16 (H) and 17 (K). The overall run average score of 16 (see above) indicates that the UK labs are performing to a high standard. Comparison of UK scores to the median (all scores): If the median scores are taken as the average for all labs submitting for the neuropathology module (UK and non-uk) then the averages for the non-uk labs are approximately one mark lower (15) than those for the UK labs (16). N.B. UK NEQAS ICC & ISH median scores are calculated to the nearest.5 of a mark. We currently have 68 active labs. Five new labs have registered, but 8 have withdrawn, including the loss of 3 UK labs due to the consolidation of services, a familiar trend. Antigens, overall submission rates, and tissue details: The Gold for the year was Glial Fibrillary Acidic Protein (GFAP); slides G (NEQAS) and H (In-house). Cases used were a low grade glioma, a glioblastoma, and a GMB/oligo. Submission rates, second antigens, and tissue were: Run 12 (99%) - CD34: Haemangiopericytoma Run 13 (96%) - CD68: Radio induced necrosis Run 14 (98%) - S1: Schwannoma Run 15 (97%) - EMA: Breast metastasis A submission is when a lab returns slides, even if they return only the Gold slides, and not the 2nd antigen, or in-house. All non-submissions were from non-uk or non-diagnostic laboratories; this constitutes a rate of between one and three labs per run. Results and run by run comparison (UK labs only) Following the first run of the year (12), one UK participant withdrew from the module, having not submitted for this run either, leaving a total of 3 UK participants; 29 were diagnostic labs, and there was one commercial laboratory. For the purposes of calculating statistics therefore, the laboratory that withdrew was ignored. N.B. The commercial sector (UK and Overseas) are welcome to join the scheme for all modules, but they are not subject to poor performance review (UK). They also provide valuable, additional validation slides for some of our modules. Poor performance (UK), failure and borderline rates Only UK diagnostic labs and the NEQAS slides, are included in the poor performance (PP) review, carried out after each run. During the last EQA year no UK laboratory achieved a status of Red or Amber for the neuropathology module. N.B. If reports were run for all labs, 5 non-uk labs would have been given a status of Red. The majority of these are for nonsubmissions, rather than failed slides due to the ICC staining. Failed and borderline scores (UK and All labs): The 2 UK labs that failed on runs 12 and 14 were both on NEQAS GFAP slides (G), because of weak or very weak demonstration of the tumour. Coincidentally, the 5 UK labs with borderline scores for run 12 were also on the NEQAS GFAP slides. Borderline scores for UK labs were more evenly distributed for the other runs, 13-15, but the GFAP samples proved the most difficult to stain to a consistently high standard. The NEQAS EMA sample (J), proved the most challenging of the 2nd antibodies, 6/17 of the borderline scores (UK) for run 15 were on the NEQAS EMA slides (J). Summary (see the e-journals for methodology details) Neuropathology participants are performing to a high standard, and there are no performance issues with UK labs. The staining of the GFAP slides was the most interesting, given how common a marker this is. CD34, CD68 and S1 had little or no problems, with only EMA of the 2nd antigens posing any real challenges to the participants. UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 4

7 The Management of Uncertainty: Increasing Stringency for Validation and Verification of Immunohistochemistry in Tissue Diagnostics Perry Maxwell, Stephen McQuaid and Manuel Salto Tellez Northern Ireland Molecular Pathology Laboratory. Belfast Health & Social Care Trust and Queen s University Belfast, UK Primary Antibody Uncertainty All laboratories performing immunohistochemistry (IHC) and undergoing accreditation under ISO regulations must validate their tests through evidence-based procedures and document analytical validation. Under ISO the degree of uncertainty needs to be determined and managed by establishing the accuracy, specificity and reproducibility of each test e.g. Clause For IHC, it should be determined if the antibody is used for diagnostic, molecular or therapeutic testing (McCourt et al 213). PCR-based molecular diagnostic validations are regarded as very rigorous, while IHC (a qualitative and subjective assessment) has been criticised for a lack of stringency. In this paper, our aim is to establish the basis on which IHC laboratories can bring the same level of completeness in the validation of antibodies for use in IHC tests. The use of standards and controls is essential for IHC. As outlined below and in Figure 1, the use of well characterised antibodies (Level 1 in our model) is a pre-requisite startingpoint for diagnostic purposes. Bordeaux et al (21) described criteria by which antibodies should be characterised at source (Levels 2 and 3 in our model), including Western blot data and knockdown or transfected cell lines models. Moreover, the purchase of suitable antibodies for diagnostic IHC from commercial sources should be guided by the stringency of validation which specific companies include in their datasheets. At optimal this would include testing by various immune methods including flow cytometry and ELISA. Further testing by the company in FFPE cell pellets from cell lines with knock-out or knock-in of the target would also be key factors to assess in the choice of commercial antibodies. Ideally well-documented staining methods for biopsy and/or surgical resection specimens with details for predicted staining patterns, pretreatment regimes and suggested concentration/dilution parameters should also be available from the company. Question 1: What is known about the antibody? The specific choice of commercially available antibodies for use in diagnostic IHC test procedures can be stratified to 3 levels based on knowledge of the antibody under test: Level 1 - antibody specificity is fully understood and publications and external quality assurance data exist to support usage for general diagnostic evaluations. Single clone of antibody only may be evaluated; an evaluation at the next 2 levels may require a higher knowledge of the biology of the biomarker, to accurately assess important aspects of antibody expression such as sub-cellular localisation, expression in background cells, etc. Level 2 - antibody specificity is known and research publications exist (Mohd Omar et al 21). At this level there is a need to examine target expression in different species (spontaneous tumour identification), xenografts or cell lines. Determination of fit for purpose characteristics may require up to 4 different clones to be tested in first approach; Level 3 - antibody specificity not tested or identified, as for novel biomarker development or pharmacodynamic assays. The evaluation of unidentified targets requires a more stringent approach with a rigorous assessment of antibody utilisation. Validations at this level must also include full specificity and sensitivity evaluations. Antibodies which fall within levels 2 and 3 should be considered to have a high degree of uncertainty and therefore should not be used in diagnostic tests. Question 2: Is there a good protocol? This forms the technical workup, with the determination of antigen retrieval and working dilutions. The correct choice and use of controls both for this and for ongoing IHC tests remains central to continuous proof of specificity and sensitivity. Use of well-designed Tissue Microarrays (TMAs) with cores from surgical specimens representing a range of expression value from no expression through high expression for a variety of antibodies and to include relevant FFPE cell pellets of cell lines represents an ideal quality control for the accuracy, specificity and reproducibility of antibodies in test procedures. The inclusion of cell lines has the advantage of homogeneity and therefore removal of variability factors such as ischaemic and fixation times which are present with surgical tissue samples. Question 3a: Does it have the expected use accuracy & specificity? Accuracy and specificity are determined through the use of reference standards, where the gold standard has been identified for a defined purpose i.e. the level of concordance. The number of cases which need to be included as reference standards is dependent upon the degree of concordance acceptable for predictive or diagnostic purposes. The reference standard is dependent on the diagnostic value of the IHC test. For instance, a novel antibody which may be included in a panel to support a diagnosis of carcinoma versus other malignancies would need to be compared against a collection of various cancers; a series of antibodies aiming to help in the diagnosis of HNPCC would require accurate concordance within a retrospective case series with cases in which MMR mutation status or at least MSI status has been established. The gold standard for use of antibodies in therapeutic IHC is more complex. The ideal validation set would include cases with known response to therapy. However, this is not feasible within many centres. The alternative is to validate these antibodies according to national and international standards and seek concordance with other laboratories where the test is already validated. The College of American Pathologists (CAP) has issued guidelines on IHC validation (Fitzgibbons et al 213). Where there are clear guidelines published for predictive testing, as for ER, PR and Her2, the number of cases to be used in order to achieve expected 95% concordance is usually over 1. Where the antibody is used for diagnostic purposes and where no clear guidelines exist, a lower level of concordance of 9% is acceptable and below this it is up to the Laboratory Director to justify the acceptance of the antibody for diagnostic use. Under ISO 15189, the Laboratory Director must be clearly identified and this would be within his/her remit. In order to achieve the lower level of concordance where there is a clear positive/negative result, then using 1 expected positive and 1 expected negative cases should be sufficient provided a simple positive or negative result is to be recorded. Where more complex scoring exists, then the number of reference standards needs to be increased accordingly (Mattocks et al 213). Question 3b: What is the reproducibility? Reproducibility is important. This can be achieved showing that the same result occurs 3 times in a single staining batch together with 3 individual batches over subsequent runs. Where more than one automated platform is used, even if these are from the same manufacturer it is necessary to show the level of inter- and intra-machine/platform reproducibility. Verification of reproducibility between antibody lots should also be performed and documented within the laboratory. Use of serial sections of control Tissue Microarrays (TMAs - see above) are ideal for this purpose. Staff and Equipment Uncertainty The role of the Laboratory Director is central to testing (Clause ). This person must demonstrate that he/she is qualified UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 5

8 Perry Maxwell, Stephen McQuaid and Manuel Salto Tellez to determine all subsequent test parameters and is in a position of authority to implement change necessary to establish and maintain quality. It is also necessary to demonstrate personnel carrying out the tests are competent (Clause ), with relevant competency documents signed ultimately by the Laboratory Director. Moreover, all equipment must be shown to be properly maintained by suitably qualified staff through a documented maintenance schedule and/or by the manufacturer or supplier under a preventative maintenance contract (Clause 5.1). In the UK, the company performing maintenance shall show that it has achieved relevant ISO standards. Conclusion In conclusion, for the validation of IHC tests for which guidelines do not exist, pre-analytical fixation and processing need to be documented, together with the number of cases used and the expected versus resultant immunoreactivities. It should be remembered that all IHC tests must produce a level of certainty regarding the accuracy and reproducibility of antibodies in use for clinical purposes. Any changes to protocols must be documented along with the evaluation of their effect on tests and interpretation. The continued monitoring of performance, regular staff competence review and authorisation by the Laboratory Director forms the framework which accompanies all IHC tests. References Bordeaux J, Welsh AW, Agarwal S et al. Antibody validation. Biotechniques 21;48: Courtney B. Introduction to ISO 15189: Overview and mapping to current CPA Standards. Fitzgibbons PL, Bradley LA, Fatheree BS et al. Principles of analytic validation of immunohistochemical assays. Arch Pathol Lab Med 214; March 19, EPub ahead of print. Mattocks CJ, Morris MA, Matthijs G et al. A standardized framework for the validation andverification of clinical molecular genetic tests. Eur J Hum Gen 21;18: McCourt CM, Boyle D, James J, Salto- Tellez M. Immunohistochemistry in the era of personalised medicine. J Clin Pathol. 213;66: Mohd Omar MF, Huang N, Ou K, Yu K, Putti TC, Jikuya H, Ichikawa T, Nishimura O, Tan P, Salto-Tellez M. Molecular-assisted immunohistochemical optimization. Acta Histochem. 21;112: Note: This article and its content is independently submitted by the authors, and views expressed are of the authors alone and not of UK NEQAS ICC & ISH. Figure 1 UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 6

9 7

10 The General Pathology Module Run 16 Perry Maxwell and Merdol Ibrahim Introduction Gold Standard Second Antibody Antigens Assessed: CK7 Calretinin Tissue Sections circulated: Normal Breast and Colon Appendix and Mesothelioma Number of Registered Participants: 338 Number of Participants This Run 325 (96%) Gold Standard: CK7 Cytokeratin 7 (CK7) is a 54 kda type II simple cytokeratin which is found in most glandular and transitional epithelia, but not stratified squamous epithelium. Antibodies against CK7 label a variety of cells in normal tissue including subsets of endothelial cells (Chu & Wiess). CK7 can also be demonstrated in a number of different tumours, including adenocarcinoma of the lung, breast, ovary (serous and endometroid), endometrium and in transitional cell carcinoma of bladder, although the level of staining very much depends on the degree of differentiation of the tumour. It is generally negative in colonic adenocarcinomas, squamous cell carcinomas, hepatomas and renal epithelial tumours. Staining with CK7 should be cytoplasmic, and it is usually used in conjunction with cytokeratin 2. (Dabbs). Features of Optimal Immunostaining: (Figs 1A, 2A, 3 &5) Normal Breast Strong and even cytoplasmic/membranous luminal staining of the glandular epithelial cells Weaker staining of the myoepithelial cells Clean background with no staining of the stromal cells Normal Colon Negative reaction in the colon Clean background with no staining of the stromal cells Features of Suboptimal Immunostaining: (Figs 1B, 2B, 4 & 6) Weak and/or uneven staining of the glandular epithelial cells of the breast Non-specific staining of stromal cells or connective tissue of the breast or colon Non-specific staining of the colon epithelial cells Excessive background staining References 1. Chu PG, Wiess LM. Keratin expression in human tissues and neoplasms. Histopathology. 22 May;4(5): Dabbs DJ. Carcinomatous differentiation and metastatic carcinoma of unknown primary. In: Diagnostic Immunohistochemistry. Dabbs D (Editor). Church Livingstone, Philadelphia 22; pg Second Antigen: Calretinin Calretinin is a 29 kda calcium-binding protein involved in calcium signalling. It is found abundantly in neurons and the thymus, and is associated with the kinetocore during the cell cycle. Outside the nervous system, calretinin is found in a number of cells with varying expression, including mesothelial cells, steroid producing cells, testicular cells, ovarian surface epithelium, some neuroendocrine cells, breast glands, and hair follicle cells. Diagnostically, calretinin is used as a positive marker for both benign mesothelium and in malignant mesothelioma (Saydan et al). Its use in the identification of mesothelioma in cytological preparations (Doglioni et al), and in the differential diagnosis between malignant mesothelioma and adenocarcinoma in FFPE cell blocks of cytological fluids, washings and aspirates (Wiezorek and Krane) have been described. Calretinin can also be used to help differentiate lung tumours (Marchevsky), and also to distinguish between different types of brain tumour, i.e. neuronal rather than glial differentiation (Leong et al). Features of Optimal Immunostaining Appendix Strong cytoplasmic and nuclear staining of the peripheral nerves and macrophages Clean background with no non-specific staining of the epithelial cells Mesothelioma Strong cytoplasmic and nuclear staining of the majority of tumour cells Clean background with no staining of the stromal cells Features of Suboptimal Immunostaining Weak or absent staining in the peripheral nerves of the appendix. Weak or absent staining in the tumour cells of the mesothelioma Excessive background staining. Excessive non-specific staining of cell types or components not expected to stain References Saydan N, et al. Expression of calretenin in human mesothlioma cell lines and cell cycle analysis by flow cytometry. Anticancer Res 21 (1A): Doglioni C, et al. Calretenin: a novel immunocytochemical marker for mesothelioma. Am J Surg pathol 1996; 2: Wieczorek TJ, et al. Diagnostic utility of calretenin immunohistochemical marker for mesothelioma. Am J Surg pathol 1996; 2: Marchevsky AM. Application of immunohistochemistry to the diagnosis of malignant mesothelioma. Arch Pathol Lab Med. 132 (3): Leong A, et al. Manual of Diagnostic Cytology (2nd ed.) Greenwich Medical Media Ltd. pp Assessment Summary: Overall, most participants achieved acceptable results for both antibodies assessed in this run (pass rates are shown in the subsequent summary graphs). For the CK7 assessment the Dako OV-TL12/3 clone was by far the most favoured antibody. This showed a pass rate of 98%, and was noted to work well on all the commercial staining platforms, including the Dako Autostainer, Ventana Benchmark and Leica Bond machines. No labs failed the CK7 assessment, and only 5% of labs received a borderline pass (scores of 1-12/2). The main reason being due to weak staining, most often caused by inappropriate/insufficient pretreatment or incorrect antibody dilution. For the Calretinin assessment, the Dako CALRET antibody clone was used by most labs. This showed a pass rate of 76%. Both the Ventana DC8 and Leica CAL6 clones were also popular, and showed pass rates of 92% and 82% respectfully. Overall, scores for the Calretinin assessment were slightly more varied, with 7% of labs actually failing the run. Again this was mainly due to very weak or little staining of the NEQAS distributed tissues, and this was often caused by using a citrate antigen retrieval buffer rather than a high ph EDTA buffer, which is the recommended pre-treatment for the Calretinin antibody. UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 8

11 RUN 16 GENERAL PATHOLOGY Module Selected Images showing Optimal and Sub-optimal Immunostaining Fig 1. Two examples of CK7 demonstration in the UK NEQAS normal breast sample. Both stained using the Leica OV-TL 12/3 antibody. (A) Optimal strong cytoplasmic/membranous staining in the glandular epithelial cells, while (B) is very weak, due to no antigen retrieval being carried out. Section (A) was pre-treated. Fig 2. 2 examples of CK7 staining in the UK NEQAS normal breast sample carried out using the Dako OV-TL 12/3 antibody. Again, (A) was given antigen retrieval and shows optimal results, while (B) is weak, due to no/insufficient antigen retrieval. Fig 3. Another example of optimal CK7 staining in the UK NEQAS normal breast. The section shows the strong cytoplasmic/membranous staining in the luminal cells, and also the weaker staining of the myoepithelial cells. Section stained with the Dako OV-TL 12/3 antibody, 1:4, on the Ventana Benchmark XT, with Protease I enzyme digestion for 12 minutes. Fig 4. Borderline pass demonstration of CK7 in the UK NEQAS normal colon. The tissue is CK negative CK7, but this example shows some inappropriate, non-specific and background staining. Section stained with the Dako OV-TL 12/3 antibody, 1:2, on the Leica Bond III, with VBS Bond enzyme digestion for 1 minutes. Fig 5. Good example of an in house breast control stained with CK7. The cytoplasmic/membranous staining is intense, while the background stromal tissue remains clean. Section stained with the Leica OV-TL 12/3 antibody, 1:25, on the Ventana XT with Protease I enzyme, digestion for 8 minutes. Fig 6. Unacceptable demonstration of CK7 on a participants in-house control. Although the ovarian tumour is strongly stained, there is also excessive inappropriate background staining, most likely due to inappropriate dilution of the antibody. 9

12 RUN 16 GENERAL PATHOLOGY Module Selected Images showing Optimal and Sub-optimal Immunostaining Fig 7. Optimal demonstration of Calretenin in the UK NEQAS distributed appendix. The section shows strong distinct cytoplasmic and nuclear staining in the peripheral nerves, while the background remains clean. Section stained with the Leica RTU CAL6 antibody, on the Leica Bond III and ER1 antigen retrieval for 2 minutes. Fig 8. Borderline pass staining of Calretenin in the UK NEQAS distributed appendix section. Although some of the peripheral nerves are staining, there are fewer elements being demonstrated than expected (compare to Fig 7). Section stained with the Leica 5A5 antibody, 1:5, on the Dako Autostainer, and pre-treatment in the PT link with low ph buffer for 2 Fig 9. Unacceptable demonstration of Calretenin in the UK NEQAS distributed appendix. Although the cells expected to stain appear to be staining, there is also non-specific and background staining, making it difficult to determine the true specific staining. Section stained with the Cell Marque pre-diluted polyclonal antibody, on the Ventana Benchmark with CC1 Fig 1. Optimal demonstration of Calretenin on the UK NEQAS mesothelioma. The section shows moderate to strong nuclear and cytoplasmic staining of virtually all the tumour cells. Section stained with the Ventana SP65 pre-diluted antibody, on the Ventana Benchmark XT with CC1 mild antigen retrieval. Fig 11. Suboptimal demonstration of Calretenin on the UK NEQAS mesothelioma: The staining intensity is too weak and the percentage of tumour cells staining is far lower than expected (compare to Fig 1). (Same protocol as Fig 8). Fig 12. Good example of an in house mesothelioma control stained with Calretenin. The image shows strong, distinct staining of the tumour cells. Section stained with the Vector 5A5 antibody, 1:5, pre-treatment in the pressure cooker for 2 minutes with citrate buffer, and detection with the Dako Envision kit. 1

13 Run 16 GENERAL PATHOLOGY Module GRAPHICAL REPRESENTATION OF PASS RATES 12 RUN 16A CK 7 on NEQAS Sections Individual 4 = (%) 12 RUN 16B CK 7 on in-house Sections Individual 4 = (%) 1 5 = (%) 6 = (%) 1 5 = 1 (%) 6 = (%) 7 = (%) 7 = (%) no. of returns = (%) 9 = (%) 1 = (%) 11 = 3 (1%) 12 = 12 (4%) 13 = 11 (3%) 14 = 24 (7%) no. of returns = 2 (1%) 9 = (%) 1 = 3 (1%) 11 = (%) 12 = 22 (7%) 13 = 14 (4%) 14 = 14 (4%) 15 = 28 (9%) 15 = 26 (8%) 2 16 = 12 (31%) 17 = 13 (32%) 2 16 = 12 (32%) 17 = 13 (32%) = 27 (8%) 19 = 9 (3%) 2 = 6 (2%) = 23 (7%) 19 = 7 (2%) 2 = 4 (1%) Summary Summary 16-2 = 247 (76%) 16-2 = 239 (74%) = 63 (19%) = 54 (17%) 1-12 = 15 (5%) 1-12 = 25 (8%) - 9 = (%) - 9 = 3 (1%) Median = 15.5 Median = 14.5 no. of returns RUN 16C Calretinin on NEQAS Sections Individual 4 = 2 (1%) 5 = (%) 6 = 1 (%) 7 = 1 (%) 8 = 11 (4%) 9 = 6 (2%) 1 = 12 (4%) 11 = 8 (3%) 12 = 3 (1%) 13 = 19 (6%) 14 = 29 (9%) 15 = 26 (8%) 16 = 88 (28%) 17 = 31 (1%) 18 = 28 (9%) 19 = 17 (5%) 2 = 3 (1%) no. of returns RUN 16D Calretinin on in-house Sections Individual 4 = 1 (%) 5 = (%) 6 = (%) 7 = 2 (1%) 8 = 4 (1%) 9 = 3 (1%) 1 = 3 (1%) 11 = 1 (3%) 12 = 37 (12%) 13 = 2 (6%) 14 = 23 (7%) 15 = 28 (9%) 16 = 112 (36%) 17 = 43 (14%) 18 = 12 (4%) 19 = 6 (2%) 2 = 4 (1%) Summary Summary 16-2 = 167 (54%) 16-2 = 177 (57%) = 74 (24%) = 71 (23%) 1-12 = 5 (16%) 1-12 = 5 (16%) - 9 = 21 (7%) - 9 = 1 (3%) Median = 12.5 Median =

14 Run 16 GENERAL PATHOLOGY Module ANTIBODIES AND OTHER TECHNICAL PARAMETERS EMPLOYED BY PARTICIPANTS IN THE GENERAL PATHOLOGY MODULE The following tables record the number of participants (N) using each primary antibody and the percentage (%) of these participants achieving acceptable staining (a score >12/2) on UK NEQAS sections. General Pathology Run: 16 General Pathology Run: 16 Primary Antibody : CK 7 Antibody Details N % Biocare CM 61 A,B, C (OV-TL 12/3) 3 67 Biocare PM 61 AA,H (OV-TL 12/3) 1 1 BioGenex MU255-UC (OV-TL 12/3) 2 1 Cell Marque 37M-96 (OV-TL 12/3) 6 83 Dako M718 (OV-TL 12/3) Dako N1626 (OV-TL 12/3) 4 1 Dako RTU Link IR619 (OV-TL 12/3) 13 1 DAKO RTU Omnis GA619 (OV-TL12/3) 1 1 Gennova AP1174C (OV-TL12/3) 1 1 Leica RTU PA942(RN7) 1 1 Linaris ( OV-TL 12/3) 1 1 NeoMarkers MS-1352-P (OV-TL 12/3) 4 1 Novcastra NCL-L-CK-56 (RN7) 19 1 Novocastra Bond RTU PA942 (RN7) 16 1 Novocastra NCL-CK7-OVTL (OV-TL 12/3) 5 8 Novocastra NCL-L-CK7-OVTL (OV-TL 12/3) Novocastra RTU-CK7-OVTL (OV-TL 12/3) 1 1 Other 1 9 Vector Labs VPC43 (OV-TL 12/3) 5 1 Ventana (SP52) Ventana, (OV-TL 12/3) 2 1 Zymed (OV-TL 12/3) 2 1 Primary Antibody : Calretinin Antibody Details N % BioGenex AR413 (R Poly) 1 Cell Marque - CMC757/758/759 3 Cell Marque 232A-74/75/76/77/78 (R Poly) 9 Dako IS627 RTU FLEX DAK-Calret Dako M7245 (DAK-Calret 1) Invitrogen (DC8) 2 95 Novocastra NCL-CALRET (5A5) Novocastra/Leica NCL-L-CALRET-566 (CAL6) Novocastra/Leica PA346 RTU (CAL6) Novocastra/Leica RTU-CALRET (5A5) 2 5 Other Thermo Fisher/Labvision RM-9113 (SP13) 6 33 Vector VP-C36 (5A5) 11 1 Ventana (SP65) Zymed/Invitrogen (DC8) General Pathology Run: 16 General Pathology Run: 16 Calretinin CK 7 Calretinin CK 7 Heat Mediated Retrieval N % N % _Leica BondMax ER _Microwave Oven 1 1 _Ventana Benk CC1 (Extended) 1 1 _Ventana Benk CC1 (Mild) 1 1 _Ventana Benk XT CC1 (Mild) 1 1 Biocare Decloaking Chamber Dako Omnis Dako Pascal Dako PTLink Lab vision PT Module Leica ER1 1 mins Leica ER1 2 mins Leica ER1 3 mins Leica ER2 1 mins Leica ER2 2 mins Leica ER2 3 mins Microwave None Other Pressure Cooker Pressure Cooker in Microwave Oven Steamer Ventana CC1 16mins Ventana CC1 2mins Ventana CC1 24mins Ventana CC1 32mins Ventana CC1 36mins Ventana CC1 4mins 2 1 Ventana CC1 48mins 1 1 Ventana CC1 52mins Ventana CC1 56mins Ventana CC1 64mins Ventana CC1 8mins Ventana CC1 extended 1 Ventana CC1 mild Ventana CC1 standard Ventana CC2 16mins 1 1 Ventana CC2 32mins Ventana CC2 64mins 1 1 Ventana CC2 standard 1 1 Water bath 68 OC Water bath OC Enzyme Mediated Retrieval N % N % AS PER KIT BioGenex Protease 1 1 Dako Pronase (S213) 1 1 Dako Proteinase K (S32) 1 1 NOT APPLICABLE Other 2 1 Trypsin-Difco 1 1 VBS Bond Enzyme VBS Bond Enzyme Ventana Protease 2 1 Ventana Protease 1 (76-218)

15 Run 16 GENERAL PATHOLOGY Module General Pathology Run: 16 General Pathology Run: 16 Calretinin CK 7 Calretinin CK 7 Detection N % N % A Menerini Polymer (MP-XCP) AS PER KIT BioGenex SS Polymer (QD 43-XAKE) Dako EnVision FLEX ( K8/1) Dako EnVision FLEX+ ( K82/12) Dako Envision HRP/DAB ( K57) Dako Envision+ HRP mouse K44/5/6/ LabVision UltraVision LP HRP (TS 125 HD) LabVision UltraVision ONE Polymer ( TL-12/5-HDJ/T Leica Bond Polymer Define (DS9713) 2 5 Leica Bond Polymer Refine (DS98) MenaPath X-Cell Plus (MP-XCP) None NOT APPLICABLE Novocastra Novolink PDS (RE7-14/15/28/29-K) 2 1 Other Power Vision DPVB999 HRP Vector Elite Universal ABC (PK-62) 1 1 Vector ImmPRESS Universal (MP-75) Ventana iview system (76-91) Ventana OptiView Kit (76-7) Ventana UltraView Kit (76-5) Chromogen N % N % A. Menarini Liquid Stable DAB kit 1 1 AS PER KIT BioGenex DAB (QD43) BioGenex Liquid DAB (HK153-5K) BioGenex liquid DBA (HK-124-7K) Dako DAB K DAKO DAB+ 1 1 Dako DAB+ Liquid (K3468) Dako DAB+ REAL Detection (K51) 3 1 Dako EnVision Plus kits Dako FLEX DAB Dako REAL EnVision K57 DAB LabVision (TA-125-HD) 1 1 Leica Bond Polymer Refine kit (DS98) menapath xcell kit DAB (MP-86) Other Sigma DAB (D5637) Sigma DAB (D595) Ventana DAB Ventana iview Ventana Ultraview DAB Vision BioSystems Bond X DAB General Pathology Run: 16 Calretinin CK 7 Automation N % N % BioGenex GenoMX 6i BioGenex Optimax Dako Autostainer Dako Autostainer Link Dako Autostainer plus Dako Autostainer Plus Link Dako Omnis LabVision Autostainer Leica Bond Max Leica Bond-III Menarini - Intellipath FLX None (Manual) Other Shandon Sequenza TechMate 5Plus 1 1 Ventana Benchmark Ventana Benchmark ULTRA Ventana Benchmark XT BEST METHODS - Gold Standard Antibody A selection from just a few of the best methods employed by participants CK 7 - Method 1 Participant scored 2/2 (UK NEQAS Slide) and 18/2 (In House slide) using this method. Primary Antibody: Ventana (SP52), 1 Mins, 37 ºC Automation: Ventana Benchmark XT Ventana UltraView DAB Main Buffer: Ventana reaction buffer (95-3) Ventana CC1 32mins Chromogen: Ventana Ultraview DAB Detection: Ventana UltraView Kit (76-5) 13

16 Run 16 GENERAL PATHOLOGY Module CK 7 - Method 2 Participant scored 19/2 (UK NEQAS Slide) and 17/2 (In House slide) using this method. Primary Antibody: Dako RTU Link IR619 (OV-TL 12/3), 2 Mins, 25 ºC Prediluted Automation: Dako Autostainer Link 48 Dako FLEX kit Main Buffer: Dako FLEX wash buffer, PH: 7.6 Dako PTLink, Buffer: EnVision FLEX TRS, PH: 9 NOT APPLICABLE Chromogen: Dako FLEX DAB, 25 ºC., Time 1: 1 Mins Detection: Dako EnVision FLEX ( K8/1), 2 Mins, 25 ºC Prediluted CK 7 - Method 3 Participant scored 2/2 (UK NEQAS Slide) and 18/2 (In House slide) using this method. Primary Antibody: Dako M718 (OV-TL 12/3), 16 Mins, 42 ºC Dilution 1: 1/2 Automation: Ventana Benchmark Ventana iview Kit Main Buffer: Ventana reaction buffer (95-3) Ventana Protease 1 (76-218), 42 ºC. Digestion Time NEQAS: 4 Mins. In-House: 4 Mins Chromogen: Ventana DAB, 42 ºC., Time 1: 8 Mins Detection: Ventana iview system (76-91), 8 Mins, 42 ºC Prediluted CK 7 - Method 4 Participant scored 2/2 (UK NEQAS Slide) and 18/2 (In House slide) using this method. Primary Antibody: Dako M718 (OV-TL 12/3), 3 Mins, 19 ºC Dilution 1: 2 Automation: Dako Autostainer Link 48 Dako FLEX+ kit Main Buffer: Dako FLEX wash buffer, PH: 7.6 Dako PTLink, Buffer: Dako FLEX TR Solution High ph, PH: 9 NOT APPLICABLE Chromogen: Dako FLEX DAB, 19 ºC., Time 1: 5 Mins, Time 2: 5 Mins Detection: Other, 3 Mins, 19 ºC Prediluted BEST METHODS - Secondary Antibody A selection from just a few of the best methods employed by participants Calretinin - Method 1 Participant scored 19/2 (UK NEQAS Slide) and 17/2 (In House slide) using this method. Primary Antibody: Novocastra/Leica PA346 RTU (CAL6), 15 Mins, 2 ºC Prediluted Automation: Leica Bond-III Leica BondMAx Refine KIT Main Buffer: Bond Wash Buffer (AR959), PH: 8 Leica ER2 2 mins, Buffer: EDTA, PH: 8 NOT APPLICABLE Chromogen: Leica Bond Polymer Refine kit (DS98), 2 ºC., Time 1: 5 Mins Detection: Leica Bond Polymer Refine (DS98), 2 Mins, 2 ºC Prediluted 14

17 Run 16 GENERAL PATHOLOGY Module Calretinin - Method 2 Participant scored 18/2 (UK NEQAS Slide) and 2/2 (In House slide) using this method. Primary Antibody: Dako M7245 (DAK-Calret 1), 2 Mins, 2 ºC Dilution 1: 1/1 Automation: Leica Bond-III Leica BondMAx Refine KIT Main Buffer: AS PER KIT Leica ER2 2 mins Chromogen: AS PER KIT, 2 ºC., Time 1: 8 Mins Detection: AS PER KIT, 1 Mins, 2 ºC Calretinin - Method 3 Participant scored 18/2 (UK NEQAS Slide) and 18/2 (In House slide) using this method. Primary Antibody: Zymed/Invitrogen (DC8), 32 Mins, 37 ºC Dilution 1: 4 Automation: Ventana Benchmark XT Ventana UltraView DAB Main Buffer: Ventana reaction buffer (95-3) Ventana CC1 standard Chromogen: Ventana Ultraview DAB Detection: Ventana UltraView Kit (76-5) Calretinin - Method 4 Participant scored 2/2 (UK NEQAS Slide) and 17/2 (In House slide) using this method. Primary Antibody: Ventana (SP65), 2 Mins Prediluted Automation: Ventana Benchmark ULTRA Ventana UltraView DAB Main Buffer: Ventana reaction buffer (95-3) Ventana CC1 8mins Chromogen: Ventana Ultraview DAB Detection: Ventana UltraView Kit (76-5) Prediluted 15

18 The Breast Hormonal Receptor Module Run 16 Merdol Ibrahim, Dawn Steele, Suzanne Parry and Keith Miller Antigen Assessed: Oestrogen Receptor (ER) Tissue Sections circulated: Number of Registered Participants: 298 Number of Participants This Run 277 (93%) Composite slide consisting of 3 breast carcinomas with different levels of receptor expression. Most slides also included normal tonsil and stomach (see table below*) Circulated Tissue The table below shows the staining characteristics of the tissue sections circulated during Run 16. This composed of three infiltrating ductal carcinomas (IDCs) with differing levels of receptor expression, along with sections of tonsil and stomach. The staining of the breast tumours were characterised using the Leica 6F11, Ventana SP1 and Dako EP1. Sections % Positivity Staining Intensity Allred / Quick Score A. IDC >95% High 8 B. IDC 4-6% Medium C. IDC % Negative D. Tonsil* 2-5% Medium See write-up for further comments E. Stomach* % Negative *Tonsil and stomach samples were also included on most slides but these were not scored in the actual assessment procedure. Please Note: Multiple tissue blocks are compiled from the same cases, however, there may on occasion be variability in staining from the Allred / Quick score shown above. This is taken into consideration when scoring participants slides and also noted prior to assessments by staining every 5th sample for quality control purposes. 6 General Guideline Used in The Assessment of Slides SCORE STAINING PATTERN Slide not returned by participant. 1 or 2 Unacceptable: E.g. False positive / false negative / non-specific or inappropriate staining / lack of staining Clinically Unacceptable. 3 Borderline Acceptable: Staining weaker than expected / background staining / weak/strong counterstain, Clinically still readable but technical improvements can be made 4 or 5 Acceptable: Demonstration of the expected proportion of nuclei stained in the invasive tumours, with roughly the expected staining intensity. Marks are also deducted when correct clinical interpretation of staining may be hindered due to factors such as: - Excessive cytoplasmic or diffuse nuclear staining - Excessively strong or weak haematoxylin counterstain - Excessive antigen retrieval resulting in morphological damage - Poor quality/inadequate choice of in-house control tissue ( poor/inadequate fixation, damaged cell morphology, over retrieval etc.) In-House Tissue Recommendations & Assessment Participants in-house control tissue MUST consist of composite breast tissue placed onto a single slide (cell line controls are an acceptable alternative). Commercial kit/assay controls are not accepted as the participants in-house control, and therefore will not be assessed. The required composite control should consist of the following samples: i. >8% tumour positivity with high intensity (Allred/Quick score 7-8) ii. 3-7% tumour positivity with low-moderate intensity (Allred/Quick score 3-6) iii. Negative tumour, ideally with normal positively stained glands (Allred/Quick score ) Participants NOT using a composite control are scored a 'borderline' pass (1-12/2). UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 16

19 The Breast Hormonal Receptor Module Run 16 Introduction Oestrogen receptor alpha (ER- ) plays a vital role in both the prognosis and predictive response of patients who may be considered for hormone therapy. Following the work of Harvey and colleagues 1, immunohistochemistry has now become the recognised gold standard for determining patient ER status. It is therefore crucial that not only the antibodies are correctly validated prior to patient-tissue use, but also proper control tissues are used to gauge the sensitivity of the test. An incorrect assay can lead to false ER staining 6,7, which can have a direct impact on patient treatment regime. Furthermore, the UK NHS Breast Screening Programme ( breastscreen/index.html) recommends using the Quick score (Allred) 1,2 to semi quantify the proportion and intensity of nuclear staining, thus further standardising the scoring criteria. Choice of Tissue for Assessments This assessment consisted not only of invasive breast tumours (samples A-C), but most slides also included tonsil and stomach tissue. The reason for using the tonsil and stomach samples was to help gauge the sensitivity of the assay. ER Staining Within the Tonsil The distributed tonsil sections showed ER expression in 2-5% of cells. This has also been indicated in the datasheets from the commercial suppliers, such as the Dako (EP1) and Ventana SP1 clones. Although not noted on the datasheet of the Leica 6F11 clone, staining was also seen in the tonsil of slides stained with this antibody. The staining was observed in some of the epithelial cells, but was mainly seen within the germinal centre lymphocytes. Quality Control of NEQAS Samples NEQAS tissue samples were tested by staining every 5th section on each of he relevant tissue blocks using the Leica 6F11 clone. This made sure that any heterogeneity in the tissue samples did not result in a participant being penalised. Furthermore, samples were also tested by the relevant commercial companies to further verify the expected level of staining. This included the Leica (clone 6F11), Dako (clone EP1) and Roche (clone SP1). We have recently also started to investigate the NEQAS distributed breast tumour samples for mrna expression using ACDbio RNAscope technology. The initial results (not shown) indicate a good correlation between protein (immunohistochemistry) and mrna expression. Assessment Results Features of Optimal Immunostaining: (Figs 1-7) Staining of the expected proportion of invasive tumour nuclei with the anticipated staining intensity Intense nuclear staining of the appropriate distribution in normal glands Cytoplasmic staining is not excessive No background staining of connective tissues or inappropriately localised staining Features of Suboptimal Immunostaining: (Figs 8-11) Inappropriate non-specific nuclear staining in the negative tumour (Fig 9B) Weak or lower expression of nuclear staining of the receptor positive tumours (Fig 8A) Nuclear staining is showing a higher expression level than expected Excessive cytoplasmic staining (fig11) Excessive background staining (Fig 8B) Excessive antigen retrieval Inappropriate staining of some cells in the tumour sections e.g. lymphocytes, fibroblasts (Fig 9A) Inappropriate non-specific staining in the stomach tissue Inappropriate non-specific or higher level of staining than expected in the tonsil tissue (Fig 1) NEQAS Slide Results Similarly to the previous ER assessment (Run 15), the results for this Run are very encouraging, with an overall pass rate of 91% (scores of >13-2/2) on the NEQAS sections. A further 8% received a borderline score (1-12/2), and only 4 labs (1%) failed the assessment. The reasons for failure during this assessment were due to weaker than expected staining within the distributed mid-er expressing sample or excessive inappropriate staining and morphology damage in the mid-expressor, and one participant showed inappropriate false positive staining in the ER-negative tumour (Fig 9B). Unexpected staining by this particular participant was also evident in the distributed tonsil section (see Fig 1), which showed a higher level of staining than the expected 2-5%. Antibody Pass Rates The most popular antibody used in this assessment run was the Ventana SP1 clone, used by 36% (n=99) of participants. This antibody showed an acceptable pass rate of 99%. The second most popular antibody was the Leica 6F11 clone, used by 28% (n=78), which showed a pass rate of 82%. The Dako EP1 antibody was used by 18% (n=51) of labs, and showed an acceptable pass rate of 98%. In-House Tissue Results 92% of participants also submitted their in-house controls for assessment. These showed an acceptable pass rate of 8%, and a further 16% of labs received a borderline pass. 4% of participants failed the assessment on their in house submitted slide. There were various reasons for a failed result, including excessive cytoplasmic staining, excessive pre-treatment, inappropriate non-specific staining, and poor tissue morphology. Several labs that received a borderline pass had not submitted the required composite control including a high, mid/low and a negative expressing tumour. Comparing NEQAS Slide and In-house Control. The NEQAS slides and in house material are scored differently, so participants should be cautious when comparing their UK NEQAS score directly to that of in-house scores. The two samples are scored on different factors: The UK NEQAS distributed slides are scored against the expected levels of staining as defined by 1) initial patient diagnosis for ER; 2) re-tested and validated staining of every 5th section from the relevant tissue blocks using the Leica 6F11 clone; 3) re-tested and validated samples by the relevant commercial companies to further verify the expected level of staining, including the Leica (clone 6F11), Dako (clone EP1) and the Ventana (clone SP1) antibodies. The in-house slides are scored on the staining quality, readability of the samples, the correct choice of tissue, and good morphological preservation. If the in-house control samples do not comprise of the required composite breast UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 17

20 The Breast Hormonal Receptor Module Run 16 tissue, containing a high (Allred 7-8), mid- (Allred 3-6) and a negative (Allred ) ER-expressing tumour, participants are given a maximum score of 12/2. The presence of normal staining glands is also encouraged. It is not possible to know the expected staining levels of the participants in-house slides and whether there are cases showing false-positive/negative staining, however assessors do try and highlight to participants where there are conflicting observation between self-assessment scores compared to the staining observed by the UK NEQAS assessors. Participants who are having difficulty in producing acceptable staining results are encouraged to contact UK NEQAS ICC so that further help and advice can be provided. References 1. Jennet M. Harvey, Gary M. Clark, C. Kent Osborne, and D. Craig Allred (1999) Estrogen Receptor Status by Immunohistochemistry Is Superior to the Ligand-Binding Assay for Predicting Response to Adjuvant Endocrine Therapy in Breast Cancer J Clin Oncol 17: Robin Leake, Diana Barnes, Sarah Pinder, Ian Ellis, Liz Anderson, Tom Anderson, Ruth Adamson, Tony Rhodes, Keith Miller and Rosemary Walker (2) J. Clin. Pathol. 2: Anderson E. The role of oestrogen and progesterone receptors in human mammary development and tumorigenesis. Breast Cancer Res. 22; 4: Fisher B, Costantino J, Redmond C, Poisson R, Bowman D, Couture J, Dimitrov NV, Wolmark N, Wickerham DL, Fisher ER, et al. (1989) A randomized clinical trial evaluating tamoxifen in the treatment of patients with node-negative breast cancer who have estrogen receptor-positive tumors. N Engl J Med 1989: Ciocca DR and Elledge R. Molecular markers for predicting response to tamoxifen in breast cancer patients. Endocrine. 2;13: Rhodes A, Jasani B, Balaton AJ, Barnes DM, Anderson E, Bobrow LG, Miller KD (21). Study of inter-laboratory reliability and reproducibility of estrogen and progesterone receptor assays in Europe. Documentation of poor reliability and identification of insufficient microwave antigen retrieval time as a major contributory element of unreliable assays. Am J Clin Pathol. 21 Jan;115(1): Ibrahim M, Dodson A, Barnett S, Fish D, Jasani B, Miller K. (28) Potential for false-positive staining with a rabbit monoclonal antibody to progesterone receptor (SP2): findings of the UK National External Quality Assessment Scheme for Immunocytochemistry and FISH highlight the need for correct validation of antibodies on introduction to the laboratory. Am J Clin Pathol. 129: Davies C, Godwin J, Gray R et al., (211) Relevance of breast cancer hormone receptors and other factors to the efficacy of adjuvant tamoxifen: patient-level meta-analysis of randomised trials. The Lancet 378: Acknowledgments We would like to thank Dako, Roche and Leica for agreeing to stain the UK NEQAS Breast hormonal receptor Gold standard slides with their respective assays. UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 18

21 RUN 16 BREAST STEROID HORMONE RECEPTOR Module Selected Images showing Optimal and Sub-optimal Immunostaining Fig 1. Optimally stained UK NEQAS distributed samples: (A) High ER expressing tumour shows intense staining in over 9% of neoplastic cells, while (B) mid-expressing ER tumour shows varying intensity of positivity in over 6% of neoplastic cells. Stained with the Dako EP1 pre-diluted antibody on the Dako Autostainer and pre-treatment in the PT link, high ph buffer. Fig 2. Optimal demonstration of ER in the UK NEQAS distributed samples. (A) The ER negative tumour remains unstained; only a certain percentage of the nuclei in the normal gland are positive. As expected, the tonsil (B) shows staining in 3-5% of the germinal centre lymphocytes, and the stomach remains unstained. Same protocol as Fig 1. Fig 3. Optimal demonstration of ER in the UK NEQAS (A) high and (B) mid -expressing tumours, showing the expected level of staining in both samples. Section stained with the Leica 6F11 antibody clone on the Leica BondMax, with antigen retrieval carried out using the ER1 buffer. Fig 4. Optimal level of expression in the UK NEQAS distributed samples stained with the Leica 6F11 antibody clone. As expected (A) the ER negative tumour remains unstained, while (B) the tonsil shows up to 5% of lymphocytes staining, and (C) the stomach is negative. (Same protocol as Fig 3). Fig 5. Good ER demonstration in the UK Neqas distributed samples. Both the (A) high and (B) mid -expressing tumours show the expected intensity and percentage of positive nuclei. The normal glands seen in section A also show the expected percentage of positive nuclei. Stained with the Ventana SP1 pre-diluted antibody on the Ventana Benchmark XT, CC1 mild AgR. Fig 6. Expected result in the UK Neqas distributed samples stained with the Ventana SP1 antibody. The negative tumour (A) remains unstained, and the tonsil shows staining in approximately 5% of the lymphocytes. (Same protocol as Fig 5). 19

22 RUN 16 BREAST STEROID HORMONE RECEPTOR Module Selected Images showing Optimal and Sub-optimal Immunostaining Fig 7. Good demonstration ER in the UK Neqas distributed high (A), mid (B) and negative (C) -expressing tumour. Each of the samples shows the expected level of staining. Section stained with the Leica 6F11 antibody, 1:5, on the Ventana ULTRA with CC1 standard antigen retrieval and iview detection kit. Fig 8. Two examples showing poor demonstration of ER in the mid-expressing tumour. The staining in (A) is much weaker and lower than expected, while the staining in (B) shows excessive background staining. Fig 9. Two examples showing suboptimal demonstration of ER in the negative expressing tumour. (A) Inappropriate non-specific staining of the lymphocytes, while (B) shows false positive staining of the tumour known to be negative. Sample (A) received a borderline pass, whereas, section (B) failed the assessment due to a potential for an incorrect diagnostic result. Fig 1. Suboptimal ER staining of the UK Neqas distributed tonsil section. The percentage of lymphocytes staining is much higher than expected, indicating that the sensitivity of the assay is too high. This section was from the same laboratory that showed false positive staining of the negative tumour (shown in Fig 9). Fig 11. Unacceptable ER staining of a participant s in house high expressing tumour: The staining is diffuse and looks more like non-specific cytoplasmic staining rather than true nuclear staining. The slide was unreadable and therefore failed the assessment. Fig 12. Good example of an in house control for ER. The multi-block section contains high, mid, and negative expressing tumours ( A-C respectively). A control containing tumours of known differing expression levels is important to gauge the sensitivity of the assay. Section stained with the Ventana SP1 clone on the Benchmark XT, CC1 mild AgR. 2

23 Run 16 BREAST STEROID HORMONE RECEPTOR Module GRAPHICAL REPRESENTATION OF PASS RATES 14 RUN 16E Oestrogen receptors on NEQAS Sections ALL PARTICIPANTS Individual 4 = (%) 5 = (%) 7 RUN 16E Oestrogen receptors on NEQAS Sections UK PARTICIPANTS Individual 4 = (%) 5 = (%) 12 6 = (%) 7 = (%) 6 6 = (%) 7 = (%) no. of returns = 2 (1%) 9 = 2 (1%) 1 = 6 (2%) 11 = 2 (1%) 12 = 13 (5%) 13 = 13 (5%) 14 = 12 (4%) 15 = 19 (7%) 16 = 126 (45%) no. of returns = (%) 9 = 2 (1%) 1 = 2 (1%) 11 = 1 (1%) 12 = 7 (5%) 13 = 1 (7%) 14 = 8 (5%) 15 = 15 (1%) 16 = 61 (41%) 2 17 = 52 (19%) 18 = 23 (8%) 1 17 = 28 (19%) 18 = 11 (7%) = 7 (3%) 2 = (%) = 2 (1%) 2 = (%) Summary Summary 16-2 = 28 (75%) 16-2 = 12 (69%) = 44 (16%) = 33 (22%) 1-12 = 21 (8%) 1-12 = 1 (7%) - 9 = 4 (1%) - 9 = 2 (1%) Median = 13.5 Median = RUN 16F Oestrogen receptors on in-house Sections ALL PARTICIPANTS Individual 4 = (%) 5 = (%) 6 = (%) 45 4 RUN 16F Oestrogen receptors on in-house Sections UK PARTICIPANTS Individual 4 = (%) 5 = (%) 6 = (%) 8 7 = (%) 8 = 3 (1%) 35 7 = (%) 8 = 1 (1%) no. of returns = 8 (3%) 1 = 3 (1%) 11 = 15 (5%) 12 = 27 (1%) 13 = 24 (9%) 14 = 25 (9%) 15 = 38 (14%) 16 = 85 (31%) 17 = 3 (11%) 18 = 1 (4%) no. of returns = 3 (2%) 1 = 3 (2%) 11 = 6 (4%) 12 = 6 (4%) 13 = 11 (8%) 14 = 2 (14%) 15 = 18 (12%) 16 = 43 (29%) 17 = 22 (15%) 18 = 1 (7%) = 5 (2%) 2 = (%) = 3 (2%) 2 = (%) Summary Summary 16-2 = 13 (48%) 16-2 = 78 (53%) = 87 (32%) = 49 (34%) 1-12 = 45 (16%) 1-12 = 15 (1%) - 9 = 11 (4%) - 9 = 4 (3%) Median = 13.5 Median =

24 Run 16 BREAST STEROID HORMONE RECEPTOR Module ANTIBODIES AND OTHER TECHNICAL PARAMETERS EMPLOYED BY PARTICIPANTS IN THE BREAST STEROID HORMONE RECEPTOR MODULE The following tables record the number of participants (N) using each primary antibody and the percentage (%) of these participants achieving acceptable staining (a score >12/2) on UK NEQAS sections. Breast Steroid Hormone Receptor Run: 16 Breast Steroid Hormone Receptor Run: 16 Primary Antibody : Oestrogen receptors Antibody Details N % Dako (EP1) M Dako (EP1) RTU Auto Plus IS Dako (EP1) RTU FLEX IR Dako FLEX (1D5) IR/IS Dako M3634 (SP1) 2 1 Dako M747 ER (1D5) 7 71 Leica Bond PA151 (6F11) 3 67 Leica/Novocastra NCL-ER-6F11 (6F11) Leica/Novocastra NCL-ER-6F11/2 1 9 Leica/Novocastra NCL-L-ER- 6F Leica/Novocastra RTU-ER-6F Menapath MP-93-CM1 1 Other 6 1 Thermo Fisher/ Neomarkers RM 911-S (SP1) Vector VP-E613/4 (6F11) 8 75 Ventana ER (6F11) 1 1 Ventana (SP1) Ventana (SP1) 32 1 Automation Oestrogen receptors N % BioGenex GenoMX 6i 1 1 Dako Autostainer 3 67 Dako Autostainer Link Dako Autostainer plus 5 6 Dako Autostainer Plus Link 5 8 Dako Omnis 1 1 LabVision Autostainer 4 1 Leica Bond Max 4 88 Leica Bond-III Menarini - Intellipath FLX 4 75 None (Manual) 3 1 Other 1 Shandon Sequenza 3 1 Ventana Benchmark 5 1 Ventana Benchmark ULTRA 62 1 Ventana Benchmark XT Breast Steroid Hormone Receptor Run: 16 Breast Steroid Hormone Receptor Run: 16 Heat Mediated Retrieval Oestrogen receptors N % Biocare Decloaking Chamber 3 33 Dako Omnis 1 1 Dako Pascal 1 1 Dako PTLink Lab vision PT Module 5 8 Leica ER1 1 mins 2 1 Leica ER1 2 mins Leica ER1 3 mins Leica ER1 4 mins 9 67 Leica ER2 1 mins 2 5 Leica ER2 2 mins Leica ER2 3 mins 8 1 Leica ER2 4 mins 3 67 Microwave 3 1 Other 2 1 Pressure Cooker 7 71 Ventana CC1 16mins 2 5 Ventana CC1 2mins 2 1 Ventana CC1 24mins 3 1 Ventana CC1 32mins 7 1 Ventana CC1 36mins 17 1 Ventana CC1 4mins 1 1 Ventana CC1 52mins 1 1 Ventana CC1 64mins 21 1 Ventana CC1 76mins 1 1 Ventana CC1 88mins 1 1 Ventana CC1 8mins 1 1 Ventana CC1 92mins 1 1 Ventana CC1 extended 5 1 Ventana CC1 mild Ventana CC1 standard 4 93 Ventana CC2 64mins 1 1 Ventana CC2 mild 1 1 Water bath OC 2 1 Enzyme Mediated Retrieval Oestrogen receptors N % AS PER KIT 4 1 NOT APPLICABLE

25 Run 16 BREAST STEROID HORMONE RECEPTOR Module Breast Steroid Hormone Receptor Run: 16 Breast Steroid Hormone Receptor Run: 16 Detection Oestrogen receptors N % A Menerini Polymer (MP-XCP) 1 1 AS PER KIT Dako EnVision FLEX ( K8/1) 8 88 Dako EnVision FLEX+ ( K82/12) Dako Envision HRP/DAB ( K57) 4 75 Dako Envision+ HRP mouse K44/5/6/7 1 1 Dako Envision+ HRP rabbit K48/9/1/ Dako REAL HRP/DAB (K51 ) 1 1 LabVision UltraVision LP HRP (TS 125 HD) 1 1 LabVision UltraVision ONE Polymer ( TL-12/5-HDJ/T 1 1 Leica Bond Polymer Refine (DS98) 71 9 MenaPath X-Cell Plus (MP-XCP) 2 5 None 2 5 NOT APPLICABLE 1 Novocastra Novolink PDS (RE7-14/15/28/29-K) 1 1 Other 1 9 Vector ImmPRESS Universal (MP-75) 1 Ventana iview system (76-91) 8 1 Ventana OptiView Kit (76-7) 9 89 Ventana UltraView Kit (76-5) Chromogen Oestrogen receptors N % AS PER KIT BioGenex liquid DBA (HK-124-7K) 1 1 Dako DAB K Dako DAB+ REAL Detection (K51) 2 5 Dako EnVision Plus kits 4 75 Dako FLEX DAB Dako REAL EnVision K57 DAB 4 75 Dako REAL K51 DAB 1 1 Leica Bond Polymer Refine kit (DS98) 7 9 menapath xcell kit DAB (MP-86) 3 67 Other Sigma DAB (D5637) 1 1 Sigma DAB (D595) 1 1 Ventana DAB 4 1 Ventana Enhanced Alk. Phos. Red Detection Kit 1 1 Ventana iview 7 1 Ventana Ultraview DAB Vision BioSystems Bond X DAB 1 1 BEST METHODS A selection from just a few of the best methods employed by participants Oestrogen receptors - Method 1 Participant scored 18/2 (UK NEQAS Slide) and 17/2 (In House slide) using this method. Primary Antibody: Dako (EP1) RTU FLEX IR84, 2 Mins, 24 ºC Prediluted Automation: Dako Autostainer Link 48 Dako FLEX+ kit Main Buffer: Dako FLEX wash buffer Dako PTLink, Buffer: Dako target retrieval solution, PH: 9 NOT APPLICABLE Chromogen: Dako FLEX DAB, 24 ºC., Time 1: 5 Mins, Time 2: 5 Mins Detection: None Oestrogen receptors - Method 2 Participant scored 17/2 (UK NEQAS Slide) and 16/2 (In House slide) using this method. Primary Antibody: Thermo Fisher/ Neomarkers RM 911-S (SP1), 4 Mins, 23 ºC Dilution 1: 125 Automation: Dako Autostainer Link 48 Dako FLEX+ kit Main Buffer: Dako FLEX wash buffer Dako PTLink, Buffer: Target retrieval solution, PH: 9 Chromogen: Dako EnVision Plus kits, 23 ºC., Time 1: 1 Mins Detection: Other, 15 Mins, 23 ºC Prediluted Oestrogen receptors - Method 3 Participant scored 18/2 (UK NEQAS Slide) and 16/2 (In House slide) using this method. Primary Antibody: Ventana (SP1), 16 Mins, 36 ºC Prediluted Automation: Ventana Benchmark ULTRA Ventana UltraView DAB Main Buffer: Ventana reaction buffer (95-3), PH: 7.4 Ventana CC1 standard, Buffer: Ultra CC1(95-224), PH: 8.5 Chromogen: Ventana Ultraview DAB, 36 ºC., Time 1: 8 Mins, Time 2: 4 Mins Detection: Ventana UltraView Kit (76-5), 8 Mins, 36 ºC Prediluted 23

26 Run 16 BREAST STEROID HORMONE RECEPTOR Module Oestrogen receptors - Method 4 Participant scored 18/2 (UK NEQAS Slide) and 16/2 (In House slide) using this method. Primary Antibody: Leica/Novocastra NCL-L-ER- 6F11, 3 Mins, RT ºC Dilution 1: 5 Automation: Leica Bond-III Leica BondMAx Refine KIT Main Buffer: Bond Wash Buffer (AR959) Leica ER1 3 mins Chromogen: Leica Bond Polymer Refine kit (DS98), Time 2: 1 Mins Detection: Leica Bond Polymer Refine (DS98), 8 Mins, RT ºC Prediluted 24

27 The Breast HER2 ICC Module Run 16 Merdol Ibrahim, Suzanne Parry and Keith Miller Antigen Assessed: Sections Circulated: HER2 Number of Registered Participants: 436 Number of Participants This Run 322 (74%) Composite slide consisting of 4 breast carcinoma cell lines (see table below) Specific Guideline Used in The Assessment of Slides: The immunohistochemical results were evaluated by UK NEQAS assessors scoring independently using an adapted method initially devised by the Clinical Trials Assay. Due to the nature of cell lines, overall percentage staining criteria can not be accurately applied when scoring cell lines. Communication with Leica Microsystems, with whom UK NEQAS ICC & ISH developed the cell lines, indicate that the expected level of membrane staining for a given cell line may range from 3-9% from the viable cell line population. For this reason when cutting sections, every 5th section is stained as a reference point to gauge the expected level of staining throughout the cell block/s. The table below demonstrates the staining patterns looked for in cell lines and in-house tissue sections during assessments. Cell line position (slide label end) Assessment of Cell Line Staining Pattern Assessment of In-House Tissue Sections* *The ASCO/CAP guidelines by Wolff et al., (213) are not currently used for scoring in-house controls A: 3+ Strong complete cell membrane staining Strong complete cell membrane staining in >3% of tumour cells B: 2+ Weak to moderate complete cell membrane staining Weak to moderate complete cell membrane staining in >1% of tumour cells C: 1+ Faint barely perceptible incomplete membrane staining Faint barely perceptible incomplete membrane in >1% of cells staining D: Negative No staining in the control cell line No staining in the control cell line Validation of Distributed UK NEQAS ICC & ISH Cell Lines IHC Validation using: Leica Oracle, Ventana Pathway 4B5 and Dako HercepTest ISH Validation using: FISH (Abbott Vysis) and DDISH (Roche) Updated Assessment and Scoring Procedure (Applicable as of Run 16) UK NEQAS Specific Membrane Scoring Algorithm: UK NEQAS ICC & ISH devised an EQA specific algorithm for scoring the cell lines so as to provide participants additional technical feedback. As well as taking into account the expected range (3-9% see above) of cell line membrane staining, the acceptable staining levels of each of the cell lines are as follows: Table below illustrates the expected level of staining for each of the circulated breast carcinoma cell lines along with the UK NEQAS scoring criteria used in assessments ONLY. Cell line Score Acceptable Level/s of Staining During Assessments Description of Staining Pattern Used By the Assessors SK-BR only The 3+ cell line has a wide threshold of complete membrane staining showing strong staining. Only this level of membrane staining is deemed acceptable for this cell line. MDA-MB or 1+/2+ or 2+/3+ or 2+/1+ or 3+/2+ i) 1+/2+ or 2+/1+: Staining is slightly weaker than expected with membrane showing more 1+ compared to 2+ (1+/2+) or 2+ membrane staining is present but also showing 1+ staining (2+/1+). ii) 2+/3+ or 3+/2+: Staining is slightly weaker than expected with membrane showing more 2+ compared to 3+ (2+/3+) or 3+ membrane staining is present but also showing 2+ staining (3+/2+). MDA-MB or /1+ or 1+/ i) /1+: Staining is more towards the weaker end of 1+ staining but still acceptable. ii) 1+/: Staining is more towards the weaker end of 1+ staining but still acceptable. MDA-MB-231 Negative or /1+ or 1+/ /1+ or 1+/ = Cells are starting to show very weak membrane staining U /Uninterpretable Assessors may also give a score of 'U' which indicates that the cell lines were 'uninterpretable due to the reasons set out below. Borderline Pass: A score of U/x e.g. U/3+ or U/2+ or U/1+ or U/ indicates a borderline uninterpretable scores indicating that the staining is just about readable and further improvements are required. UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 25

28 The Breast HER2 ICC Module Run 16 Numerical Scoring Criteria Once the team of assessors have assessed the membrane interpretation for each of the NEQAS samples, they will award a potential score out of 5 marks; based on the interpretability of the membrane staining and technical staining quality. An overall pass mark is then awarded by combining the four assessors scores to give scores out of 2. Score and Interpretation 16-2/2: Excellent 13-15/2: Acceptable Interpretation Overall the staining is at the expected level for each of the samples. Some slight technical issues noted by some of the assessors, but overall the staining is suitable for interpretation 1-12/2: Borderline Acceptable 4-9/2: Unacceptable Overall the samples are borderline interpretable (still clinically relevant) indicating that technical improvements need to be made. Marks may have been deducted due to: Weaker/stronger staining than expected, cytoplasmic staining, morphological damage etc Overall the samples are of unacceptable quality for clinical interpretation and technical improvements need to be made. Marks may have been deducted due to: False positive/negative membrane staining, excessive cytoplasmic staining, excessive morphological damage, staining of normal glands Introduction HER2 immunohistochemistry has been routinely used as a predictive marker in breast cancer for more than 1 years. Patients with HER2 positive tumours generally have a poor overall prognosis. However, patients with metastatic disease showed an improved rate of survival when the humanized monoclonal antibody Trastuzumab (Herceptin) was given alone or added with chemotherapy treatment (Slamon et al., 1998). Herceptin, in the metastatic setting was made available in the UK in 2 by the UK National Institute for Clinical Excellence (NICE). In 25, Herceptin in the adjuvant setting was also shown to be effective in primary breast cancers by reducing the risk of recurrence and mortality (see articles by Wolff et al., (26) and Walker et al., (28), and in 26 NICE approved primary breast screening for HER2. In the UK alone this has resulted in approximately 4, breast cancers per year being tested for HER2. With so many more cases now being tested worldwide, it is imperative that correct protocols and methodologies are followed and adhered to. The UK NEQAS HER2 IHC module assesses the technical quality of staining achieved by laboratories and provides feedback to laboratories to help improve the analytical phase of patient diagnosis. Two recommended HER2 testing guidelines are: 1. The ASCO/ CAP guidelines by Wolff et al. (27), and 2. the UK guidelines by Walker et al., (28). Please note, the ASCO/CAP guidelines have recently been updated (Wolff et al., 213). Both publications provide invaluable guidelines covering interpretation, tissue fixation and antibody validation. The article by Walker et. al., also provides guidelines on the minimum number of tests per year that labs should be carrying out in order to provide a robust HER2 IHC service. The recommendations are 25 tests for HER2 IHC and 1 tests for HER2 ISH. Participants who continue to have difficulty in producing acceptable staining results should be aware that the members of staff at UK NEQAS ICC are always willing to assist any laboratory that is struggling to meet the required standard of HER2 immunostaining, and would be happy to receive requests for assistance via in the first instance (see front of journal for contact details). clinical interpretation. UK NEQAS ICC & ISH therefore recommends the following: In-house control tissue (or cell lines) must include 3+, 2+ and 1+/ invasive breast cancer cases. However, it has become quite apparent that as patient tumour size and respective biopsies become smaller laboratories may be having difficulty finding appropriate invasive control material. It is therefore acceptable, if laboratories are having problems in finding appropriate invasive control material, to submit DCIS tissue showing differing levels of membrane staining as an acceptable alternative. Laboratories must indicate on their datasheet which component of the tumour they have scored otherwise the invasive component (if present) will be assessed. Important: The Ventana OptiView detection system has not been validated for use with the 4B5 HER2 for IHC. Any laboratory using this system for breast HER2 testing should be aware that they are doing so off label usage. References 1. Slamon D, Leyland-Jones B, Shak S, et al. Addition of Herceptin (humanised anti-her2 antibody) to first line chemotherapy for (HER2+/MBC) markedly increases anticancer activity: a randomised, multinational controlled phase III trial. Proc ASCO 1998;17:98a. 2. Piccart-Gebhart MJ, Procter M, Leyland-Jones B, et al. Trastuzumab after adjuvant chemotherapy in HER2-positive breast cancer. N Engl J Med 353: , Bartlett JM, Ibrahim M, Jasani B, et al. External quality assurance of HER2 FISH testing: results of a UK NEQAS pilot scheme. J Clin Pathol 27 6 (7): Walker RA, Bartlett JM, Dowsett M, Ellis IO, Hanby AM, Jasani B, Miller K, Pinder SE. HER2 testing in the UK: further update to recommendations. J Clin Pathol (7): Wolff AC, Hammond MEH, Schwartz JN, et al. American Society of Clinical Oncology/College of American Pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. J Clin Oncol 27;25: Wolff AC, et al. Recommendations for Human Epidermal Growth Factor Receptor 2 Testing in Breast Cancer: American Society of Clinical Oncology/ College of American Pathologists Clinical Practice Guideline Update. J Clin Pathol. 213; 31 (31): Acknowledgments We would like to thank Dako, Ventana and Leica for agreeing to stain the UK NEQAS Breast HER2 ICC Gold standard samples with their respective kits/assays. In-House Control Tissue Recommendations Correct choice of in-house control tissue and good morphological preservation is paramount to gauge the sensitivity of the HER2 test, which can have an impact on UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 26

29 RUN 16 BREAST HER2 ICC Module Selected Images showing Optimal and Sub-optimal Immunostaining Fig 1. Appropriate level of staining of the UK NEQAS SK-BR3 (3+) cell line. The example shows strong membrane staining, which is completely circumferential with minimal cytoplasmic staining. Section stained with the Dako HercepTest. Fig 2. Two examples showing the appropriate level of staining of the UK NEQAS SK-BR3 (3+) cell line. Both sections show strong complete membranous staining. (A) Stained with the Ventana Pathway, and (B) with the Leica Oracle kit. Fig 3. Two examples of inappropriate staining of the UK NEQAS distributed SK-BR3 (3+) cell line. (A) The level of staining is lower than expected and more representative of a 2+. (B) Excessive membrane and cytoplasmic staining, making the sample difficult to interpret. Fig 4. Optimal HER2 staining of the MDA-MB-453 (2+) cell line: The majority of cells show complete membrane staining which is less intensive than that seen with the 3+ cell line. Section stained with the Leica Oracle kit. Fig 5. Two examples showing the appropriate level of HER2 staining of the MDA-MB-453 (2+) cell line: Both sections show the expected weak to moderate complete membrane staining. (A) Stained with the Dako HercepTest and (B) stained with the Ventana Pathway kit. Fig 6. Two examples of inappropriate levels of staining of the UK NEQAS MDA-MB-453 (2+) cell line. (A) The staining is too strong and more representative of a 3+, while the staining in (B) is too low and looks more like a 1+ (compare both images to Figs 4&5). 27

30 RUN 16 BREAST HER2 ICC Module Selected Images showing Optimal and Sub-optimal Immunostaining Fig 7. Two examples showing the expected level of staining of the UK NEQAS 1+ cell line (MDA-MB-175). Over 1% of the tumour cells show fine incomplete membrane staining. Fig 8. Inappropriate levels of staining of the UK NEQAS MDA-MB-175 (1+) cell line. The staining in (A) is too strong and more representative of a 2+, while (B) shows no staining, and is therefore assessed as a negative result. Fig 9. Unacceptable staining of the UK NEQAS MBA-MB-231 cell line. This sample should be n n n n n n inappropriate excessive membrane and cytoplasmic staining. Fig 1. Examples of unacceptable staining: (A & C) Excessive cytoplasmic staining and morphology damage seen on the participants own in house control tissue is also shown on the UK NEQAS cell lines (B & D). Samples A & B were 3+ and C & D 2+. F! " n n n n n # n n pathway kit (see Fig 12 also). F F! n n n and negative (B) levels of expression. 28

31 Run 16 BREAST HER2 ICC Module GRAPHICAL REPRESENTATION OF PASS RATES no. of returns RUN 16E1 HER-2 on NEQAS Sections Individual 4 = 19 (6%) 5 = 5 (2%) 6 = 8 (2%) 7 = 5 (2%) 8 = 27 (8%) 9 = 14 (4%) 1 = 12 (4%) 11 = 21 (7%) 12 = 18 (6%) 13 = 18 (6%) 14 = 23 (7%) 15 = 4 (12%) 16 = 73 (23%) 17 = 21 (7%) 18 = 11 (3%) 19 = 6 (2%) 2 = 1 (%) no. of returns RUN 16Fi HER-2 on in-house Sections Individual 3 = (%) 4 = 7 (2%) 5 = 1 (%) 6 = 6 (2%) 7 = 9 (3%) 8 = 34 (11%) 9 = 15 (5%) 1 = 12 (4%) 11 = 18 (6%) 12 = 28 (9%) 13 = 23 (8%) 14 = 21 (7%) 15 = 29 (1%) 16 = 75 (25%) 17 = 7 (2%) 18 = 5 (2%) 19 = 3 (1%) 2 = 3 (1%) Summary Summary 16-2 = 112 (35%) 16-2 = 93 (31%) = 81 (25%) = 73 (25%) 1-12 = 51 (16%) 1-12 = 58 (2%) - 9 = 78 (24%) - 9 = 72 (24%) Median = 12. Median = 12. ANTIBODIES AND OTHER TECHNICAL PARAMETERS EMPLOYED BY PARTICIPANTS IN THE BREAST HER2 ICC MODULE The following tables record the number of participants (N) using each primary antibody and the percentage (%) of these participants achieving acceptable staining (a score >12/2) on UK NEQAS sections. 29

32 Run 16 BREAST HER2 ICC Module Breast HER2 ICC Run: 16 Breast HER2 ICC Run: 16 HER-2 Primary Antibody N % Biocare CME 342 A,B (EP145Y) 2 BioGenex (EP145Y) rb mono 6 Biogenex AM134-5M (CB11) 3 Cell Marque CMA 61 (CB11) 3 Dako A485 C-erB-2 (poly) Dako HercepTest K524 (poly) 5 2 Dako HercepTest K527 (poly) 5 4 Dako Link HercepTest SK1 (poly) Labvision / Neomarkers RM-913 (SP3) 1 2 Leica Oracle HER2 Bond IHC (CB11) Neomarkers MS-73-P 1 Novocastra NCL-CB11 (CB11) 3 Novocastra NCL-L-CB11 (CB11) 3 Novocastra NCL-L-CBE356 (1A7) 2 1 Novocastra RTU-CB11 (CB11) 3 Automation N % BioGenex GenoMX 6i 3 33 Dako Autostainer 4 25 Dako Autostainer Link Dako Autostainer plus 5 2 Dako Autostainer Plus Link 6 5 LabVision Autostainer 3 Leica Bond Max Leica Bond-III Menarini - Intellipath FLX 1 None (Manual) 37 5 Other 4 25 Shandon Sequenza 2 5 Ventana Benchmark Ventana Benchmark ULTRA Ventana Benchmark XT 8 8 Ventana NexES 1 1 Novocastra RTU-CBE-356 (1A7) 1 Other Vector CB11 VP-C Ventana Confirm 79/ (4B5) Ventana Pathway 79-1 (4B5) 9 89 Ventana Pathway (4B5)

33 Run 16 BREAST HER2 ICC Module Breast HER2 ICC Run: 16 HER-2 Breast HER2 ICC Run: 16 HER-2 Heat Mediated Retrieval N % Biocare Decloaking Chamber 8 Dako Pascal 1 Dako PTLink Lab vision PT Module 3 Leica ER1 1 mins 4 5 Leica ER1 2 mins Leica ER1 25 mins 8 1 Leica ER1 3 mins 3 1 Leica ER2 2 mins 1 Microwave None 1 Other 9 67 Pressure Cooker 6 Steamer 1 Ventana CC1 16mins 2 1 Ventana CC1 2mins 3 33 Ventana CC1 32mins 1 7 Ventana CC1 36mins Ventana CC1 4mins 3 33 Ventana CC1 52mins 3 67 Ventana CC1 56mins 3 1 Ventana CC1 64mins 7 1 Ventana CC1 8mins 5 6 Ventana CC1 mild Ventana CC1 standard Water bath OC Detection N % AS PER KIT Biocare SLAB (STU HRP 7H,L1) 1 BioGenex HRP (HK 519-6K) 2 BioGenex SS Polymer (QD 42-YIKE) 1 BioGenex SS Polymer (QD 43-XAKE) 3 Dako HerCep Test (K524) 4 25 Dako EnVision FLEX ( K8/1) 11 Dako EnVision FLEX+ ( K82/12) 4 5 Dako Envision HRP/DAB ( K57) 7 Dako Envision+ HRP rabbit K48/9/1/11 1 Dako HerCep Test Autor (K527) 4 5 Dako HerCep Test Autor (SK1) 9 78 Dako REAL HRP/DAB (K51 ) 1 LabVision UltraVision LP HRP (TL 125 HLJ) 1 LabVision UltraVision LP HRP (TS 125 HD) 1 LabVision UltraVision ONE Polymer ( 2 TL-12/5-HDJ/T Leica Bond Polymer Refine (DS98) Novocastra Novolink PDS 1 (RE7-14/15/28/29-K) Other Ventana iview system (76-91) 1 7 Ventana OptiView Kit (76-7) 6 67 Ventana UltraView Kit (76-5) Breast HER2 ICC Run: 16 HER-2 Breast HER2 ICC Run: 16 HER-2 Enzyme Retrieval N % AS PER KIT Enzyme digestion + HIER 1 NOT APPLICABLE Other 3 33 Ventana Protease 2 1 Ventana Protease 1 (76-218) 1 Chromogen N % AS PER KIT BioGenex DAB (QD43) 2 BioGenex Liquid DAB (HK153-5K) 2 BioGenex liquid DBA (HK-124-7K) 3 33 DAKO DAB+ 2 5 Dako DAB+ Liquid (K3468) 5 2 Dako DAB+ REAL Detection (K51) 2 Dako EnVision Plus kits 2 1 Dako FLEX DAB Dako REAL EnVision K57 DAB 9 11 LabVision DAB 1 Leica Bond Polymer Refine kit (DS98) Other Sigma DAB (D5637) 1 Sigma DAB (D595) 1 Ventana DAB 4 5 Ventana iview 8 75 Ventana Ultraview DAB

34 Run 16 BREAST HER2 ICC Module BEST METHODS A selection from just a few of the best methods employed by participants HER-2 - Method 1 Participant scored 16/2 (UK NEQAS Slide) and 16/2 (In House slide) using this method. Primary Antibody: Leica Oracle HER2 Bond IHC (CB11), 3 Mins, RT ºC Prediluted Automation: Leica Bond-III Leica BondMAx Refine KIT Main Buffer: Bond Wash Buffer (AR959) Leica ER1 25 mins, PH: 6 Chromogen: Leica Bond Polymer Refine kit (DS98), RT ºC., Time 1: 1 Mins Detection: Leica Bond Polymer Refine (DS98), 1 Mins, RT ºC Prediluted HER-2 - Method 2 Participant scored 19/2 (UK NEQAS Slide) and 16/2 (In House slide) using this method. Primary Antibody: Ventana Pathway (4B5), 12 Mins Prediluted Automation: Ventana Benchmark ULTRA Ventana UltraView DAB Main Buffer: Ventana reaction buffer (95-3) Ventana CC1 mild Chromogen: Ventana Ultraview DAB Detection: Ventana UltraView Kit (76-5), 8 Mins Prediluted HER-2 - Method 3 Participant scored 17/2 (UK NEQAS Slide) and 16/2 (In House slide) using this method. Primary Antibody: Dako Link HercepTest SK1 (poly), 3 Mins, 25 ºC Prediluted Automation: Dako Autostainer Link 48 Dako FLEX+ kit Main Buffer: Dako FLEX wash buffer Dako PTLink, Buffer: HercepTest Eitope Retrieval Solution NOT APPLICABLE Chromogen: Dako EnVision Plus kits, 25 ºC., Time 1: 1 Mins Detection: Dako HerCep Test Autor (SK1), 3 Mins, 25 ºC Prediluted 32

35 The Gastric HER2 ICC Module Run 16 Merdol Ibrahim and Suzanne Parry Antigen Assessed: The NEQAS distributed sections consisted of a composite slide containing 4 surgical /resections of intestinal gastric carcinoma: HER2 for IHC 2 NEQAS blocks were used to distribute slides for this assessment. The expected expression levels of each tissue sample were: Block1: A. 3+ B. 2+ C. 2+, 1+ or depending on the serial section received by the laboratory D. (Please note: A few laboratories received slides with no tumour in section D) Number of Registered Participants 117 Number of Participants This Run 76 (65%) Block2: A. 3+ B. 2+ C. 1+ D. 1+ or depending on the serial section received by the laboratory Score (negative) 1+ (negative) 2+ (equivocal*) 3+ (positive) Table: 1 Gastric HER2 ICC Membrane Scoring Guidelines by Hoffman et al., (28) and Rüschoff et al., (21) Surgical / resections As used in NEQAS assessments No staining in < 1% of tumour cells Faint barely perceptible incomplete membrane staining in >1% of cells staining Weak/ moderate complete, basolateral or lateral membrane reactivity in > 1% of tumour cells Strong complete, basolateral or lateral membrane reactivity in > 1% of tumour cells Biopsies No staining in any of the tumour cells Tumour cells with faint barely perceptible incomplete membrane staining, irrespective of percentage of tumour cells stained Tumour cells with weak/ moderate complete, basolateral or lateral membrane reactivity irrespective of % tumour cells stained Tumour cells with strong, basolateral or lateral membrane reactivity irrespective of percentage of tumour cells stained * Equivocal cases should be refluxed to ISH testing. Note: in the UK, NICE guidelines recommend that only 3+ cases are put forward for Trastuzumab treatment, see Validation of Distributed Samples IHC Validation of Distributed Samples The NEQAS tissue sections are validated prior to sending out: Gold standard slides are stained at every 25th serial level as part of the tissue quality control and to assess the HER2 expression level throughout the blocks. The slides are also stained with each of the commercially available kits, i.e. the Ventana Pathway 4B5, Dako HercepTest and Leica Oracle systems This is taken into consideration when scoring participants slides and also noted prior to assessments when staining every 5th sample for quality control purposes. Important: The Leica Oracle HER2 IHC System is not currently supported for gastric HER2 IHC in the UK and any laboratory using the kit for gastric purposes should be aware that they are doing so off label usage and would need to fully validate the Leica Oracle system prior to diagnostic use. Table 2: Section Label From slide Label End A (Block 1) A (Block 2) B (Block 1) B (Block 2) C (Block 1) C (Block 2) D (Block 1) D (Block 2) Staining Pattern with IHC ,1+ or or HER2 status by ISH* Amplified Amplified Amplified Amplified Non-Amplified Non-Amplified Non-Amplified Non-Amplified UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 33

36 The Gastric HER2 ICC Module Run 16 Assessment Procedure Assessment is carried out around a multi-header microscope with each slide being assessed by 4 independent assessors and a microscope lead. Each of the NEQAS samples (A-D) (as well as in-house samples where submitted) are individually assessed, with each assessor providing membrane interpretation (see table 3), and an individual score out of 5 (see table 4) based on interpretability of membrane staining and technical feedback. The four assessors scores are then combined to give an overall score out of 2 and it s interpretation (see table 4). Table 3: UK NEQAS Specific Membrane Scoring Criteria UK NEQAS ICC & ISH uses an EQA specific scoring criteria when scoring the tissue sections, so as to provide participants additional technical feedback, which is illustrated below. Expected membrane staining Scoring criteria used by UK NEQAS ICC & ISH 3+ i) 3+: as expected Ii) 2+/3+ or 3+/2+: Staining is slightly weaker than expected with membrane showing more 2+ compared to 3+ (2+/3+) or 3+ membrane staining is present but also showing 2+ staining (3+/2+). 2+ i) 1+/2+ or 2+/1+: Staining is slightly weaker than expected with membrane showing more 1+ compared to 2+ (1+/2+) or 2+ membrane staining is present but also showing 1+ staining (2+/1+). ii) 2+/3+ or 3+/2+: Staining is slightly weaker than expected with membrane showing more 2+ compared to 3+ (2+/3+) or 3+ membrane staining is present but also showing 2+ staining (3+/2+). 1+ i) /1+: Staining is more towards the weaker end of 1+ staining but still acceptable. ii) 1+/: Staining is more towards the weaker end of 1+ staining but still acceptable. Neg. /1+ or 1+/ = Staining starting to show very weak membrane staining U = Uninterpretable: Assessors may also give a score of 'U' which indicates that the cell lines / tissue sections were 'uninterpretable due to the reasons set out below. U/x = Borderline interpretable. A score of U/x e.g. U/3+ or U/2+ or U/1+ or U/ indicates that the staining is just about readable and further improvements are required. Any other membrane score other that assigned for each of the expected scores are deemed as unacceptable Individual Assessor Score Out of 5 Marks Table 4: Numeric Scoring Criteria used to Calculate the Overall Pass Mark Combined Assessor Out of 2 Marks Slide not submitted for assessment 1 & = Unacceptable = Borderline 4 & = Acceptable 16-2 = Acceptable/ Excellent Standard Introduction Immunohistochemical testing of HER2 status is now routinely used in breast cancer testing and is recognised as a prognostic and predictive marker, generally used alongside breast hormonal receptor markers ER/PR. Patients who are HER2 positive (IHC 3+ and IHC 2+/ISH+ ) have been shown to benefit from Herceptin (Trastuzumab) therapy and increased overall survival rate. More recently the Trastuzumab for Gastric Cancer (ToGA) study, which investigated Trastuzumab in HER2 positive advanced gastric cancer (Bang et al., 21) showed overall median survival of nearly 3 months. Similar to breast cancer, the ToGA trial showed an increased benefit from Trastuzumab treatment for patients Overall Score Interpretation Overall the samples are of unacceptable quality for clinical interpretation and technical improvements need to be made. Marks may have been deducted due to: False positive / negative membrane staining Excessive cytoplasmic staining Excessive morphological damage Excessive staining of normal glands Overall the samples are borderline interpretable indicating that technical improvements need to be made. Marks may have been deducted due to: Weaker / stronger than expected membrane staining Some cytoplasmic staining Morphological damage Overall the samples show acceptable membrane staining and are suitable for interpretation. Further comments are also provided on individual participant reports. Other factors which may lead to a low score include: excessive/ insufficient haematoxylin staining; poor quality of in-house control tissue; poor/inadequate choice of control tissue; poor/inadequate fixation; excessive antigen retrieval etc. showing 3+ IHC and IHC 2+/FISH+ expression. The initial development of the HER2 scoring criteria was developed as a precursor to the ToGA trial (Hoffman et al., 28) with the study group modifying the breast HER2 IHC scoring algorithm to compensate for the incomplete membrane staining and greater tumour heterogeneity seen in gastric cancers. A different scoring system was also established for resection and core biopsies as illustrated in table 1. A more recent article by Rüschoff et al., (21) has validated the scoring procedure further with a detailed approach to stepwise HER2 IHC scoring in gastric cancers. The article also compared the Ventana 4B5 assay with the HercepTest and indicated a UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 34

37 The Gastric HER2 ICC Module Run 16 tendency towards higher sensitivity for 4B5 detecting positive HER2 amplification but there was no difference between both test platforms with respect to the ISH-negative cases, with the one equivocal case testing negative at some sites and as positive at others. Rüschoff and colleagues also used both fluorescent (Dako PharmDX or PathVysion Vysis) and chromogen (Ventana DISH) to confirm their IHC findings. Assessment Results Features Of Acceptable Staining: (Figs 1, 4, 7, 8, & 9) Membrane staining of the invasive tumour with the expected expression level Cytoplasmic staining not excessive No background staining of stromal tissues or inappropriately localised staining Features Of Suboptimal or Unacceptable Staining: (Figs 2, 3, 5, 6 & 1) Weaker/lower or stronger/higher than the expected expression level of membrane staining in the invasive tumour (Figs 2, 5 & 1) False positive or negative membrane staining Excessive cytoplasmic staining Excessive background staining or inappropriately localised staining (Fig 3B) Morphological damage (Fig 6) Excessive staining of normal glands Additional Comments a. The Ventana 4B5 is known to stain intestinal metaplasia (not illustrated) and participants are not penalised when this staining is observed. b. The Leica Oracle HER2 IHC System is not currently supported for gastric HER2 IHC in the UK and any laboratory using the kit for gastric purposes should be aware that they are doing so off label usage and would need to fully validate the Leica Oracle system prior to diagnostic use. Pass Rates Only seventy-six (65%) of participants took part in the current assessment, as several of our overseas participants were unable to send their slides. Results from the NEQAS distributed samples showed an overall acceptable pass rate of 49%; a decrease of 19% compared to the last assessment (run 15). A further 2% of participants received a borderline pass. The unacceptable rate was slightly higher than that of the previous run, with 31% of labs receiving a score of less than 1 (not including non-submissions). The main reasons for unacceptable scores were due to weak staining or lower membrane expression than expected. Inappropriate excessive membrane staining was another reason for unacceptable results, which was mainly caused by excessive antigen retrieval. Some participants were also marked down for excessive counterstain if the haematoxylin was too heavy, and therefore obscuring the membrane staining and making interpretation difficult. Of the 76 participants who submitted their NEQAS slides, 61 labs (8%) also submitted their in-house controls. The pass rates for these were better than a those achieved for the NEQAS slides, with 72% (N=44) of participants receiving an acceptable pass, and a further 23% (N=14) receiving a borderline pass. This was also better than the previous run inhouse tissue pass rates, where 66% achieved an acceptable pass. In the current run only 3 labs (5%) received a fail on their in house material. This was mainly due to poor tissue quality. Several laboratories did not submit the required composite control material, consisting of a 3+, 2+ and 1+/ expressing gastric/breast control tumour samples. These labs were given a maximum score of 3/5 from each assessor (12/2 overall). Methodologies As with previous assessment runs, the majority of participants (71%) are using the Ventana 4B5 assay for their gastric HER2 testing. This method showed an overall acceptable rate of 6% on the Neqas distributed tissue for this run. The Dako HercepTest was used by 9% (n=7), and this showed an overall pass rate of 33%. The rest of the labs (2%) used a home-brew method, and these showed an overall pass rate of 7%. A variety of antibodies and antigen retrieval protocols, including the Dako Pascal and microwave, were used in the home-brew methods. Please also note, when using a standardised kit/assay, such as the Dako HercepTest or Ventana s 4B5, it is important to use the manufacturers recommended protocol. Control Tissue and Recommendations It was left an open choice for laboratories to submit either gastric or breast control material, although gastric controls are the preferred tissue choice for this module. However, laboratories were not penalised if they had submitted breast carcinomas. Whether gastric or breast tissue is used, laboratories should still submit in-house controls showing 3+, 2+ and 1+/ levels of membrane expression. Due to the heterogenic nature of many gastric tumours, it is acceptable for labs to submit a heterogeneous in-house control with areas of both 3+ and 2+ membrane expression for example, as long as the participant clearly indicates the areas where the assessment should be carried out. Important: The Ventana OptiView detection system has not been validated for use with the 4B5 HER2 for IHC. Any laboratory using this system for gastric HER2 testing should be aware that they are doing so off label usage. References: 1. Hofmann M, Stoss O, Shi D et al. Assessment of a HER2 scoring system for gastric cancer: results from a validation study. Histopathology (7): Rüschoff J, Dietel M, Baretton G etc al. HER2 diagnostics in gastric cancerguideline validation and development of standardized immunohistochemical testing. Virchows Arch (3): Bang YJ, Van Cutsem E, Feyereislova A, Chung HC, Shen L, Sawaki A, Lordick F, Ohtsu A, Omuro Y, Satoh T, Aprile G, Kulikov E, Hill J, Lehle M, Rüschoff J, Kang YK. ToGA Trial Investigators. Trastuzumab in combination with chemotherapy versus chemotherapy alone for treatment of HER2- positive advanced gastric or gastro-oesophageal junction cancer (ToGA): a phase 3, open-label, randomised controlled trial. Lancet (9742): Acknowledgments We would like to thank Dako, Ventana and Leica for agreeing to stain the UK NEQAS Breast HER2 ICC Gold standard samples with their respective kits/assays, and for also providing the ISH UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 35

38 RUN 16 GASTRIC HER2 ICC Module Selected Images showing Optimal and Sub-optimal Immunostaining $%& '( )*+%,-. /+-%%& :; &-/+=%B +?,1?=/C D:E /-,*.4 : >.1BG 'H and (B) sample A block 2. The membrane staining is strong and completely circumferential with minimal cytoplasmic staining. Both sections stained with the Ventana 4B5 on the Benchmark platform with CC1 mild pre-treatment. Fig 2. Suboptimal HER2 staining of the UK NEQAS distributed 3+ gastric tumour (sample A block 2). The staining is weak and more representative of a 2+ (compare to Fig 1B). Section stained with the HercepTest, but unfortunately not all the methodology data was submitted by the participant. Fig 3. Two examples of unacceptable staining of the UK NEQAS distributed 3+ gastric tumour (sample A block 2). (A) Excessive counterstain making it difficult to determine the membrane staining. (B) Non-specific inappropriate staining throughout the section, and again, this makes it difficult to determine the membranes. $%& I( )*+%,-. /+-%%& :; <%/+=%>?+4< JA &-/+=%B +?,1?=/C D:E /-,*.4 K >.1BG 'H and (B) sample B block 2. The majority of cells show weak to moderate complete membrane staining. Both sections stained with the Ventana 4B5 on the Benchmark platform with CC1 mild pre-treatment. Fig 5. Suboptimal demonstration of HER2 in the UK NEQAS distributed 2+ gastric tumour (sample B block 2). The staining is weak and more representative of a 1+ (compare with Figs 4A & 4B). Section stained with the Ventana 4B5 on the Benchmark with CC1 for 32 minutes. Fig 6. Poor demonstration of HER2 in the UK NEQAS distributed 2+ gastric tumour (sample B block 2). The section shows morphology damage, most likely caused by an inappropriate pre-treatment protocol. 36

39 RUN 16 GASTRIC HER2 ICC Module Selected Images showing Optimal and Sub-optimal Immunostaining Fig 7. Optimal HER2 staining of the UK NEQAS 2+ gastric tumour (sample C block 1), showing the expected weak to moderate complete membrane staining. Please note, due to the heterogenic nature of the tissue, sample C changed staining expression through the block, and some participants would have received a section of 1+ or expression level. Fig 8. Expected level of staining of the UK NEQAS distributed 1+ gastric tumour (sample C block 2). The membrane staining is incomplete and barely perceptible. Section stained with the Ventana 4B5 on the Benchmark XT with CC1 mild retrieval. Fig 9. Optimal staining of the UK NEQAS distributed 1+ gastric tumour (sample D block 2). Similarly to sample C (Figs 7 & 8), due to the heterogenic nature of the tissue, sample D changed staining expression through the block, and some participants would have received a section of /negative expression. Fig 1. Suboptimal demonstration of HER2 in the UK NEQAS distributed 1+ gastric tumours. (A) sample C block 2, and (B) sample D block 2. Both these tumours look more representative of /negative expression. The serial sections of these samples were known to be of 1+ elmneoopqr SeTeS UVeR WQXMYNeZ [Q [Ve \]QSZ O[YRZYNZO^_ Fig 11. Good example of an in house composite gastric HER2 control. (A) 3+: Intense, crisp membrane staining and (B) 2+: Weak to moderate membrane staining. (See Fig 12 also). Fig 12. Good example of an in house composite gastric HER2 control. (A) 1+: Faint, incomplete membrane staining, and (B), the negative tumour remains unstained. 37

40 Run 16 GASTRIC HER2 ICC Module GRAPHICAL REPRESENTATION OF PASS RATES no. of returns RUN 16Gn Gastric HER2 IHC on NEQAS Sections Individual 4 = 7 (9%) 5 = (%) 6 = 3 (4%) 7 = (%) 8 = 14 (18%) 9 = (%) 1 = 4 (5%) 11 = 2 (3%) 12 = 9 (12%) 13 = 2 (3%) 14 = 8 (11%) 15 = 5 (7%) 16 = 15 (2%) 17 = 3 (4%) 18 = 3 (4%) 19 = 1 (1%) 2 = (%) no. of returns RUN 16Gh Gastric HER2 IHC on in-house Sections Individual 3 = (%) 4 = 1 (2%) 5 = (%) 6 = (%) 7 = 1 (2%) 8 = 1 (2%) 9 = (%) 1 = 2 (3%) 11 = (%) 12 = 12 (2%) 13 = 2 (3%) 14 = 4 (7%) 15 = 2 (3%) 16 = 21 (34%) 17 = 6 (1%) 18 = 4 (7%) 19 = 2 (3%) 2 = 3 (5%) Summary Summary 16-2 = 22 (29%) 16-2 = 36 (59%) = 15 (2%) = 8 (13%) 1-12 = 15 (2%) 1-12 = 14 (23%) - 9 = 24 (32%) - 9 = 3 (5%) Median = 13. Median = 14. ANTIBODIES AND OTHER TECHNICAL PARAMETERS EMPLOYED BY PARTICIPANTS IN THE GASTRIC HER2 ICC MODULE The following tables record the number of participants (N) using each primary antibody and the percentage (%) of these participants achieving acceptable staining (a score >12/2) on UK NEQAS sections. Gastric HER2 ICC Run: 16 Primary Antibody N % Biocare CME 342 A,B (EP145Y) 1 Dako A485 C-erB-2 (poly) 3 33 Dako A485 C-erB-2 (poly) 1 Dako HercepTest K524 (poly) 1 Dako Link HercepTest SK1 (poly) 2 5 Dako Link HercepTest SK1 (poly) 4 5 Labvision / Neomarkers RM-913 (SP3) 1 Neomarkers MS-73-P 1 Ventana Confirm (4B5) Gastric HER2 ICC Run: 16 Gastric HER2 IHC Automation N % BioGenex GenoMX 6i 1 Dako Autostainer Link Dako Autostainer plus 2 Dako Autostainer Plus Link 3 67 Leica Bond Max 2 5 Menarini - Intellipath FLX 1 Ventana Benchmark 3 67 Ventana Benchmark ULTRA Ventana Benchmark XT Ventana Confirm 79/ (4B5) 3 33 Ventana Pathway 79-1 (4B5) 2 5 Ventana Pathway (4B5) 2 1 Ventana Pathway (4B5)

41 Run 16 GASTRIC HER2 ICC Module Gastric HER2 ICC Run: 16 Heat Mediated Retrieval Gastric HER2 IHC N % Biocare Decloaking Chamber 1 Dako PTLink 7 43 Lab vision PT Module 1 Leica ER1 2 mins 2 5 Microwave 1 1 Other 1 Ventana CC1 16mins 1 Ventana CC1 32mins 4 5 Ventana CC1 36mins Ventana CC1 52mins 1 Ventana CC1 56mins 1 1 Ventana CC1 64mins 3 67 Ventana CC1 8mins 1 Ventana CC1 mild Ventana CC1 standard 4 25 Ventana CC2 standard 1 Gastric HER2 ICC Run: 16 Detection Gastric HER2 IHC N % AS PER KIT 8 63 Dako HerCep Test Autor (SK1) 3 1 Leica Bond Polymer Refine (DS98) 2 5 None 1 Other 1 Ventana iview system (76-91) 4 25 Ventana OptiView Kit (76-7) 2 5 Ventana UltraView Kit (76-5) Gastric HER2 ICC Run: 16 Gastric HER2 IHC Enzyme Retrieval N % AS PER KIT 2 Enzyme digestion + HIER 1 NOT APPLICABLE 16 5 Ventana Protease 1 Ventana Protease 1 (76-218) 2 5 Gastric HER2 ICC Run: 16 Gastric HER2 IHC Chromogen N % AS PER KIT DAKO DAB+ 1 1 Dako FLEX DAB 3 67 Dako REAL EnVision K57 DAB 1 Leica Bond Polymer Refine kit (DS98) 2 5 Other 1 Ventana DAB 2 5 Ventana iview 4 25 Ventana Ultraview DAB

42 Run 16 GASTRIC HER2 ICC Module BEST METHODS A selection from just a few of the best methods employed by participants Gastric HER2 IHC - Method 1 Participant scored 18/2 (UK NEQAS Slide) and 19/2 (In House slide) using this method. Primary Antibody: Ventana Pathway (4B5), 12 Mins, 37 ºC Prediluted Automation: Ventana Benchmark XT Ventana UltraView DAB Main Buffer: Ventana reaction buffer (95-3) Ventana CC1 mild Chromogen: AS PER KIT Detection: AS PER KIT Gastric HER2 IHC - Method 2 Participant scored 15/2 (UK NEQAS Slide) and 18/2 (In House slide) using this method. Primary Antibody: Dako Link HercepTest SK1 (poly), 3 Mins, 25 ºC Prediluted Automation: Dako Autostainer Plus Link Dako FLEX kit Main Buffer: Dako FLEX wash buffer Dako PTLink, Buffer: Dako Citrate, PH: 6 Chromogen: Dako FLEX DAB, 25 ºC., Time 1: 1 Mins, Time 2: 1 Mins Detection: Dako HerCep Test Autor (SK1), 3 Mins, 25 ºC Prediluted Gastric HER2 IHC - Method 3 Participant scored 16/2 (UK NEQAS Slide) and 2/2 (In House slide) using this method. Primary Antibody: Ventana Confirm (4B5), 2 Mins, 37 ºC Prediluted Automation: Ventana Benchmark ULTRA Ventana UltraView DAB Main Buffer: Ventana reaction buffer (95-3) Ventana CC1 36mins Chromogen: Ventana Ultraview DAB, Time 1: 1 Mins Detection: Ventana UltraView Kit (76-5), 8 Mins, 2 ºC Prediluted 4

43 The Lymphoma Module Run 16 David Blythe Gold Standard Second Antibody Antigens Assessed: CD5 CD1 Tissue Sections circulated: Reactive tonsil & lymph node (MCL) Reactive tonsil & lymph node (FL) Number of Registered Participants: 217 Number of Participants This Run 27 (95%) Introduction Gold Standard: CD5 CD5 is a 67-kDa transmembrane glycoprotein, which is involved in B- and T-cell receptor signal transduction (Taylor). Normal cell types which are positive for this antigen include thymocytes, the majority of T-cells and a small number of B- cells (B-1 lymphocytes). This marker is useful in a variety of diagnostic situations (Taylor; Bishop; Dabbs): B-cell Lymphomas Positive in the vast majority (>9%) of B-cell chronic lymphocytic leukaemia and B-cell small lymphocytic lymphomas Positive in the majority of mantel cell lymphomas Negative in almost all other low grade B-cell lymphomas e.g. follicular lymphoma T-cell Lymphomas Positive in most (approximately 85%) of T-cell acute lymphoblastic leukaemia and lymphoblastic lymphomas Distinguishes T-cell lymphoma (CD5 +ve) from extranodal T/ NK cell lymphoma (CD5 -ve) Thymic Carcinoma Positive in most thymic carcinomas (6-1%), distinguishing thymic carcinomas from pulmonary carcinomas which are usually CD5 ve. Features of Optimal Immunostaining: (Figs 1,3 & 5) Strong, predominantly membranous staining of the majority of the T-cells in the inter-follicular area of the tonsil Moderate staining of the scattered B-cells in the mantle zone of the tonsil Strong, predominantly membranous staining of the tumour cells of the MCL Darker staining of any normal scattered T-cells within the MCL Clean background Features of Suboptimal Immunostaining: (Figs 2,4 & 6) Weak staining of cells expected to stain, especially T-cells in the inter-follicular areas. Poor localisation of staining to cell membranes. Disruption of normal cellular detail, due to excessive or incorrect antigen retrieval. References 1. Taylor CR, Burns J. The demonstration of plasma cells and other immunoglobulin-containing cells in formalin-fixed, paraffin-embedded tissues using peroxidise-labelled antibody. J Clin Pathol 19 74; 27: Bishop PW. Immunohistochemistry vade mecum. mmunohistochemistry.info/ 3. Dabbs DJ. Immunohistology of metastatic carcinomas of unknown primary. In: Dabbs DJ (ed) Diagnostic immunohistochemistry (end ed). 26. Second Antigen: CD1 Neprilysin (CD1) is a transmembrane glycoprotein expressed in a wide variety of tissues and is particularly abundant in kidney. It is also detected in mature neutrophil granulocytes and in several epithelial cell types including liver, prostate, intestine, and breast myoepithelium, as well as some stromal ells and their tumours. It is a common acute lymphocytic leukaemia antigen, which is an important cell surface marker in the diagnosis of human acute lymphocytic leukaemia (ALL). This protein is present on leukaemic cells of pre-b phenotype, which represent 85% of cases of ALL, but is also found in a proportion of T-cell ALL s (Chu and Arber)¹, (Kaufamann et al) ². Since it is expressed by early B, pro-b and pre-b lymphocytes and by lymph node germinal centres, CD1 is of use in haematological diagnosis: It is expressed in the following diseases; angioimmunoblastic T cell lymphoma, Burkitt lymphoma, CML in blast crisis (9%), diffuse large B- cell lymphoma (variable), follicular centre cells (7%), hairy cell leukaemia (1%), and some myelomas (Jaffe et al) ³. It tends to be negative in AML, CLL, mantle cell lymphoma, and marginal zone lymphoma. Features of Optimal Immunostaining: (Figs 7,9 & 12) Moderate-strong membrane staining of all germinal centre B -lymphocytes and some dispersed in the peri-follicular region of the reactive tonsil. Intense membrane staining in most of the neoplastic B-cells of the follicular lymphoma. Clean background Features of Suboptimal Immunostaining: (Figs 8,1 & 11) Weak, uneven, partially missing staining of relevant cells. Poor / diffuse localisation of staining. Excessive background staining, or non-specific staining of cell types or components not expected to stain. Tissue destruction due to harsh heat induced antigen retrieval. References 1. Chui P, Arber DA. Paraffin-section detection of CD1 in 55 nonhematopoietic neoplasms. Frequent expression in renal cell carcinoma and endometrial stromal sarcoma. Am J Clin Pathol 2; 113: Kaufamann O, et al. Immunohistochemical detection of CD1 with monoclonal anditoby 56C6 on paraffin sections. Am J Clin Pathol 1999; 11: Jaffe ES, Harris NL, Stein H, Vardiman JW (Eds.) WHO Classification of Tumours. Pathology and Genetics of tumours of Haematopoietic and Lymphoid Tissues. Lyon, IARC Press 21. Assessment Summary: Overall the results obtained were very good, and showed an improvement on the pass rates compared to those achieved when CD5 was last assessed (Run 1). However, for many participants the marks on the UK NEQAS material was higher than that for the in-house. This was mainly down to the ability to see B-cell demonstration in the mantle zone of the UK NEQAS tonsil. In-house control material submitted by participants included tonsil (but without any germinal centres), breast carcinomas, spleen, Hodgkin's lymphoma, follicular cell lymphoma, other miscellaneous lymphoma tissue and a bone marrow trephine biopsy. Although staining was present using this range of tissue it was difficult for the assessors to accurately gauge the mantle cell demonstration. UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 41

44 The Lymphoma Module Run 16 The assessors felt that participants of the lymphoma scheme should possess more suitable control material, such as reactive tonsil. Interestingly for the CD1 assessment (slide P) more participants submitted tonsil as the in-house control than they did for the CD5. Being able to see good demonstration of the B-cells in the mantle zone enabled laboratories to score 16/2 and above. Where mantle cell staining was pale this resulted in a final mark in the 12-14/2 range, and for laboratories just demonstrating T-cells the marks fell to 8-1/2. Generally the staining of the mantle cell lymphoma tissue on the UK NEQAS slides was good. This clearly showed the far greater tumour cell expression of the antigen compared to normal mantle cells, although tumour cells and reactive T-cells were clearly distinguishable. Some participants lost marks due to poor localisation, this was generally associated with less intense antigen retrieval strategies such as a steamer, microwave oven and use of CC1 (Ventana platform) for only 16 minutes. Interestingly, excessive antigen retrieval was more of an issue with in-house slides rather than the UK NEQAS slides. Some of the top scoring laboratories submitted both tonsil and mantle cell lymphoma tissue as their in-house control slides. CD1 (Slides N and P) Generally most participants showed good demonstration on both the UK NEQAS and their in-house material. The most frequent assessor comments related to "staining could be stronger", excessive antigen retrieval and poor localisation. As with the CD5 assessment, some laboratories submitted various carcinomas as their in-house control together with kidney, placenta and multi-blocks without lymphoid tissue. This made their assessment difficult, as the assessors felt that the controls were not ideal for the lymphoma scheme. Footnote: It was pleasing to note that the use of the Surgipath Xtra slides (as an alternative to the Thermo Shandon slides) has addressed the issue that some participants were getting negative or patchy staining on the UK NEQAS slides. UK NEQAS apologies for the inconvenience caused to those participants. UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 42

45 RUN 16 LYMPHOMA Module Selected Images showing Optimal and Sub-optimal Immunostaining Fig 1. Optimal demonstration of CD5 in the UK NEQAS reactive tonsil. All the T-cells of the inter-follicular areas and scattered T-cells in the germinal centres show strong predominantly membranous staining. Stained with the Novocastra 4C7 antibody, 1:2, pre-treated in the Dako PT Link with high ph buffer. Fig 2. Suboptimal demonstration of CD5 in the UK NEQAS reactive tonsil: The reaction is weak and the number of T-cells staining is lower than expected (compare to Fig 1). This is most likely due to insufficient pre-treatment. Section stained with the Novocastra 4C7 antibody, 1:15, on the Ventana Benchmark XT with no antigen retrieval. Fig 3. Good demonstration of CD5 in the UK NEQAS distributed mantle cell lymphoma. Both the neoplastic cells and scattered normal T-cells show strong membranous staining (shown to better advantage in the high power insert). Same protocol as Fig 1. Fig 4. Poor demonstration of CD5 in the UK NEQAS mantle cell lymphoma: Staining of the tumour and normal T-cells is weak (compare to Fig 3). Same protocol as Fig 2. Fig 5. Good demonstration of CD5 on an in house reactive tonsil control. Even at low power it is clear to see the strong membrane staining of the T-cells, and also the B-cells in the mantle zone. Section stained with the Dako 4C7 pre-diluted antibody, pre-treated in the Dako PT link. Fig 6. Suboptimal demonstration of CD5 on an in house reactive tonsil section. The example shows weak and poorly localized staining. Stained with the Gennova antibody, 1:25, on the Ventana ULTRA, CC2 64 minutes pre-treatment. 43

46 RUN 16 LYMPHOMA Module Selected Images showing Optimal and Sub-optimal Immunostaining Fig 7. Optimal demonstration of CD1 in the UK NEQAS reactive tonsil. The section shows strong membrane and cytoplasmic staining of the B-cells in the lymphoid germinal centre. Section stained with the Dako 56C6 pre-diluted antibody on the Omnis with high ph buffer pre-treatment. Fig 8. Suboptimal demonstration of CD1 in the UK NEQAS reactive tonsil. The germinal centre B-cells show only very weak staining (compare to Fig 7). Section stained with the Leica 56C6 antibody, 1:3, on the Ventana Benchnark XT with CC1 standard antigen retrieval. Fig 9. Optimal demonstration of CD1 in the UK NEQAS distributed follicular lymphoma. The strong membrane and cytoplasmic staining is shown to better advantage in the high power insert. Same protocol as Fig 7. Fig 1. Poor demonstration of CD1 in the UK NEQAS distributed follicular lymphoma. The staining is very weak with many of the neoplastic cells not staining (compare to Fig 9). Section stained with the Labvision antibody, 1:3, on the Leica Bond III with ER1 antigen retrieval for 2 minutes. Fig 11. Suboptimal demonstration of CD1 in the UK NEQAS distributed follicular lymphoma. The tissue morphology has been damaged, most likely caused by excessive heat antigen retrieval. Section stained with the Ventana SP67 pre-diluted antibody on the Benchmark, no details on the antigen retrieval were provided. `ab cdf ghhi jklmopj hq lr sar thuvjw xhrvap yhrxzhp vxlarji {axt }c~f tj bjzmarlp yjppv show strong and distinct staining while the background remains clean. Section stained with the Ventana SP67 pre-diluted antibody on the Benchmark XT with CC1 standard antigen retrieval. 44

47 Run 16 LYMPHOMA Module GRAPHICAL REPRESENTATION OF PASS RATES 7 RUN 16L CD5 on NEQAS Sections Individual 4 = (%) 1 RUN 16M CD5 on in-house Sections Individual 4 = (%) = (%) 6 = (%) 7 = (%) 8 = 2 (1%) 8 5 = (%) 6 = (%) 7 = 1 (%) 8 = 7 (3%) no. of returns = 1 (%) 1 = 2 (1%) 11 = 5 (2%) 12 = 14 (7%) 13 = 11 (5%) no. of returns = (%) 1 = 1 (%) 11 = 7 (3%) 12 = 1 (5%) 13 = 12 (6%) = 15 (7%) 15 = 14 (7%) 16 = 67 (32%) 17 = 35 (17%) 2 14 = 18 (9%) 15 = 14 (7%) 16 = 9 (44%) 17 = 27 (13%) = 14 (7%) 19 = 2 (1%) 2 = 7 (3%) = 11 (5%) 19 = 6 (3%) 2 = 2 (1%) Summary Summary 16-2 = 143 (69%) 16-2 = 136 (66%) = 4 (19%) = 44 (21%) 1-12 = 21 (1%) 1-12 = 18 (9%) - 9 = 3 (1%) - 9 = 8 (4%) Median = 14. Median = 14. no. of returns RUN 16N CD1 on NEQAS Sections Individual 4 = (%) 5 = (%) 6 = (%) 7 = (%) 8 = 4 (2%) 9 = 1 (%) 1 = (%) 11 = 4 (2%) 12 = 22 (11%) 13 = 12 (6%) 14 = 13 (6%) 15 = 12 (6%) 16 = 74 (36%) 17 = 32 (15%) 18 = 21 (1%) 19 = 7 (3%) 2 = 5 (2%) no. of returns RUN 16P CD1 on in-house Sections Individual 4 = (%) 5 = (%) 6 = 1 (%) 7 = 1 (%) 8 = 3 (1%) 9 = 6 (3%) 1 = 5 (2%) 11 = 12 (6%) 12 = 22 (11%) 13 = 1 (5%) 14 = 15 (7%) 15 = 22 (11%) 16 = 75 (36%) 17 = 2 (1%) 18 = 7 (3%) 19 = 6 (3%) 2 = 1 (%) Summary Summary 16-2 = 139 (67%) 16-2 = 19 (53%) = 37 (18%) = 47 (23%) 1-12 = 26 (13%) 1-12 = 39 (19%) - 9 = 5 (2%) - 9 = 11 (5%) Median = 14.5 Median =

48 Run 16 LYMPHOMA Module ANTIBODIES AND OTHER TECHNICAL PARAMETERS EMPLOYED BY PARTICIPANTS IN THE LYMPHOMA MODULE The following tables record the number of participants (N) using each primary antibody and the percentage (%) of these participants achieving acceptable staining (a score >12/2) on UK NEQAS sections. Lymphoma Run: 16 Lymphoma Run: 16 Primary Antibody : CD5 Antibody Details N % Biogenex AM43-5M (4C7) 1 1 Biogenex MU43-UC (4C7) 1 Cell Marque 25M/S-x (4C7) 2 5 Cell Marque (SP19) 1 1 Dako IR81 (SP19) 2 1 Dako IR82 RTU FLEX Link (4C7) 1 9 Dako IS82 RTU Auto Plus (4C7) 2 1 Dako M3641 (4C7) Labvision MS-393-S (4C7) 1 1 Labvision RM-9119 (SP19) 3 33 Leica Bond RTU PA168 (4C7) NeoMarkers MS-393-S (4C7) 1 1 Novocastra Bond PA168 (4C7) 5 1 Novocastra NCL-CD5-4C7 (4C7) 41 8 Novocastra NCL-L- CD5-4C7 (4C7) 63 9 Novocastra RTU-CD5-4C7 (4C7) 1 1 Other 3 67 Vector VP-C322 (4C7) 1 1 Ventana (SP19) 7 1 Ventana CD (SP19) 26 1 Primary Antibody : CD1 Antibody Details N % BioGenex MU/AM451 (56C6) 1 1 Cell Marque 11M-16/18 (56C6) 1 1 Dako M727 (SS2/36) 4 1 Dako M738 (56C6) Dako RTU FLEX Auto Plus IS648 (56C6) 4 1 Dako RTU FLEX Link IR648 (56C6) 9 89 Labvision MS728S (56C6) 1 Leica RTU PA27 (56C6) Novocastra NCL-CD1-27 (56C6) Novocastra NCL-L-CD1-27 (56C6) Novocastra RTU-CD1-27-R (56C6) Other 7 57 Vector VP-C328 (56C6) 1 1 Ventana (56C6) 2 5 Ventana (SP67) Lymphoma Run: 16 Lymphoma Run: 16 CD1 CD5 CD1 CD5 Heat Mediated Retrieval N % N % _Leica Bond III ER2 1 1 _Leica BondMax ER _Microwave Oven 1 1 _Ventana Benk CC1 (Extended) 1 1 _Ventana Benk CC1 (Standard) 2 1 _Ventana Benk ULTRA CC1 (Exten.) _Ventana Benk ULTRA CC1 (Stan.) 1 1 _Ventana Benk XT CC1 (Mild) 1 1 _Ventana Benk XT CC1 (Standard) Biocare Decloaking Chamber Dako Omnis Dako PTLink Lab vision PT Module Leica ER1 1 mins Leica ER1 2 mins Leica ER1 3 mins 4 1 Leica ER2 1 mins 2 1 Leica ER2 2 mins Leica ER2 3 mins Microwave None Other Pressure Cooker Steamer 1 Ventana CC1 16mins 1 Ventana CC1 32mins Ventana CC1 36mins 4 1 Ventana CC1 4mins Ventana CC1 48mins Ventana CC1 52mins 2 1 Ventana CC1 56mins Ventana CC1 64mins Ventana CC1 72mins Ventana CC1 8mins 1 1 Ventana CC1 92mins 4 1 Ventana CC1 extended Ventana CC1 mild Ventana CC1 standard Ventana CC2 24mins 1 1 Ventana CC2 64mins Ventana CC2 76mins 1 1 Ventana CC2 extended 1 1 Water bath OC Enzyme Mediated Retrieval N % N % AS PER KIT NOT APPLICABLE

49 Run 16 LYMPHOMA Module Lymphoma Run: 16 Lymphoma Run: 16 CD1 CD5 CD1 CD5 Detection N % N % AS PER KIT ƒ ˆ Š Œ Ž š 1 1 ƒ ˆ Š Œ Ž œ š 1 1 Dako EnVƒžƒ Ÿ œ ( K8/1) Dako EnVƒžƒ Ÿ œ+ ( K82/12) Dako Envision HRP/DAB ( K57) LabVision UltraVision LP HRP (TS 125 HD) Leica Bond Polymer Refine (DS98) Š œ Š ž Š œ Šš Ž Vector ImmPRESS Universal (MP-75) Ventana iview system (76-91) V ƒview Kit (76-7) Ventana UltraView Kit (76-5) Chromogen N % N % AS PER KIT ƒ ˆ ƒ ƒª «š ƒ ˆ ƒ ƒª A (HK-124-7K) ƒ ƒª ±š Dako DAB+ REAL Detection (K51) 1 1 Dako EnVision Plus kits Ÿ œ Dako REAL EnVision K57 DAB LabVision DAB 1 Leica Bond Polymer Refine kit (DS98) ˆ² ƒ Š ± š Ž Sigma DAB (D5637) Sigma DAB (D595) Ventana DAB Ventana iview Ventana Ultraview DAB Vƒžƒ ƒ Œž ž ª œ Lymphoma Run: 16 CD1 CD5 Automation N % N % ƒ ˆ œ ƒ Dako Autostainer Dako Autostainer Link Dako Autostainer plus Dako Autostainer Plus Link ƒž LabVision Autostainer ƒ² ª ˆ ƒ² ª œ 1 1 Leica Bond-III Žƒ ƒ ƒ Ÿ œ None (Manual) ª ³ TechMate 5Plus 1 1 Ventana Benchmark Ventana Benchmark ULTRA V ² Ž œ BEST METHODS - Gold Standard Antibody A çèéèêëìíî ïðíñ òóçë ô ïèõ íï ëöè èçë ñèëöíøç èñùéíúèø ú ùôðëìêìùôîëç CD5 - Method 1 Participant scored 2/2 (UK NEQAS Slide) and 2/2 (In House slide) using this method. Primary Antibody: À Á Â Ã Ä ÀźÆŵÇÆÈÅÉ ÊÈÅÉË Ì ÍΠϾÐÃÌ ÑÒ ÓÅ µ¾ôõ ¾ Ð ÒÖ ÑÎÎ Automation: Dako Autostainer Plus Link µ ¹º»¼½ ¾ Main Buffer: µ ¹º»¼ ÃØ ÙÕffer Dako PTLink, BufferÖ µ Ú¾ÛØ ÜÚ ÀÝT APPLICABLE Chromogen: Detection: µ ¹º»¼ µþßì ÑÒ ÓÅàÌ Time ÒÖ Ç Ï¾ÐÃÌ Time ÑÖ Ç Ï¾Ðà Dako EnV¾Ã¾ Ð ¹º»¼½ Ê áâîîñãòñë Ì ÒÇ Ï¾ÐÃÌ ÑÒ ÓÅ ääåæ¾ôõ åæ CD5 - Method 2 Participant scored 19/2 (UK NEQAS Slide) and 18/2 (In House slide) using this method. Primary Antibody: À Á Â Ã Ä ÀÅºÆºÆ ÅµÇÆÈÅÉ ÊÈÅÉË Ì ÒÇ Ï¾Ðà µ¾ôõ ¾ Ð ÒÖ ÇÎ Automation: Leica Bond-III ºå¾Â ß ÐæÏÞû üåý¾ðå áþÿ Main Buffer: Bond Wash Buffer (AR959) Leica ER2 2 mins Chromogen: Detection: Leica Bond Polymer Refine kit (DS98), Time ÒÖ ÒΠϾÐà Leica Bond Polymer Refine (DS98), 8 Mins Prediluted 47

50 Run 16 LYMPHOMA Module CD5 - Method 3 Participant scored 19/2 (UK NEQAS Slide) and 18/2 (In House slide) using this method. Primary Antibody: Ventana CD (SP19), 32 Mins, 37 ºC Prediluted Automation: Ventana Benchmark XT Ventana Optiview Main Buffer: Ventana reaction buffer (95-3), PH: 7.6 Ventana CC1 4mins, Buffer: CC1, PH: 7.8 Chromogen: Other, Time 1: 8 Mins Detection: Ventana OptiView Kit (76-7), 8 Mins, 37 ºC Prediluted CD5 - Method 4 Participant scored 2/2 (UK NEQAS Slide) and 16/2 (In House slide) using this method. Primary Antibody: Labvision RM-9119 (SP19), 56 Mins Dilution 1: 25 Automation: Ventana Benchmark XT Ventana UltraView DAB Main Buffer: Ventana reaction buffer (95-3) Ventana CC1 standard Chromogen: Ventana Ultraview DAB, Time 1: 8 Mins, Time 2: 8 Mins Detection: Ventana UltraView Kit (76-5) Prediluted BEST METHODS - Secondary Antibody A selection from just a few of the best methods employed by participants CD1 - Method 1 Participant scored 19/2 (UK NEQAS Slide) and 18/2 (In House slide) using this method. Primary Antibody: Dako M738 (56C6), 3 Mins, 21 ºC Dilution 1: 5 Automation: Dako Autostainer Link 48 Dako FLEX+ kit Main Buffer: Dako FLEX wash buffer Dako PTLink, Buffer: FLEX TRS HIGH PH NOT APPLICABLE Chromogen: Dako FLEX DAB, 21 ºC., Time 1: 1 Mins Detection: Dako EnVision FLEX+ ( K82/12), 2 Mins, 21 ºC Prediluted CD1 - Method 2 Participant scored 18/2 (UK NEQAS Slide) and 17/2 (In House slide) using this method. Primary Antibody: Vector VP-C328 (56C6) Dilution 1: 5 Automation: Main Buffer: Dako Autostainer Dako FLEX+ kit Dako FLEX wash buffer Chromogen: Detection: Dako PTLink, Buffer: DAKO HIGH PH NOT APPLICABLE Dako FLEX DAB, 2 ºC., Time 1: 1 Mins, Time 2: 1 Mins Dako EnVision FLEX+ ( K82/12), 3 Mins, 2 ºC Prediluted 48

51 Run 16 LYMPHOMA Module CD1 - Method 3 Participant scored 18/2 (UK NEQAS Slide) and 19/2 (In House slide) using this method. Primary Antibody: Ventana (56C6) Automation: Ventana Benchmark XT Ventana UltraView DAB Main Buffer: Ventana reaction buffer (95-3) Chromogen: Ventana Ultraview DAB Detection: CD1 - Method 4 Participant scored 19/2 (UK NEQAS Slide) and 19/2 (In House slide) using this method. Primary Antibody: Novocastra NCL-L-CD1-27 (56C6), 12 Mins, 36 ºC Dilution 1: 2 Automation: Ventana Benchmark ULTRA Ventana Optiview Main Buffer: Ventana reaction buffer (95-3) Ventana CC1 64mins NOT APPLICABLE Chromogen: AS PER KIT, 36 ºC., Time 1: 8 Mins Detection: Ventana OptiView Kit (76-7), 8 Mins, 36 ºC Prediluted 49

52 Mercury FREE Microscopy The safer, cheaper solution for your lab All mercury and metal halide bulbs contain potentially dangerous levels of Mercury. On top of that they are inefficient, difficult to dispose of and expensive to run. Lumencor s light engines from Imsol are unique, solid state illuminators, ideally suited for use with fluorescence microscopes and bioanalytical instruments without the need for mercury. Performance Metrics: Mercury-based light sources vs. light engines in a 72h microscopy experiment Measured 13w metal halide Light engine On time 72 hours (shuttered) 72 seconds (shutterless) Energy used 11.5 KWh.1 KWh CO 2 23lbs.2lbs Unit lifetime used 4%.1% Experiment cost If you are interested in reducing your costs AND your carbon footprint by upgrading to Solid State Illumination, contact Image Solutions UK Ltd. for a free consultation. Imsol House, 2 Cable Court, Fulwood, Lancs, PR2 9YW Phone: Fax: Web: info@imsol.co.uk 5

53 The Neuropathology Module Run 16 Neil Bilbe Gold Standard Second Antibody Antigens Assessed: MIB-1 CK7 Tissue Sections circulated: Number of Registered Participants: 68 Number of Participants This Run 67 (99%) Introduction Meningioma and Metastatic Squamous Cell Carcinoma Gold Standard: MIB-1 Antigen Ki-67 is a nuclear protein that is associated with and may be necessary for cellular proliferation. Furthermore it is associated with ribosomal RNA transcription. Inactivation of antigen Ki-67 leads to inhibition of ribosomal RNA synthesis. Ki-67 is an excellent marker to determine the growth fraction of a given cell population. The fraction of Ki-67-positive tumour cells (the Ki-67 labelling index) is often correlated with the clinical course of cancer. The best-studied examples in this context are carcinomas of the prostate, brain, breast and nephroblastoma. For these types of tumours, the prognostic value for survival and tumour recurrence have repeatedly been proven in uni- and multivariate analysis. Ki-67 and MIB-1 monoclonal antibodies are directed against different epitopes of the same proliferation-related antigen. Ki- 67 and MIB1 may be used on fixed sections. MIB-1 is used in clinical applications to determine the Ki-67 labelling index. One of its primary advantages over the original Ki-67 antibody (and the reason why it has essentially supplanted the original antibody for clinical use) is that it can be used on formalinfixed paraffin-embedded sections, after heat-mediated antigen retrieval. Features of Optimal Immunostaining: (Figs 2, 5 & 6) Intense and well-localised nuclear staining of tumour cells Clean background No non-specific staining Adequate counterstain Features of Suboptimal Immunostaining: (Figs 1, 2 &4) Weak, uneven, or lower than expected level of staining in the tumour cells Diffuse staining High background or non-specific staining of cell types not expected to stain Excessive or very weak counterstain References 1. Hsu DW et al, Use of MIB-1 (Ki-67) immunoreactivity in differentiating grade II and grade III gliomas. J Neuropathol Exp Neurol Aug;56 (8): Ralte AM, et al, Clinicopathological features, MIB-1 labeling index and apoptotic index in recurrent astrocytic tumours. Pathol Oncol Res. 21; 7(4): S. H Torp. Proliferative activity in human glioblastomas: evaluation of different Ki-67 equivalent antibodies. Mol Pathol 1997;5: Second Antigen: CK7 Cytokeratin 7 (CK7) is a 54 kd type II simple cytokeratin which is found in most glandular and transitional epithelia but not stratified squamous epithelium. Antibodies against CK7 label a variety of cells in normal tissue including subsets of endothelial cells. CK7 can also be demonstrated in a number of different tumours including adenocarcinoma of the lung, breast, ovary (serous and endometroid), endometrium and in transitional cell carcinoma of bladder although the level of staining very much depends on the degree of differentiation of the tumour. It is generally negative in colonic adenocarcinomas, and primary squamous cell carcinomas, hepatomas and renal epithelial tumours. Staining with CK7 should be cytoplasmic. It is usually used in conjunction with cytokeratin 2. Metastatic Breast Carcinoma Features of Optimal Immunostaining: (Figs 7,8 &12) Specific staining in the tumour cells, and other epithelia Clean background with no non-specific staining Good contrast with counterstain Features of Suboptimal Immunostaining: (Figs 9,1 & 11) Weak or uneven staining of the metastatic tumour Diffuse, poorly-localised staining. Excessive background or non-specific specific staining Weak or overtly heavy counterstain References 1. Chu P, Wu E, Weiss LM. Cytokeratin 7 and cytokeratin 2 expression in epithelial neoplasms: A survey of 435 cases. Mod Pathol. 2;13: Gyure KA, Morrison AL. Cytokeratin 7 and 2 expression in choroid plexus tumors: Utility in differentiating these neoplasms from metastatic carcinomas. Mod Pathol. 2;13: Pekmezci M, Perry A. Neuropathology of brain metastases. Surg Neurol Int 213;4, Suppl S4: Assessment Summary: A total of 264 slides were assessed, with only one laboratory not submitting at all. Two out of the 67 laboratories did not return any in-house controls. All laboratories stained for MIB-1, but three used an alternative antibody for CK7: Pan CK, AE1/AE3 or CAM 5.2. Three slides were assessed as failed; all were on NEQAS sections, one on the MIB-1 (G), and two on the CK7 (J). All three were for very weak or non demonstration of tumour cells. The main problem with the staining on the NEQAS Gold sections (G) was the level of nuclear staining in the meningioma tissue; whereas most participants achieved some diagnostically acceptable level of staining (apart from the failed slide) in the metastatic squamous cell carcinoma. Interestingly, the lab who failed on the NEQAS MIB-1 (G) did not submit any details of their methodology, and the two participants who failed the NEQAS CK7 (J) both used little used or obscure antibodies (Gennova OV-TL 1313, and PROGEN Ks7.18). With the 2nd antigen NEQAS slides (CK7), which consisted of a case of metastatic breast disease, some patchy staining was seen within one area of the tumour. Comparison with the validation slides suggested that this may be tissue related, rather than due to any technical issues. Participants were not marked down if this was observed, and similar to the Gold standard at the same serial level. This contributed to 7/65 labs (1.5%) being assessed as borderline (1-12) for CK7 (J), compared to 5/67 (7.5%) of those for MIB-1 (G). The in-house slides were relatively straightforward. No slides failed, and only three slides were scored as borderline (1-12) all of these were on the MIB-1 stained slides (H). Thus, all in-house control slides for the CK7 (K) scored 13. All in all, a fairly trouble-free assessment. The level of laboratories using a CNS or some form of neurological sample for in-house slides is still relatively low: 11 out of 65 labs -17% for the MIB-1 in-house controls (H) 9 out of 65 labs - 14% for the CK7 in-house controls (K) 51 UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent

54 RUN 16 NEUROPATHOLOGY Module Selected Images showing Optimal and Sub-optimal Immunostaining Fig 1. Sub-optimal staining on the NEQAS metastatic tumour. There is noticeable background staining, otherwise the nuclear staining is crisp and selective. Monoclonal Dako Ki67 antibody, 1/5 dilution, stained on the Leica Bond III, with ER2 for 2 mins and the Leica Refine kit. Fig 2. Optimal staining of the NEQAS slide on the metastatic squamous carcinoma tissue. The slide is clean, and the tumour nuclei are clearly demonstrated with excellent differential expression of the antigen. Dako monoclonal MIB1, on Ventana Benchmark ULTRA and CC1 standard. Fig 3. Sub-optimal demonstration of the metastatic component. Weak and incomplete staining, with little or no differential staining of nuclei. Assessed as a borderline pass. Stained with the D F! " # $ # % &% ' % missing completely. Slide failed the NEQAS assessment. The participant did not submit any details of the antibody employed, or methodology used, which prevents us providing any suitable feedback. Fig 5. Slide showing excellent demonstration of nuclei in the NEQAS meningioma sample. Staining is precise and clean. Obtaining this level of expression proved problematic for many labs. Ventana RTU Ki67 for 16 mins, on the Ventana Benchmark ULTRA, and Optiview kit for 8 mins. Fig 6. Excellent demonstration of a metastatic tumour in an in-house control slide. The non-tumour areas are clean with no discernable non-specific staining. Ventana RTU CONFIRM for 16 mins, on the Ventana Benchmark ULTRA, with CC1 for 52 mins, and the Ultraview kit 2nd layer for 8 mins. 52

55 RUN 16 NEUROPATHOLOGY Module Selected Images showing Optimal and Sub-optimal Immunostaining Fig 7. Good demonstration of CK7 antigen in the NEQAS breast metastatic tissue. The area of patchy staining (top left-hand portion) was a feature of the sample, and not due to incomplete staining. Novacastra monoclonal at 1/1, 15 mins at RT, on Leica Bond with ER1 for 3 mins. Fig 8. Optimal demonstration of CK7 in the NEQAS tissue. There is crisp staining of tumour, to a level comparable to the NEQAS Gold standard slide(s). Dako OV-TL monoclonal antibody at 1/2, on the Leica Bond-III, with ER2 for 2 mins, and the Leica Bond polymer Refine kit. Fig 9. Sub-optimal demonstration of the NEQAS breast metastasis. Staining is weak, and in some areas uneven (compare with fig 8). Slide assessed as a borderline pass. Dako OV-TL, 1 in 2, on the Ventana Benchmark ULTRA, with hot CC1 for 36 mins, and the Ultraview kit for 12 mins at RT. Fig 1. Weak staining throughout the NEQAS tumour sample (low power). Slide was assessed as just about acceptable for diagnostic purposes, and given a borderline score. Primary antibody was too dilute. Dako OV-TL at 1 in 2, on Bond Max, using ER2 for 2 mins, and a pre-diluted Refine kit. Fig 11. Poor demonstration of CK7 in NEQAS slide. Staining is too weak for reliable diagnosis, (+3 8,-1 /9-,9) c1558 ()1 c,-251/15a 9+8/(*+13L /*8 85*31.(*513 /1 ( /6 M1++,N( OV-TL antibody at 1/2, Ventana Benchmark XT, CC1 standard, and the pre-diluted Ultra View kit. Fig 12. Excellent example of CK7 demonstration on an in-house control slide, from a case of c()c*+,-(,. /1 2(),/*3 45( /*8891 :(8.*;13 *+ <=>?@A.,) BC,9)86 E(G, HIJ7K diluted 1/3 for 32 mins on the Ventana Benchmark XT, with CC1 mild. 53

56 Run 16 NEUROPATHOLOGY Module GRAPHICAL REPRESENTATION OF PASS RATES 14 RUN 16G Ki-67 / MIB1 on NEQAS Sections Individual 4 = (%) 14 RUN 16H Ki-67 / MIB1 on in-house Sections Individual 4 = (%) 12 5 = (%) 6 = (%) 12 5 = (%) 6 = (%) 1 7 = (%) 8 = (%) 1 7 = (%) 8 = (%) no. of returns = 1 (1%) 1 = 3 (4%) 11 = (%) 12 = 2 (3%) 13 = 3 (4%) no. of returns = (%) 1 = 1 (2%) 11 = 1 (2%) 12 = 1 (2%) 13 = 1 (2%) 4 14 = 4 (6%) 15 = 7 (1%) 4 14 = 5 (8%) 15 = 5 (8%) 2 16 = 13 (19%) 17 = 9 (13%) 2 16 = 12 (18%) 17 = 11 (17%) = 3 (4%) 19 = 8 (12%) 2 = 14 (21%) = 9 (14%) 19 = 13 (2%) 2 = 6 (9%) Summary Summary 16-2 = 47 (7%) 16-2 = 51 (78%) = 14 (21%) = 11 (17%) 1-12 = 5 (7%) 1-12 = 3 (5%) - 9 = 1 (1%) - 9 = (%) Median = 15. Median = 15. no. of returns RUN 16J CK7 on NEQAS Sections Individual 4 = (%) 5 = (%) 6 = (%) 7 = (%) 8 = 2 (3%) 9 = (%) 1 = (%) 11 = 1 (1%) 12 = 6 (9%) 13 = 2 (3%) 14 = 3 (4%) 15 = 1 (15%) 16 = 17 (25%) 17 = 9 (13%) 18 = 2 (3%) 19 = 5 (7%) 2 = 1 (15%) no. of returns RUN 16K CK7 on in-house Sections Individual 4 = (%) 5 = (%) 6 = (%) 7 = (%) 8 = (%) 9 = (%) 1 = (%) 11 = (%) 12 = (%) 13 = (%) 14 = 2 (3%) 15 = 5 (8%) 16 = 17 (26%) 17 = 9 (14%) 18 = 11 (17%) 19 = 8 (12%) 2 = 13 (2%) Summary Summary 16-2 = 43 (64%) 16-2 = 58 (89%) = 15 (22%) = 7 (11%) 1-12 = 7 (1%) 1-12 = (%) - 9 = 2 (3%) - 9 = (%) Median = 15. Median =

57 Run 16 NEUROPATHOLOGY Module ANTIBODIES AND OTHER TECHNICAL PARAMETERS EMPLOYED BY PARTICIPANTS IN THE NEUROPATHOLOGY MODULE The following tables record the number of participants (N) using each primary antibody and the percentage (%) of these participants achieving acceptable staining (a score >12/2) on UK NEQAS sections. Neuropathology Run: 16 Neuropathology Run: 16 Primary Antibody : Ki-67 / MIB1 Antibody Details N % Dako FLEX RTU (MIB1) IR Dako M7187 (Ki-67 ) 2 1 Dako M724 (MIB1) 4 95 Dako N1574 (Ki67) 1 1 Leica/Novocastra (MM1) NCL-Ki67-CE 4 1 Leica/Novocastra RTU (K2) PA Leica/Novocastra RTU (MM1) PA NeoMarkers/Thermo Sci (SP6) RM Other 3 67 Ventana RTU (3-9) Primary Antibody : CK7 Antibody Details N % Biocare CM 61 A,B, C (OV-TL 12/3) 1 Cell Marque 37M-96 (OV-TL 12/3) 1 Dako M718 (OV-TL 12/3) Dako RTU Link IR619 (OV-TL 12/3) 2 1 Linaris ( OV-TL 12/3) 2 1 Novcastra NCL-L-CK-56 (RN7) 6 1 Novocastra Bond RTU PA942 (RN7) 2 1 Novocastra NCL-CK7-OVTL (OV-TL 12/3) 1 1 Other 8 75 Vector Labs VPC43 (OV-TL 12/3) 1 Ventana (SP52) 7 1 Ventana, (OV-TL 12/3) 1 1 Neuropathology Run: 16 Neuropathology Run: 16 CK7 Ki-67 / MIB1 CK7 Ki-67 / MIB1 Heat Mediated Retrieval N % N % _Ventana Benk ULTRA CC1 (Stan.) 1 1 Biocare Decloaking Chamber Dako Omnis Dako PTLink Leica ER1 2 mins 3 67 Leica ER1 3 mins Leica ER2 1 mins 1 1 Leica ER2 2 mins Leica ER2 3 mins Microwave None 1 1 Other 1 1 Ventana CC1 16mins 1 1 Ventana CC1 32mins Ventana CC1 36mins Ventana CC1 4mins 1 1 Ventana CC1 52mins Ventana CC1 56mins 1 1 Ventana CC1 64mins Ventana CC1 8mins Ventana CC1 mild Ventana CC1 standard Ventana CC2 64mins 1 1 Ventana CC2 mild 1 1 Water bath OC 1 1 Enzyme Mediated Retrieval N % N % Menarini protease XXlV (P838) 1 1 NOT APPLICABLE VBS Bond Enzyme Ventana Protease 1 Ventana Protease 1 (76-218)

58 Run 16 NEUROPATHOLOGY Module Neuropathology Run: 16 Neuropathology Run: 16 CK7 Ki-67 / MIB1 CK7 Ki-67 / MIB1 Detection N % N % AS PER KIT Biocare polymer (M4U534) Dako EnVision FLEX ( K8/1) Dako EnVision FLEX+ ( K82/12) Dako Envision HRP/DAB ( K57) Leica Bond Polymer Refine (DS98) None 1 1 NOT APPLICABLE 1 1 Other Ventana iview system (76-91) Ventana OptiView Kit (76-7) Ventana UltraView Kit (76-5) Chromogen N % N % AS PER KIT Dako DAB+ REAL Detection (K51) 1 1 Dako EnVision Plus kits Dako FLEX DAB Dako REAL EnVision K57 DAB Leica Bond Polymer Refine kit (DS98) Other Ventana DAB Ventana iview Ventana Ultraview DAB Neuropathology Run: 16 CK7 Ki-67 / MIB1 Automation N % N % Dako Autostainer Link Dako Autostainer Plus Link Dako Omnis Leica Bond Max Leica Bond-III Menarini - Intellipath FLX None (Manual) Ventana Benchmark Ventana Benchmark ULTRA Ventana Benchmark XT BEST METHODS - Gold Standard Antibody A selection from just a few of the best methods employed by participants Ki-67 / MIB1 - Method 1 Participant scored 2/2 (UK NEQAS Slide) and 19/2 (In House slide) using this method. Primary Antibody: Leica/Novocastra (MM1) NCL-Ki67-CE Dilution 1: 1 Automation: Leica Bond-III Leica BondMAx Refine KIT Main Buffer: Chromogen: Detection: Bond Wash Buffer (AR959) Leica ER2 3 mins Leica Bond Polymer Refine kit (DS98) AS PER KIT Ki-67 / MIB1 - Method 2 Participant scored 2/2 (UK NEQAS Slide) and 19/2 (In House slide) using this method. Primary Antibody: Dako M724 (MIB1), 32 Mins, 37 ºC Dilution 1: 1 Automation: Ventana Benchmark ULTRA Ventana UltraView DAB Main Buffer: Ventana reaction buffer (95-3) Ventana CC1 64mins NOT APPLICABLE Chromogen: Ventana Ultraview DAB, 37 ºC., Time 1: 12 Mins, Time 2: 12 Mins Detection: Ventana UltraView Kit (76-5), 12 Mins, 37 ºC Prediluted 56

59 Run 16 NEUROPATHOLOGY Module Ki-67 / MIB1 - Method 3 Participant scored 2/2 (UK NEQAS Slide) and 16/2 (In House slide) using this method. Primary Antibody: NeoMarkers/Thermo Sci (SP6) RM 916, 24 Mins, 37 ºC Dilution 1: 3 Automation: Ventana Benchmark XT Ventana UltraView DAB Main Buffer: Ventana reaction buffer (95-3) Ventana CC1 64mins, Buffer: CC1 Chromogen: Ventana Ultraview DAB, 37 ºC., Time 1: 8 Mins Detection: Ventana UltraView Kit (76-5), 8 Mins, 37 ºC Prediluted Ki-67 / MIB1 - Method 4 Participant scored 2/2 (UK NEQAS Slide) and 2/2 (In House slide) using this method. Primary Antibody: Ventana RTU (3-9) , 16 Mins, 36 ºC Prediluted Automation: Ventana Benchmark ULTRA Ventana Optiview Main Buffer: Ventana reaction buffer (95-3) Ventana CC1 56mins NOT APPLICABLE Chromogen: AS PER KIT, 36 ºC., Time 1: 8 Mins, Time 2: 8 Mins Detection: Ventana OptiView Kit (76-7), 8 Mins, 36 ºC Prediluted BEST METHODS - Secondary Antibody A selection from just a few of the best methods employed by participants CK7 - Method 1 Participant scored 16/2 (UK NEQAS Slide) and 2/2 (In House slide) using this method. Primary Antibody: Novcastra NCL-L-CK-56 (RN7), 15 Mins, 25 ºC Dilution 1: 5 Automation: Leica Bond-III Leica BondMAx Refine KIT Main Buffer: Bond Wash Buffer (AR959) Leica ER2 2 mins Chromogen: Leica Bond Polymer Refine kit (DS98) Detection: Leica Bond Polymer Refine (DS98), 8 Mins, 25 ºC CK7 - Method 2 Participant scored 2/2 (UK NEQAS Slide) and 19/2 (In House slide) using this method. Primary Antibody: Dako M718 (OV-TL 12/3), 24 Mins, 37 ºC Dilution 1: 4 Automation: Ventana Benchmark XT Ventana UltraView DAB Main Buffer: Ventana reaction buffer (95-3) None Ventana Protease 1 (76-218), 37 ºC. Digestion Time NEQAS: 12 Mins. In-House: 12 Mins Chromogen: Ventana Ultraview DAB Detection: Ventana UltraView Kit (76-5) 57

60 Run 16 NEUROPATHOLOGY Module CK7 - Method 3 Participant scored 2/2 (UK NEQAS Slide) and 16/2 (In House slide) using this method. Primary Antibody: Linaris ( OV-TL 12/3), 15 Mins, RT ºC Dilution 1: 5 Automation: Leica Bond-III Leica BondMAx Refine KIT Main Buffer: Bond Wash Buffer (AR959) VBS Bond Enzyme 1, RT ºC. Digestion Time NEQAS: 1 Mins. In-House: 1 Mins Chromogen: AS PER KIT, Time 1: 1 Mins, Time 2: 1 Mins Detection: Leica Bond Polymer Refine (DS98), 8 Mins, RT ºC CK7 - Method 4 Participant scored 2/2 (UK NEQAS Slide) and 2/2 (In House slide) using this method. Primary Antibody: Ventana (SP52), 24 Mins, 37 ºC Prediluted Automation: Ventana Benchmark ULTRA Ventana Optiview Main Buffer: Ventana reaction buffer (95-3) Ventana CC1 32mins NOT APPLICABLE Chromogen: AS PER KIT, 36 ºC., Time 1: 8 Mins, Time 2: 8 Mins Detection: Ventana OptiView Kit (76-7), 8 Mins, 36 ºC Prediluted 58

61 The Cytology Module Run 16 Neil Bilbe Gold Standard Second Antibody Antigens Assessed: CD45 Ki67 Tissue Sections circulated: Lymph Node FNA and Effusion Effusion with Carcinosis Number of Registered Participants: 75 Number of Participants This Run 73 (97%) Introduction Gold Standard: CD45 Protein tyrosine phosphatase, receptor type, C also known as PTPRC is an enzyme that, in humans, is encoded by the PTPRC gene. PTPRC is also known as CD45 antigen, originally called leukocyte common antigen (LCA). It is a type I transmembrane protein that is in various forms present on all differentiated hematopoietic cells except erythrocytes and plasma cells that assists in the activation of those cells (a form of co-stimulation). It is expressed in lymphomas, B-cell chronic lymphocytic leukaemia, hairy cell leukaemia, and acute non-lymphocytic leukaemia. A monoclonal antibody to CD45 is used in routine immunohistochemistry to differentiate between lymphomas and carcinomas. Features of Optimal Immunostaining: (Fig 6) Strong, cell membrane staining of lymphocytes Clean background No non-specific staining of other cell types not expected to stain Adequate nuclear counterstain Features of Suboptimal Immunostaining: (Figs 1-5) Weak, diffuse or partial membrane staining of lymphocytes Uneven staining Excessive background staining Non-specific staining of cell types or components not expected to stain Inadequate nuclear counterstain References 1. Leong A, et al. Manual of Diagnostic Cytology (2nd ed.) Greenwich Medical Media Ltd. pp Prasad RR, et al. Fine-needle aspiration cytology in the diagnosis of superficial lymphadenopathy: an analysis of 2,418 cases. Diagn Cytopathol. 1996; 15: Hehn ST, et al. Utility of fine-needle aspiration as a diagnostic technique in lymphoma. J Clin Oncol. 24; 22: Gong JZ, et al. Diagnostic impact of core-needle biopsy on fine-needle aspiration of non-hodgkin lymphoma. Diagn Cytopathol. 24; 31: Levien PH, et al. Role of fine-needle aspiration cytology in breast lymphoma. Diagn Cytopathol. 24; 3: Second Antigen: Ki67 Antigen Ki-67 is a nuclear protein that is associated with and may be necessary for cellular proliferation. Furthermore it is associated with ribosomal RNA transcription. Inactivation of antigen Ki-67 leads to inhibition of ribosomal RNA synthesis. Ki-67 is an excellent marker to determine the growth fraction of a given cell population. The fraction of Ki-67-positive tumour cells (the Ki-67 labelling index) is often correlated with the clinical course of cancer. The best-studied examples in this context are carcinomas of the prostate, brain and the breast and nephroblastoma. For these types of tumours, the prognostic value for survival and tumour recurrence have repeatedly been proven in uni- and multivariate analysis. Ki-67 and MIB-1 monoclonal antibodies are directed against different epitopes of the same proliferation related antigen. Ki- 67 and MIB-1 may be used on fixed sections. MIB-1 is used in clinical applications to determine the Ki-67 labelling index. References 1. P A Hall, et al. The prognostic value of Ki67 immunostaining in non- Hodgkin s lymphoma. J Pathol 1988; 154: D C Brown, et al. Proliferation in non-hodgkin's lymphoma: a comparison of Ki67 staining on fine needle aspiration and cryostat sections. J Clin Pathol 199;43: Features of Optimal Immunostaining: (Figs 9,1 & 12) Intense and well-localised nuclear staining of tumour cells Clean background No non-specific staining Adequate counter-stain Features of Suboptimal Immunostaining: (Figs 7,8 & 11) Weak, uneven, or lower than expected level of staining in the tumour cells Diffuse staining High background or non-specific staining of cell types not expected to stain Excessive or very weak counterstain Assessment Summary: A total of 285 slides were assessed. Two labs did not submit, and three labs did not return any in-house controls (slides S and U). 227 (79.6%) slides passed, averaging 16.4/2. 8 slides (2.8%) were assessed as inadequate for diagnostic purposes (fail). 5 of these were on NEQAS slides (CD45 =3, Ki67 =2) and 3 on in-house (CD45 =2, Ki67 =1). There was no pattern to the reason for the failed slides, with both weak, or no staining, and everything stained comments given. Borderline rates varied between 1% (U) to 27.4% (R) but with a slightly higher ratio of NEQAS slides to in-house (1.7:1); this in part may be due to the fact that 41.4% (29/7) of laboratories did not submit a cytological preparation, instead choosing a FFPE section for their in-house control: In-house control material used for the cytology module: UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 59

62 RUN 16 CYTOLOGY Module Selected Images showing Optimal and Sub-optimal Immunostaining Fig 1. Sub-optimal staining of the NEQAS cytospin. The lymphocytes are demonstrated, although staining could be crisper, and there is some non-specific staining of tumour cells. This may be due an initial, but inadvertent use of xylene. Dako CD45, at 1:5, on the Ventana Benchmark ULTRA. Fig 2. Sub-optimal but acceptable staining of the NEQAS slide. Lymphocytes are nicely demonstrated, the background is clean, but the counterstain could be better, to aid contrast. OPQR SOTUV WXWYYV RZ [\Z]PZP ^\Z_`bPdQ efv gh]` P id\jkhlm]\k nl]dpoh\g Qh] prd q bhzr P] stu C. Fig 3. Borderline outcome on the NEQAS cytospin. Although lymphocytes are stained, the rest of the cell population (tumour, RBCs, macrophages) have also stained to a lesser or greater degree. Novacastra CD45, 1:2 at RT, on the Leica Bond Max, with ER1 for 3 mins, and RTU Refine kit. Fig 4. Borderline outcome on the NEQAS cytospin. Slide is suitable for diagnosis, but there appears to be pronounced nuclear staining, despite the lymphocytes being demonstrated Pk\vmP]\lwx OPQR SOTU P] WXWqYYV RZ y\h_p ^RZk zp{v prd sy bhzr P] YuSV ZR d\]dh\oplv }m] Refine kit for 15 mins. Fig 5. Poor differential staining of lymphocytes on the NEQAS cytospin. The level of non-specific staining on this slide is unacceptable, and unreliable for diagnostic use. Dako CD45 (no dilution given) on Dako Autostainer Link 48 for 3 mins, and the Dako Envision detection kit for 3 mins. Fig 6. Optimal staining on an in-house cytospin, alcohol fixed, pleural effusion. Staining is crisp, the background is clean, and the counterstain intensity is good. Gentaur LCA/CD45, 1/2, 32 mins, on Ventana Benchmark, with no retrieval, but with a pre-diluted Ventana APAAP detection kit. Ki67. 6

63 RUN 16 CYTOLOGY Module Selected Images showing Optimal and Sub-optimal Immunostaining Fig 7. Weak demonstration of tumour nuclei in the NEQAS cytospin. RBCs stained non-specifically. Slide assessed as a borderline pass. Participant queried if this is due to coating of slides and applied extra alcohol at 2nd attempt. Novacastra Ki67, 1:2, on Bond-III, ER1 3 mins, and RTU Refine kit. Fig 8. Poor demonstration of Ki67 on NEQAS slide. Besides the RBCs, all other cells have pronounced cytoplasmic staining, making this unsuitable for diagnostic use. Slide was failed by the assessors. Dako Ki67, 1:5, on the Leica Bond-III, with ER2 for 2 mins using the Refine kit. Fig 9. Good differential demonstration of nuclei in the NEQAS cytospin. Staining is crisp and highly selective with a clean background. Novacastra RTU (K2) Ki67, on the Leica Bond-III, no retrieval, with the Leica Bond Polymer Refine kit. Fig 1. Good demonstration of nuclei in the NEQAS cytospin, however there is marked tissue damage and nuclear vacuolation, resulting in an overall borderline score. Dako MIB1, 1:5, 32 mins on the Ventana Benchmark XT, Ventana CC1 std at ph 7.8 for 6 mins and a RTU Ultraview kit. Fig 11. Weak nuclear and non-specific staining of RBCs, considered just sufficient for diagnostic purposes and awarded a low borderline score. Dako MIB1, 1:1 for 3 mins, using a MW pre-treatment ph6 for 15 mins on a Dako Autostainer Link 48 and a Dako RTU Envision layer. Fig 12. Excellent staining on an in-house control slide, sample is a cytospin from a case of ~ ƒƒ ˆ Š ˆŒ Œ Ž ƒ ƒšœ š ˆ Œƒ œ ž ŸŒ ~ š ƒš no retrieval for 32 mins, on a Ventana Benchmark, and the Ventana iview Kit. 61

64 Run 16 CYTOLOGY Module GRAPHICAL REPRESENTATION OF PASS RATES 2 RUN 16R CD45 / LCA on NEQAS Sections Individual 4 = 1 (1%) 5 = (%) 24 RUN 16S CD45 / LCA on in-house Sections Individual 4 = (%) 5 = (%) 16 6 = (%) 7 = (%) 8 = 1 (1%) 2 6 = (%) 7 = (%) 8 = (%) no. of returns = 1 (1%) 1 = (%) 11 = (%) 12 = 2 (27%) 13 = 1 (14%) 14 = 5 (7%) 15 = 8 (11%) no. of returns = 2 (3%) 1 = 1 (1%) 11 = 1 (1%) 12 = 9 (13%) 13 = 4 (6%) 14 = 1 (1%) 15 = 6 (9%) 4 16 = 14 (19%) 17 = 4 (5%) 4 16 = 22 (31%) 17 = 5 (7%) 18 = 3 (4%) 18 = 6 (9%) = 3 (4%) 2 = 3 (4%) = 6 (9%) 2 = 7 (1%) Summary Summary 16-2 = 27 (37%) 16-2 = 46 (66%) = 23 (32%) = 11 (16%) 1-12 = 2 (27%) 1-12 = 11 (16%) - 9 = 3 (4%) - 9 = 2 (3%) Median = 14.5 Median = RUN 16T Ki67 on NEQAS Sections Individual 4 = (%) 24 RUN 16U Ki67 on in-house Sections Individual 4 = (%) 5 = (%) 5 = (%) 16 6 = (%) 7 = (%) 2 6 = (%) 7 = (%) 8 = 2 (3%) 8 = 1 (1%) no. of returns = (%) 1 = 2 (3%) 11 = 2 (3%) 12 = 8 (11%) 13 = 9 (13%) 14 = 4 (6%) 15 = 4 (6%) no. of returns = (%) 1 = 1 (1%) 11 = (%) 12 = 6 (9%) 13 = 3 (4%) 14 = 3 (4%) 15 = 1 (1%) 4 16 = 18 (25%) 17 = 8 (11%) 4 16 = 22 (31%) 17 = 16 (23%) 18 = 2 (3%) 18 = 6 (9%) = 7 (1%) 2 = 6 (8%) = 3 (4%) 2 = 8 (11%) Summary Summary 16-2 = 41 (57%) 16-2 = 55 (79%) = 17 (24%) = 7 (1%) 1-12 = 12 (17%) 1-12 = 7 (1%) - 9 = 2 (3%) - 9 = 1 (1%) Median = 14.5 Median =

65 Run 16 CYTOLOGY Module ANTIBODIES AND OTHER TECHNICAL PARAMETERS EMPLOYED BY PARTICIPANTS IN THE CYTOLOGY MODULE The following tables record the number of participants (N) using each primary antibody and the percentage (%) of these participants achieving acceptable staining (a score >12/2) on UK NEQAS sections. Cytology Run: 16 Cytology Run: 16 Primary Antibody : CD45 / LCA Antibody Details N % Dako M742 (clone UCHL1) CD45RO 1 Dako M71 (clones 2B11+PD7/26) Dako M754 (clone 4KB5) CD45RA 1 1 Dako M833 (clone PD7/26) CD45RB 1 1 Dako RTU FLEX LINK IR751 (2B11 + PD7/26) 3 1 Leica/Novocastra Bond RTU PA42 (X16/99) 2 5 Leica/Novocastra NCL-L-LCA (X16/99) 2 5 Leica/Novocastra RTU-LCA-RP (RP2/18 + 2/22) 1 Other 2 5 Ventana CONFIRM (RP2/18) 6 1 Primary Antibody : Ki67 Antibody Details N % Dako 724 (MIB-1) Dako A47 (Polyclonal) 1 1 Dako FLEX RTU IR626 (MIB-1) 3 1 Dako M7187 (Ki-67) 5 6 Leica/Novocastra (MM1) NCL-Ki67-CE 7 86 Leica/Novocastra RTU (K2) PA Leica/Novocastra RTU (MM1) PA Neomarkers/Thermo Sci (SP6) RM Ventana RTU (3-9) Cytology Run: 16 Cytology Run: 16 Primary Antibody : CD45 / LCA Primary Antibody : Ki67 Antigen Retrieval N % Antigen Retrieval N % YES YES 2 27 NO NO Breakdown of participants reporting YES N Breakdown of participants reporting YES N Heat Mediated Heat Mediated Enzyme Enzyme Both 14 Both 2 Not Specified Not Specified Cytology Run: 16 Cytology Run: 16 Heat Mediated Retrieval Heat Mediated Retrieval Cytology Run: 16 Cytology Run: 16 Enzyme Mediated Retrieval Enzyme Mediated Retrieval 63

66 Run 16 CYTOLOGY Module Cytology Run: 16 Cytology Run: 16 Detection CD45 / LCA Ki67 N % N % AS PER KIT BioGenex SS Polymer (QD 43-XAKE) Dako EnVision FLEX ( K8/1) Dako EnVision FLEX+ ( K82/12) Dako Envision HRP/DAB ( K57) 1 Dako Envision+ HRP mouse K44/5/6/ Dako REAL ( K55) 1 LabVision UltraVision LP HRP (TS 125 HD) Leica Bond Intense R Detection (DS9263) 1 1 Leica Bond Polymer Refine (DS98) None 1 1 NOT APPLICABLE Other Power Vision DPVB999 HRP Ventana iview system (76-91) Ventana OptiView Kit (76-7) Ventana UltraView Kit (76-5) Chromogen CD45 / LCA Ki67 N % N % AS PER KIT BioGenex DAB (QD43) DAKO DAB+ 2 1 Dako EnVision Plus kits 1 1 Dako FLEX DAB Dako REAL EnVision K57 DAB Dako REAL K55 Alkaline phosphatase 1 1 Leica Bond Polymer Refine kit (DS98) Other Sigma DAB (D4168) Sigma DAB (D5637) Ventana DAB Ventana Enhanced Alk. Phos. Red Detection Kit Ventana iview Ventana Ultraview DAB Cytology Run: 16 Automation CD45 / LCA Ki67 N % N % BioGenex Optimax Dako Autostainer Dako Autostainer Link Dako Autostainer plus 1 1 Dako Autostainer Plus Link LabVision Autostainer Leica Bond Max Leica Bond X 1 1 Leica Bond-III None (Manual) Ventana Benchmark Ventana Benchmark ULTRA Ventana Benchmark XT

67 Run 16 CYTOLOGY Module BEST METHODS - Gold Standard Antibody A selection from just a few of the best methods employed by participants CD45 / LCA - Method 1 Participant scored 16/2 (UK NEQAS Slide) and 17/2 (In House slide) using this method. Primary Antibody: Ventana CONFIRM (RP2/18), 32 Mins, 37 ºC Prediluted Automation: Ventana Benchmark XT Ventana Enhanced Alk. Phos. Red Detection Kit Main Buffer: Ventana reaction buffer (95-3), PH: 7.6 Ventana CC1 mild, Buffer: CC1, PH: 7.8 Chromogen: Ventana Enhanced Alk. Phos. Red Detection Kit, Time 1: 8 Mins Detection: Other, 8 Mins, 37 ºC Prediluted CD45 / LCA - Method 2 Participant scored 18/2 (UK NEQAS Slide) and 16/2 (In House slide) using this method. Primary Antibody: Ventana CONFIRM (RP2/18), 37 ºC Automation: Ventana Benchmark ULTRA AS PER KIT Main Buffer: Ventana reaction buffer (95-3) None NOT APPLICABLE Chromogen: Ventana DAB Detection: Ventana OptiView Kit (76-7), 37 ºC CD45 / LCA - Method 3 Participant scored 17/2 (UK NEQAS Slide) and 19/2 (In House slide) using this method. Primary Antibody: Ventana CONFIRM (RP2/18), 28 Mins, 37 ºC Prediluted Automation: Ventana Benchmark ULTRA Ventana UltraView DAB Main Buffer: Ventana reaction buffer (95-3) NOT APPLICABLE Chromogen: Ventana Ultraview DAB, 37 ºC., Time 1: 8 Mins Detection: Ventana UltraView Kit (76-5), 8 Mins, 37 ºC Prediluted CD45 / LCA - Method 4 Participant scored 2/2 (UK NEQAS Slide) and 19/2 (In House slide) using this method. Primary Antibody: Dako M71 (clones 2B11+PD7/26), 3 Mins, 38 ºC Dilution 1: 1 Automation: Ventana Benchmark XT Ventana iview Kit Main Buffer: Ventana reaction buffer (95-3) Chromogen: Ventana iview Detection: Ventana iview system (76-91) 65

68 Run 16 CYTOLOGY Module BEST METHODS - Secondary Antibody A selection from just a few of the best methods employed by participants Ki67 - Method 1 Participant scored 2/2 (UK NEQAS Slide) and 2/2 (In House slide) using this method. Primary Antibody: Dako M7187 (Ki-67), 37 ºC Dilution 1: 2 Automation: Ventana Benchmark XT Ventana UltraView DAB Main Buffer: Ventana reaction buffer (95-3) Ventana CC1 standard, Buffer: CC1, PH: 8 Chromogen: Ventana Ultraview DAB, 37 ºC., Time 1: 8 Mins Detection: Ventana UltraView Kit (76-5), 8 Mins, 37 ºC Prediluted Ki67 - Method 2 Participant scored 2/2 (UK NEQAS Slide) and 17/2 (In House slide) using this method. Primary Antibody: Leica/Novocastra (MM1) NCL-Ki67-CE, 28 Mins, 36 ºC Dilution 1: 2 Automation: Ventana Benchmark ULTRA Ventana Optiview Main Buffer: Ventana reaction buffer (95-3), PH: 7.47 Ventana CC1 48mins NOT APPLICABLE Chromogen: Other, 37 ºC., Time 1: 8 Mins Detection: Ventana OptiView Kit (76-7) Prediluted Ki67 - Method 3 Participant scored 2/2 (UK NEQAS Slide) and 18/2 (In House slide) using this method. Primary Antibody: Ventana RTU (3-9) , 36 Mins, 37 ºC Prediluted Automation: Ventana Benchmark ULTRA Ventana Optiview Main Buffer: Ventana reaction buffer (95-3) None NOT APPLICABLE Chromogen: Ventana DAB Detection: Ventana OptiView Kit (76-7), 37 ºC Ki67 - Method 4 Participant scored 19/2 (UK NEQAS Slide) and 2/2 (In House slide) using this method. Primary Antibody: Ventana RTU (3-9) , 8 Mins, 2 ºC Dilution 1: 1 Automation: Ventana Benchmark Ventana UltraView DAB Main Buffer: Ventana reaction buffer (95-3) Ventana CC1 standard Chromogen: Ventana Ultraview DAB, Time 1: 16 Mins, Time 2: 2 Mins Detection: Ventana UltraView Kit (76-5), 8 Mins, 2 ºC 66

69 The Alimentary Tract Module: GIST Run 16 Suzanne Parry Introduction First Antibody Antigens Assessed: CD117 DOG-1 Tissue Sections circulated: Number of Registered Participants: 11 Second Antibody GIST, Appendix & Desmoid GIST, Appendix & Desmoid Number of Participants This Run 17 (97%) CD (84%) DOG-1 Gold Standard: CD117 CD117 (c-kit) is a transmembrane tyrosine kinase growth factor receptor involved in cell differentiation. It is expressed by melanocytes, mast cells and interstitial cells of Cajal, and is also expressed in gastrointestinal stromal tumours (GIST). Until recently, patients with GIST faced a limited choice of treatment regimes, which included radiotherapy and surgery with limited success. However, Glivec (Imatinib), which was originally developed for chronic myeloid leukaemia (CML), has proved effective in the targeted treatment of GISTs by directly inhibiting the action of CD117. Assessment Procedure Composite slides comprising of a GIST, appendix and a desmoid tumour, were distributed to all participants for them to stain with CD117 using their routine protocol. Assessment was carried out around a multi-header microscope with each slide being assessed by 4 independent assessors and a microscope lead. All 3 samples on the slide were viewed, and then a final overall score out of 5 was given by each individual assessor. These four scores were then combined to give an overall score out of 2. Features of Optimal Immunostaining: (See Figs 1, 2, 4 & 6A) Good localisation of CD117 to cells of the GIST (Fig 4) Good localisation of CD117 to mast cells in the appendix and desmoid sections (Figs 1 & 6) Good localisation of CD117 to interstitial cells of Cajal in the appendix (Fig 2) No staining of the desmoid tumour (Fig 6) Features of Suboptimal Immunostaining: (See Figs 3, 5 & 6B) Weak and/or patchy staining of the tumour cells of the GIST (Fig 5) Little or no staining of the mast cells Excessive background / non specific staining (Figss 3 & 6B) Staining of the desmoid tumour (Fig 6B) Second Antibody: DOG-1 Discovered on GIST 1 (DOG-1) antibody was initially described in 24 4 and has now started to be recognized as a more specific marker of GISTs than CD117 4,5,6. A recent study has shown DOG-1 sensitivity of 99% and found to be strong and easier to interpret than CD The groups also showed that seven CD117 negative GIST cases were found to be DOG-1 positive, and six of the tumours had PDGFRA mutations. Furthermore, DOG-1 is not expressed in Kaposi sarcomas of the GI tract which have been shown to stain for CD In more cases, combined CD117 and DOG-1 immunohistochemistry has been suggested as a sufficient panel to confirm diagnosis, with CD117 and DOG-1 negative cases tested further with a panel of antibodies including SMA, desmin, S1 and molecular analysis, should be considered 6. Tissue Distribution Circulated tissue were from sequential sections also used for the CD117 distribution (as shown above), comprising of a GIST, appendix and a desmoid tumour. Slides were assessed in the same manner as for the CD117 assessment. Features of Optimal Immunostaining: (See Figs 7 & 8) Good localisation of DOG-1 to cells of the GIST (Figs 7 & 8) Good localisation of DOG-1- to interstitial cells of Cajal No staining of desmoid tumour Features of Suboptimal Immunostaining: (See Figs 9-11) Weak and/or patchy staining of the tumour cells of the GIST (Fig 1) Excessive background or non specific staining (Fig 9) Staining of the desmoid tumour Staining of the mast cells (Note: Mast cells are not expected to stain with DOG-1) References: 1.Cordless et al., Biology of Gastrointestinal Stromal Tumours. J Clin Oncol 24, 22(18): Blay et al., Consensus meeting for the management of gastrointestinal stromal tumors. Report of the GIST Consensus Conference of 2-21 March 24, under the auspices of ESMO. Ann Oncol 25 6: Hasegawa T et al, Gastrointestinal stromal tumors: consistent CD117 immunostaining for diagnosis, and prognostic classification based on tumor size & MIB-1 grade. Hum Pathol. 22 Jun;33(6): Rudolph et al, Gastrointestinal mesenchymal tumors - immunophenotypic classification and survival analysis. 22 Sep;441(3): Epub 22 Jul 6. 5.Miettinen et al, Immunohistochemical spectrum of GISTs at different sites and their differential diagnosis with a reference to CD117 (KIT). Mod Pathol. 2 Oct;13(1): Assessment Summary: Results from the CD117 assessment show an increase in the acceptable pass rate from 64% in the previous assessment run (15) to 81% in the current run. There were also less labs receiving a borderline pass, at 15% compared to 33% in the last run. 5 labs (5%). The fail rate of 5% was similar to the last run. The main reasons for the low marks were due to very weak staining of the GIST or non-specific and background staining predominantly in the desmoid tumour. Inappropriate antibody dilution factor and antigen retrieval were the main causes of these issues. The most popular CD117 antibody choice still remains the Dako polyclonal, used by 87% of participants and showed a pass rate of 8%. Results from the DOG-1 assessment showed an even higher pass rate, with 95% of labs achieving an acceptable pass, and only 4% receiving a borderline. No labs failed the assessment. Weak or uneven staining were the reasons for the lower marks given to the borderline labs. The Leica K9 antibody still remains the most popular DOG-1 antibody. This was used by 77% of participants and showed a pass rate of 98%. This antibody was noted to work well on all the commercial staining platforms, including the Dako Autostainer, Ventana Benchmark and Leica Bond machines. The Ventana SP31 antibody was the next most popular antibody, used by 1 participants, and this showed a pass rate of 1%. Please also note: NEQAS are now advising laboratories to submit normal appendix and gastric tissue for their in house controls. UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 67

70 RUN 16 ALIMENTARY TRACT PATHOLOGY Module Selected Images showing Optimal and Sub-optimal Immunostaining Fig 1. Good demonstration of CD117 in the UK NEQAS ICC distributed appendix. The mast cells show distinct membranous staining, while the background remains clean. Stained with the Dako polyclonal antibody, 1:3, pre-treatment in the PT link with high ph buffer for 2 minutes. Fig 2. Optimal demonstration of the Cajal cells in the muscularis propria of the UK NEQAS distributed appendix section. The smooth muscle cells remain unstained. Section stained with the Dako polyclonal antibody, on the Ventana ULTRA with no pre-treatment and Optiview detection kit. Fig 3. Suboptimal CD117 staining in the UK NEQAS distributed appendix (compare to Fig 2). The section shows non-specific staining of the smooth muscle and background staining throughout the section. Stained with the Dako polyclonal antibody, 1:3, on the Leica BondMax with no pre-treatment. Fig 4. Good example of CD117 staining of the UK NEQAS distributed GIST: The cytoplasmic and membranous staining of the tumour cells is strong and well localised. Section stained with the Dako polyclonal antibody, 1:3, with CC1 standard pre-treatment on the Ventana Benchmark XT. Fig 5. Suboptimal staining of the UK NEQAS distributed GIST section (compare to Fig 4). The staining is very weak and patchy, most likely caused by an inappropriate pre-treatment method. Section stained with the Dako polyclonal antibody, 1:5 on the Dako Autostainer. Pre-treatment ª«± ª ²ª ³ µ ª ¹ º Fig 6. CD117 staining in the UK NEQAS distributed desmoid tumour. (A) Optimal staining of the mast cells, while the tumour remains negative as expected. (B) Suboptimal staining, showing excessive background and non-specific staining of the desmoid tumour. 68

71 RUN 16 ALIMENTARY TRACT PATHOLOGY Module Selected Images showing Optimal and Sub-optimal Immunostaining Fig 7. Good demonstration of DOG-1 in the UK NEQAS GIST. The tumour cells show strong crisp staining, while the non-tumour endothelial cells remain unstained. Section stained with the Leica K9 antibody, 1:1, pre-treated with ER2 for 2 minutes on the Leica BondMax. Fig 8. Optimal staining of DOG-1 in the UK NEQAS GIST, showing strong and well localised staining of the tumour cells. Section stained with the Leica K9 antibody, 1:4, on the Ventana Benchmark XT with CC1 standard pre-treatment. Fig 9. Borderline pass staining of DOG-1 in the UK NEQAS distributed appendix. The section shows inappropriate non-specific staining of lymphocytes. Section stained with the Cell Marque SP31 pre-diluted antibody on the Leica Bond III, with 3 minutes pre-treatment with ER2. Fig 1. Poor demonstration of DOG-1 in the UK NEQAS GIST (compare to Figs 7&8). The staining is patchy and much weaker than expected. This is most likely due to insufficient pre-treatment. Section stained with the Leica K9 antibody, 1:1, on the Ventana Benchmark XT with no pre-treatment. Fig 11. Suboptimal staining of DOG-1 in the UK NEQAS GIST. Although the tumour is demonstrated with the optimal level of intensity, the section also shows heat artefact, caused by excessive heat antigen retrieval. Stained with the Leica K9 antibody, 1:2, on the Dako Autostainer, pre-treated in a pressure cooker within a microwave. Fig 12. Excellent example of an in house control submitted for assessment. The tissue included a GIST along with normal gastric material. Both the GIST (A) and cells of Cajal (B) are nicely demonstrated. The participant received a score of 2/2 for providing ideal control material with optimal staining. 69

72 Run 16 ALIMENTARY TRACT PATHOLOGY Module GRAPHICAL REPRESENTATION OF PASS RATES 2 RUN 16V CD117 on NEQAS Sections Individual 4 = (%) 35 RUN 16W CD117 on in-house Sections Individual 4 = (%) 16 5 = (%) 6 = (%) 7 = (%) 8 = 4 (4%) = (%) 6 = (%) 7 = (%) 8 = 3 (3%) no. of returns = 1 (1%) 1 = (%) 11 = 3 (3%) 12 = 13 (12%) 13 = 1 (9%) no. of returns = (%) 1 = 1 (1%) 11 = (%) 12 = 4 (4%) 13 = 6 (6%) 4 14 = 12 (11%) 15 = 16 (15%) 16 = 19 (18%) 17 = 16 (15%) = 8 (7%) 15 = 8 (7%) 16 = 34 (32%) 17 = 17 (16%) = 7 (7%) 19 = 5 (5%) 2 = 1 (1%) = 14 (13%) 19 = 8 (7%) 2 = 4 (4%) Summary Summary 16-2 = 48 (45%) 16-2 = 77 (72%) = 38 (36%) = 22 (21%) 1-12 = 16 (15%) 1-12 = 5 (5%) - 9 = 5 (5%) - 9 = 3 (3%) Median = 14.5 Median = 15. no. of returns RUN 16Vb DOG1 on NEQAS Sections Individual 4 = (%) 5 = (%) 6 = (%) 7 = (%) 8 = (%) 9 = (%) 1 = (%) 11 = (%) 12 = 4 (4%) 13 = 8 (9%) 14 = 7 (8%) 15 = 4 (4%) 16 = 17 (18%) 17 = 12 (13%) 18 = 1 (11%) 19 = 19 (21%) 2 = 11 (12%) no. of returns RUN 16Wb DOG1 on in-house Sections Individual 4 = (%) 5 = (%) 6 = (%) 7 = (%) 8 = (%) 9 = (%) 1 = 4 (4%) 11 = 1 (1%) 12 = 4 (4%) 13 = 8 (9%) 14 = 5 (5%) 15 = 1 (11%) 16 = 17 (18%) 17 = 1 (11%) 18 = 16 (17%) 19 = 6 (7%) 2 = 11 (12%) Summary Summary 16-2 = 69 (75%) 16-2 = 6 (65%) = 19 (21%) = 23 (25%) 1-12 = 4 (4%) 1-12 = 9 (1%) - 9 = (%) - 9 = (%) Median = 16. Median = 15. 7

73 Run 16 ALIMENTARY TRACT PATHOLOGY Module ANTIBODIES AND OTHER TECHNICAL PARAMETERS EMPLOYED BY PARTICIPANTS IN THE ALIMENTARY TRACT PATHOLOGY MODULE The following tables record the number of participants (N) using each primary antibody and the percentage (%) of these participants achieving acceptable staining (a score»12/2) on UK NEQAS sections. Alimentary Tract Pathology Run: 16 Alimentary Tract Pathology Run: 16 Primary Antibody : CD117 Antibody Details N % Cell Marque 117R/S-xx (YR145) 2 1 Dako A452 (rb poly) 93 8 Epitomics AC-29 (EP1) 1 1 Leica/Novocastra NCL-CD117 (T595) 1 1 Leica/Novocastra NCL-L-CD117 (T595) 2 5 Leica/Novocastra RTU-CD117 (T595) 1 NeoMarker RB 938 (rb poly) 1 1 Ventana (9.7) 6 1 Primary Antibody : DOG1 Antibody Details N % Abcam TMEM16A (ab53212) 1 1 Biocare CM 385 (1.1) 1 1 Cell Marque 244R-14/15/16 (SP31) 3 1 Cell Marque 244R-17/18 (SP31) 3 67 Leica NCL-L-DOG-1 (K9) Leica PA219 (K9) 16 1 Menarini MP-385-CM1/1 1 1 Other 1 Thermo RM-9132-R7 (SP31) 1 1 Ventana (SP31) Alimentary Tract Pathology Run: 16 Alimentary Tract Pathology Run: 16 CD117 DOG1 CD117 DOG1 Heat Mediated Retrieval N % N % ¼½¾ ÀÁ Á ¾ à ÄÄÅ ÆÇÈÉÊË 1 1 Dako Omnis Dako PTLink Leica ER1 1 mins Leica ER1 2 mins Leica ER1 3 mins Leica ER2 1 mins Leica ER2 2 mins Leica ER2 3 mins Microwave 1 1 None 1 6 Other 1 1 Pressure Cooker Pressure Cooker in Microwave Oven 1 1 Steamer Ventana CC1 16mins Ventana CC1 24mins 1 1 Ventana CC1 32mins Ventana CC1 36mins 4 75 Ventana CC1 48mins 2 1 Ventana CC1 52mins Ventana CC1 56mins 1 1 Ventana CC1 64mins Ventana CC1 mild Ventana CC1 standard Ventana CC2 44mins 1 1 Ventana CC2 standard 1 Enzyme Mediated Retrieval N % N % AS PER KIT 1 1 MP BioMedicals Trypsin NOT APPLICABLE

74 Run 16 ALIMENTARY TRACT PATHOLOGY Module Alimentary Tract Pathology Run: 16 Alimentary Tract Pathology Run: 16 CD117 DOG1 CD117 DOG1 Detection N % N % AS PER KIT BioGenex SS Polymer (QD 42-YIKE) 1 1 BioGenex SS Polymer (QD 43-XAKE) 1 1 Dako EnVision FLEX ( K8/1) 1 1 Dako EnVision FLEX+ ( K82/12) Dako Envision HRP/DAB ( K57) Dako Envision+ HRP mouse K44/5/6/7 2 1 Dako Envision+ HRP rabbit K48/9/1/ Leica Bond Intense R Detection (DS9263) 1 1 Leica Bond Polymer Refine (DS98) MenaPath X-Cell Plus (MP-XCP) None 1 1 Other Ventana iview system (76-91) Ventana OptiView Kit (76-7) Ventana UltraView Kit (76-5) Chromogen N % N % AS PER KIT BioGenex Liquid DAB (HK153-5K) 1 1 BioGenex liquid DBA (HK-124-7K) 1 1 Dako EnVision Plus kits Dako FLEX DAB Dako REAL EnVision K57 DAB Leica Bond Polymer Refine kit (DS98) menapath xcell kit DAB (MP-86) Other Ventana DAB Ventana iview Ventana Ultraview DAB Alimentary Tract Pathology Run: 16 CD117 DOG1 Automation N % N % BioGenex GenoMX 6i Dako Autostainer Link Dako Autostainer plus Dako Autostainer Plus Link Dako Omnis LabVision Autostainer Leica Bond Max Leica Bond-III Menarini - Intellipath FLX Ventana Benchmark Ventana Benchmark ULTRA Ventana Benchmark XT BEST METHODS - Gold Standard Antibody A selection from just a few of the best methods employed by participants CD117 - Method 1 Participant scored 19/2 (UK NEQAS Slide) and 19/2 (In House slide) using this method. Primary Antibody: Dako A452 (rb poly), 32 Mins, 36 ºC Dilution 1: 3 Automation: Ventana Benchmark ULTRA Ventana UltraView DAB Main Buffer: Ventana reaction buffer (95-3), PH: 7.2 Ventana CC1 64mins NOT APPLICABLE Chromogen: Ventana Ultraview DAB Detection: Ventana UltraView Kit (76-5), 36 ºC Prediluted CD117 - Method 2 Participant scored 2/2 (UK NEQAS Slide) and 2/2 (In House slide) using this method. Primary Antibody: Dako A452 (rb poly), 3 Mins, ROOM ºC Dilution 1: 1/3 Automation: Dako Autostainer Link 48 Dako FLEX+ kit Main Buffer: Dako FLEX wash buffer, PH: 7.6 Dako PTLink, Buffer: Dako EnVision High ph antigen retrieval sln. NOT APPLICABLE Chromogen: Dako EnVision Plus kits, Room ºC., Time 1: 5 Mins, Time 2: 5 Mins Detection: Dako EnVision FLEX+ ( K82/12), 3 Mins, ROOM ºC Prediluted 72

75 Run 16 ALIMENTARY TRACT PATHOLOGY Module CD117 - Method 3 Participant scored 17/2 (UK NEQAS Slide) and 18/2 (In House slide) using this method. Primary Antibody: Dako A452 (rb poly), 2 Mins, 37 ºC Dilution 1: 1 Automation: Dako Omnis Other Main Buffer: AS PER KIT Dako Omnis, Buffer: DAKO HIGH PH TRS Chromogen: AS PER KIT, 37 ºC., Time 1: 5 Mins Detection: AS PER KIT, 2 Mins, 37 ºC Prediluted CD117 - Method 4 Participant scored 17/2 (UK NEQAS Slide) and 16/2 (In House slide) using this method. Primary Antibody: Ventana (9.7), 6 Mins, 36 ºC Automation: Ventana Benchmark ULTRA Ventana UltraView DAB Main Buffer: Ventana reaction buffer (95-3) Ventana CC1 standard Chromogen: Ventana Ultraview DAB Detection: Ventana UltraView Kit (76-5) BEST METHODS - Secondary Antibody A selection from just a few of the best methods employed by participants DOG1 - Method 1 Participant scored 2/2 (UK NEQAS Slide) and 17/2 (In House slide) using this method. Primary Antibody: Leica PA219 (K9) Prediluted Automation: Leica Bond-III Leica BondMAx Refine KIT Main Buffer: Bond Wash Buffer (AR959) Leica ER1 2 mins Chromogen: Leica Bond Polymer Refine kit (DS98) Detection: Leica Bond Polymer Refine (DS98) Prediluted DOG1 - Method 2 Participant scored 2/2 (UK NEQAS Slide) and 18/2 (In House slide) using this method. Primary Antibody: Leica NCL-L-DOG-1 (K9), 15 Mins, RT ºC Dilution 1: 2 Automation: Leica Bond Max Leica BondMAx Refine KIT Main Buffer: Bond Wash Buffer (AR959) Leica ER2 2 mins NOT APPLICABLE Chromogen: Leica Bond Polymer Refine kit (DS98), RT ºC., Time 1: 1 Mins, Time 2: 1 Mins Detection: Leica Bond Polymer Refine (DS98), 8 Mins, RT ºC Prediluted 73

76 Run 16 ALIMENTARY TRACT PATHOLOGY Module DOG1 - Method 3 Participant scored 2/2 (UK NEQAS Slide) and 2/2 (In House slide) using this method. Primary Antibody: Leica NCL-L-DOG-1 (K9), 6 Mins, 37 ºC Dilution 1: 1/5 Automation: Ventana Benchmark XT Ventana UltraView DAB Main Buffer: Ventana reaction buffer (95-3), PH: 7.6 Ventana CC1 standard Chromogen: Ventana Ultraview DAB, 37 ºC., Time 1: 8 Mins, Time 2: 8 Mins Detection: Ventana UltraView Kit (76-5), 8 Mins, 37 ºC DOG1 - Method 4 Participant scored 2/2 (UK NEQAS Slide) and 17/2 (In House slide) using this method. Primary Antibody: Thermo RM-9132-R7 (SP31), 3 Mins, 2 ºC Dilution 1: 1 Automation: Dako Autostainer Link 48 Dako FLEX+ kit Main Buffer: AS PER KIT Dako PTLink NOT APPLICABLE Chromogen: AS PER KIT, 2 ºC., Time 1: 5 Mins, Time 2: 5 Mins Detection: AS PER KIT, 2 Mins, 2 ºC 74

77 Lynch Syndrome/HNPCC Run 16 Merdol Ibrahim and Suzanne Parry Gold Standard Antigens Assessed: MLH1 PMS2 Tissue Sections circulated: Number of Registered Participants: 73 Number of Participants This Run 67 (92%) General Introduction In Lynch/ Hereditary Non-Polyposis Colorectal Cancer (HNPCC) syndrome, tumours that have lost mismatch repair (MMR) function, immunohistochemistry (IHC) can be used to show abnormal MMR protein expression. Lynch/HNPCC syndrome sufferers inherit one germline mutant MMR allele and one normal, wild-type MMR allele. During tumour formation, the normal allele is inactivated by mutation or loss or other mechanism, thus leaving no expression from functional alleles. MMR protein abnormal expression in Lynch/HNPCC tumours may thus be detected by IHC in two patterns: (i) Complete loss of expression (when there is no expression of that MMR protein or only expression of a truncated or mutant protein to which the antibody does not bind) or (ii) Patchy/weak expression (if the mutation generates a prematurely truncated but variably stable protein, or a protein with alterations to the epitope recognised by the antibody, e.g. missense mutation). It is important to realise, however, that a small proportion, perhaps 5-2% of Lynch/HNPCC-related tumours do not exhibit any abnormality on analysis by IHC even though they have lost MMR function, as manifested by microsatellite instability (MSI) and this may be due to mutations that functionally inactivate the MMR protein but allow its expression as a stable protein with nuclear localisation and intact epitope. Mismatch Repair Markers Where possible, IHC should be carried out using an automated staining machine. Manual IHC is prone to significant variation, both within and between batches. Whereas automated immunostaining generates more consistent and reproducible staining patterns. There are 4 MMR antibodies available to aid in the differential identification of colorectal cancers. These are MLH1, MSH2, MSH6 & PMS2, and the use of all 4 antibodies is recommended rather than just using the 2 markers (MLH1 & MSH2). There is a heterodimeric association of proteins, such that a loss of MLH1 expression is almost always accompanied by the loss of PMS2 expression, and a loss of MSH2 is almost always accompanied with the loss of MSH6. Therefore, the expression changes to the heterodimeric binding partners of these proteins acts as a very useful confirmatory finding. The antigenic epitopes for the four anti-mmr antibodies are particularly fixation-sensitive, so immunostaining patterns should only be assessed in well-fixed regions of the tissue section, for example where the adjacent normal epithelium or the intra- or peri-tumoural lymphoid cells or stromal cells, such as fibroblasts, show clear, strong nuclear immunopositivity for all four of the MMR proteins. Any poorly fixed parts of the tumour should not be used to confirm a diagnosis. Clinical Interpretation & Reporting Guidelines We recommend reporting MMR IHC findings in tumours as either: a) Normal: Demonstrating strong immunopositivity of the tumour cell nuclei for all four MMR proteins, similar in staining intensity to that of the adjacent normal epithelium or intratumoural lymphoid cells or stromal cells. Normal Appendix & Colonic Tumour Second Antibody Normal Appendix, & Colonic Tumour. b) Negative: Showing complete loss of staining of one or more MMR proteins, relative to the strongly intensive immunopositivity of the adjacent normal epithelium or intra-tumoural lymphoid cells or stromal cells. c) Patchy/weak: Staining intensity of one or two of the MMR proteins is either weak or patchy in the tumour cell nuclei, compared to adjacent normal epithelium or intra-tumoural lymphoid cells or stromal cells, that show strong nuclear positivity in a well fixed region of the tumour. This patchy/weak staining may sometimes be accompanied by staining in the cytoplasm rather than the nucleus, whereas, the adjacent normal epithelium or intra-tumoural activated lymphocytes or stromal cells show the usual pattern of strong nuclear immunopositivity for the MMR protein(s). This suggests that destabilisation of the MMR protein complexes has occurred and the proteins are no longer bound to the nuclear DNA. Reports Reports should conclude with a statement that evidence of abnormal MMR expression by IHC has, or has not, been found. If abnormal MMR expression has been found the comment is made that this is compatible with, but not necessarily diagnostic of Lynch/HNPCC syndrome, and that, if not already organised, referral of the individual and their family should be made to clinical genetic services. If abnormal MMR expression has not been found, the covering statement is made that MSI studies should be carried out to include or exclude MMR functional abnormality in the tumour (because of the proportion of tumours with MSI that do not exhibit an abnormality of MMR protein expression). Because abnormality of MMR protein expression is evidence in itself of loss of MMR function in that tumour, a comment may also be made that MSI studies are not needed. A small minority of tumours may show unusual combinations of abnormal MMR protein expression, e.g. loss of both MLH1 and MSH2, or three out of the four MMR proteins. MSH2, or three out of the four MMR proteins. It is suggested that the IHC be repeated and then a clinico-pathological review of the case be carried out as necessary, to decide whether to proceed with testing of other tumours from the same individual or family, or proceeding directly to mutation detection studies. Further Discussion on MMR proteins Abnormality of MMR protein expression is not in itself necessarily diagnostic of Lynch /HNPCC syndrome, as around 15% of sporadic colon cancers show loss of MLH1 staining due to acquired promoter methylation, and some sporadic colorectal tumours will also lose MSH2, MSH6 or PMS2. While there is evidence that abnormal MMR expression in rectal cancers is more predictive of Lynch syndrome, it is by no means an absolute. Where available and appropriate, it can be suggested that testing for the BRAF V6E mutation is a way of distinguishing a proportion of MLH1-negative tumours as sporadic. It is important to appreciate the full interpretation of the tumour, where the intensity of immunopositivity in tumour 75 UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent

78 The Alimentary Tract Module: Lynch Syndrome/HNPCC Run 16 cell nuclei can be directly compared with the strong nuclear staining intensity in well fixed adjacent normal epithelium or lymphoid cells or stromal cells. Artefactually distorted staining patterns due to poor fixation (affecting both stromal cells as well as tumour cells) in central parts of the tumour are irrelevant. It should be borne in mind that geneticists may use the immunohistochemical expression pattern of MMR proteins to interpret mutations of uncertain significance, and that both false -negative and false-positive staining (or interpretation) may potentially cause misinterpretation of DNA sequence findings. This could result in a mutation being interpreted as pathogenic when it is, in fact, neutral in effect, and thus result in inappropriate predictive genetic testing being offered to the family. Hence, care is needed in both staining and interpretation. Assessment Procedure: Composite slides comprising of a normal appendix and a colonic tumour (with loss of MMR protein), were distributed to all participants for them to stain with MSH2 (1st Antibody) and MSH6 (2nd antibody) using their routine protocol. Assessment was carried out around a multi-header microscope with each slide being assessed by 4 independent assessors and a microscope lead. Both samples on the slide were viewed, and then a final overall score out of 5 was given by each individual assessor. These four scores were then combined to give an overall score out of 2. Features of Optimal Immunostaining: Appendix: (See figs 1, 2, 7, 8 & 12) Strong nuclear staining of the epithelium at the base and lower third-half of the crypts with fading of nuclear staining intensity in the middle and upper thirds of the crypts, to negative or weak staining of epithelial nuclei at the luminal surface Strong staining of lymphocytes, particularly in the lymphoid follicles. Tumour with loss of MMR protein: (See figs 4 & 1) Strong nuclear staining in the normal intra-tumoural and stromal cells. No staining in the tumour cells. Features of Sub-Optimal Immunostaining: Appendix: (See figs 3 & 9) Weak, uneven, partially missing staining of relevant cells. Excessive background or non-specific specific staining. Tumour with loss of MMR protein: (See figs 5, 6 & 11) Weak or no staining in the normal intra-tumoural and stromal cells. Excessive background or non-specific staining. Unexpected staining in the tumour cells. Assessment Summary: Although the acceptable pass rate (score of 13/2) for MSH2 was higher than for Run 13, the overall percentage of labs that passed the assessment (including borderline passes) was lower than for Run 13, i.e. an overall 69% pass rate in the current assessment and an overall 79% pass rate in Run 13. More labs failed this assessment (31%) (a score of 9/2) compared to Run 13, where 21% of labs showed unacceptable staining. The MSH6 antibody showed similar results to the MSH2 antibody, with an overall pass rate of 63%, of which 46% were at an acceptable level of staining and 17% were borderline. This was marginally lower than the previous assessment (Run 13), which showed an overall pass rate of 65%. The fail rate for the current run was disappointingly high at 37%. The Ventana antibodies were the most popular choice for both MSH2 (Clone G ) & MSH6 (Clone 44), used by the majority of labs, and showed pass rates of 57% and 53% respectively. Use of the Dako MSH2 and MSH6 antibodies also increased for this assessment Run: The Dako MSH2 (Clone FE11) was used by 1 labs, compared to only 5 labs in Run 13. Labs using this antibody showed a pass rate of 54%. The Dako MSH6 (Clone EP49) antibody was used by 17 labs for this Run, compared to just 8 labs previously, and this showed a pass rate of 72%. The main reason for sub-optimal results was due to weak or very little staining in the appendix or the normal cells in the tumour sections, or showed excessive background staining. The reasons for failure was due to no staining of the intratumour and stromal cells in the tumour section, or because of unexpected staining in the tumours cells (known to be negative). Non-specific unexpected staining of endothelial cells was also seen. This was noticeably related to (but not all) participants using the Cell Marque and Ventana G antibody clone. Laboratories did not fail for this, as the diagnostic result was not compromised, however, a mark was deducted from slides showing this staining. In-House Control Tissue Recommendations The assessors recommend the use of appendix or normal colon as appropriate control tissues to use for immunohistochemistry for all 4 MMR proteins. Although tonsil has also been used as a control tissue, the very intense nuclear staining signal in both the lymphoid follicles and the overlying squamous epithelium make it difficult to gauge the sensitivity of the stain. Therefore, appendix or colon are more appropriate tissues to use for control purposes. References 1. Vasen HF, Möslein G, Alonso A et al.,(27) Guidelines for the clinical management of Lynch syndrome (hereditary non-polyposis cancer). J Med Genet. 44 6): Free download from: jmg.bmj.com/cgi/content/full/44/6/ Dr Philippa Brice. Biomarkers in familial colorectal cancer screening. Expert workshop, 14th February 26. Public Health Genetics Unit, Cambridge, UK. Free download from: file/2743/. 3. Arends MJ, Frayling I. Mismatch Repair Deficiency in Hereditary and Sporadic Colorectal Cancer. In: The Effective Management of Colorectal Cancer (4 th Edition), UK Key Advances in Clinical Practice series. Eds: Cunningham D, Topham C, & Miles A. ISBN X. 25. Chapter 2, pp Frayling IM, Happerfield L, Mattocks C, Oakhill K, Arends MJ. Application of Molecular Diagnostics to Hereditary Nonpolyposis Colorectal Cancer (chapter 32). In: Molecular Diagnostics, For the Clinical Laboratorian (2 nd Edition). Eds: Coleman WB & Tsongalis GJ. Humana Press Inc., NJ. 25. ISBN: , ISBN13: ; ISBN1: pp Poulogiannis G, Frayling IM, Arends MJ. DNA Mismatch Repair Deficiency in Sporadic Colorectal Cancer and Lynch Syndrome. Histopathology 21; 56: a. 6. Vasen F, et al. Revised guidelines for the clinical management of Lynch syndrome (HNPCC): recommendations by a group of European experts. Gut 213;62: UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 76

79 77

80 RUN 16 HNPCC Module Selected Images showing Optimal and Sub-optimal Immunostaining Fig 1. Optimal demonstration of MLH1 in the UK NEQAS distributed appendix. The image shows strong nuclear staining of the epithelial and stromal cells, with a fading of staining intensity towards the luminal surface of the epithelial crypts. Section stained with the Leica ES5 antibody, 1:3, on the Leica Bond III with ER2 pre-treatment for 3 minutes. Fig 2. Optimal demonstration of MLH1 in the 2 UK NEQAS distributed colonic tumours: (A) with normal MLH1 expression, shows strong nuclear staining of virtually all the tumour cells. (B) Loss of MLH1 expression, therefore, all the tumour cells are negative. Both sections show distinct staining in the lymphocytes and stromal cells. Both sections stained with the Dako Fig 3. Unacceptable MLH1 staining of the UK NEQAS distributed appendix. The staining is excessive with a high level of background. This is most likely due to excessive antigen retrieval. Section stained with the Leica ES5 antibody, 1:1, on the Ventana Benchmark XT, with CC1 pre-treatment for 64 minutes and Optiview detection. Fig 4. Two examples showing suboptimal MLH1 staining in the UK NEQAS distributed samples. Both (A) the appendix and (B) the colonic tumour show weak nuclear staining (compare to Figs 1&2). This is most likely due to insufficient antigen retrieval. Sections stained with the Leica ES5 antibody, 1:1, on the Leica BondMax with ER1 pre-treatment for 1 minutes. Fig 5. Poor demonstration of MLH1 in the UK NEQAS colonic tumour with loss of MLH1. The section shows inappropriate cytoplasmic staining in the neoplastic cells. Although the tumour nuclei appear negative, the interpretation is compromised. Again, excessive antigen retrieval is most likely the cause of this reaction. Same protocol as Fig 3. Fig 6. Suboptimal staining of MLH1 in the UK NEQAS colonic tumour with loss of MLH1. Although the tumour nuclei are negative as expected, there is also some inappropriate non-specific staining of endothelial cells. This staining pattern was observed in several sections stained with a particular batch of the G antibody clone. 78

81 RUN 16 HNPCC Module Selected Images showing Optimal and Sub-optimal Immunostaining Fig 7. Optimal demonstration of PMS2 in the UK NEQAS distributed appendix, showing strong to moderate nuclear staining in the basal and lower half of the epithelial crypts, which fades towards the luminal surface. Section stained with the Dako EP51 RTU antibody, on the Leica BondMax, pre-treatment with ER2 for 2 minutes. Fig 8. Optimal demonstration of PMS2 in the 2 UK NEQAS colonic tumours. (A) With normal expression of PMS2, shows moderate to strong nuclear staining of the tumour cells. (B) With loss of PMS2, therefore, the tumour cells remain negative. Both sections show the expected distinct staining in the lymphocytes and stromal cells. Same protocol as Fig 7. Fig 9. Suboptimal demonstration of PMS2 in the UK NEQAS distributed appendix section. Although most of the cells expected to stain are demonstrated, the staining is weak (compare to Fig 7). Section stained with the Ventana EPR3947 prediluted antibody on the Benchmark XT. Fig 1. Poor demonstration of PMS2 in the UK NEQAS distributed colonic tumours. In (A), with normal expression of PMS2, the proportion of positive cells and the staining intensity is much weaker than expected. In (B), with loss of PMS2, none of the stromal cells are staining, and therefore the staining result is unreliable. (Compare both sections to Fig 8). Fig 11. Poor demonstration of PMS2 in the UK NEQAS distributed (A) colonic tumour with normal expression, and (B) with loss of PMS2. Both sections show excessive background staining, which is most likely due to excessive antigen retrieval. Sections stained with the Pharmingen A16-4 antibody, on the Leica BondMax, with pre-treatment for 4 minutes with Fig 12. Granular and punctate staining pattern frequently seen on sections stained using the Ventana EPR3947 antibody and platform. Although the staining pattern did not compromise the result, the assessors made a comment on this. 79

82 Run 16 HNPCC Module GRAPHICAL REPRESENTATION OF PASS RATES 16 RUN 16Y MLH1 on NEQAS Sections Individual 4 = (%) 16 RUN 16Z MLH1 on in-house Sections Individual 4 = (%) 14 5 = (%) 6 = (%) 14 5 = (%) 6 = (%) 12 7 = (%) 8 = 6 (9%) 12 7 = (%) 8 = 6 (9%) no. of returns = (%) 1 = 1 (1%) 11 = 3 (4%) 12 = 9 (13%) 13 = 3 (4%) no. of returns = (%) 1 = 1 (1%) 11 = 3 (4%) 12 = 6 (9%) 13 = 5 (7%) 4 14 = 7 (1%) 15 = 7 (1%) 4 14 = 5 (7%) 15 = 9 (13%) 2 16 = 15 (22%) 17 = 11 (16%) 2 16 = 16 (24%) 17 = 5 (7%) = 2 (3%) 19 = 1 (1%) 2 = 2 (3%) = 5 (7%) 19 = 3 (4%) 2 = 3 (4%) Summary Summary 16-2 = 31 (46%) 16-2 = 32 (48%) = 17 (25%) = 19 (28%) 1-12 = 13 (19%) 1-12 = 1 (15%) - 9 = 6 (9%) - 9 = 6 (9%) Median = 14.5 Median = 14.5 no. of returns RUN 16Yb PMS2 on NEQAS Sections Individual 4 = (%) 5 = (%) 6 = (%) 7 = (%) 8 = 2 (3%) 9 = 2 (3%) 1 = 4 (6%) 11 = 8 (12%) 12 = 1 (14%) 13 = 5 (7%) 14 = 2 (3%) 15 = 7 (1%) 16 = 2 (29%) 17 = 7 (1%) 18 = 1 (1%) 19 = (%) 2 = 1 (1%) no. of returns RUN 16Zb PMS2 on in-house Sections Individual 4 = (%) 5 = (%) 6 = (%) 7 = (%) 8 = 1 (1%) 9 = 3 (4%) 1 = 2 (3%) 11 = 3 (4%) 12 = 4 (6%) 13 = 7 (1%) 14 = 8 (12%) 15 = 12 (18%) 16 = 15 (22%) 17 = 12 (18%) 18 = 1 (1%) 19 = (%) 2 = (%) Summary Summary 16-2 = 29 (42%) 16-2 = 28 (41%) = 14 (2%) = 27 (4%) 1-12 = 22 (32%) 1-12 = 9 (13%) - 9 = 4 (6%) - 9 = 4 (6%) Median = 13.5 Median = 13. 8

83 Run 16 HNPCC Module ANTIBODIES AND OTHER TECHNICAL PARAMETERS EMPLOYED BY PARTICIPANTS IN THE HNPCC MODULE The following tables record the number of participants (N) using each primary antibody and the percentage (%) of these participants achieving acceptable staining (a score >12/2) on UK NEQAS sections. HNPCC Run: 16 Primary Antibody : MLH1 Antibody Details N % BD Pharmingen (G168-15) 8 63 BD Pharmingen (G ) 2 Biocare medical CM/PM 22 (G168-15) 2 1 Dako Flex RTU IR79/IS79 (ES5) 7 71 Dako M364 (ES5) 5 1 Leica Bond RTU PA61 (ES5) 3 67 Novocastra NCL-L-MLH1 (ES5) Other 1 1 Ventana (M1) 2 55 Primary Antibody : PMS2 Antibody Details N % BD Bio/Pharmingen (A16-4) Cell Marque 288M -16 (MRQ28) 1 Cell Marque 288R -17/18 (EPR3947) 3 67 Dako M3647 (EP51) 8 88 Dako RTU FLEX IR87 (EP51) 6 67 Leica/Novoca NCL-L-PMS2 (MOR4G) 3 67 Other 2 5 Ventana (EPR3947) HNPCC Run: 16 HNPCC Run: 16 MLH1 PMS2 MLH1 PMS2 Heat Mediated Retrieval N % N % Biocare Decloaking Chamber Dako Omnis Dako PTLink Leica ER1 2 mins 3 67 Leica ER1 3 mins 2 5 Leica ER2 2 mins Leica ER2 3 mins Leica ER2 4 mins Pressure Cooker Steamer Ventana CC1 2mins 1 Ventana CC1 32mins Ventana CC1 36mins 1 Ventana CC1 4mins Ventana CC1 48mins 1 2 Ventana CC1 56mins Ventana CC1 64mins Ventana CC1 72mins Ventana CC1 8mins Ventana CC1 88mins Ventana CC1 92mins 4 5 Ventana CC1 extended Ventana CC1 standard Ventana CC2 72mins 1 1 Ventana CC2 standard 1 1 Enzyme Mediated Retrieval N % N % AS PER KIT NOT APPLICABLE

84 Run 16 HNPCC Module HNPCC Run: 16 HNPCC Run: 16 MLH1 PMS2 MLH1 PMS2 Detection N % N % AS PER KIT Biocare polymer (M4U534) Dako EnVision FLEX+ ( K82/12) Dako Envision+ HRP mouse K44/5/6/ Leica Bond Polymer Define (DS9713) 1 1 Leica Bond Polymer Refine (DS98) Other Ventana OptiView Kit (76-7) Ventana UltraView Kit (76-5) Chromogen N % N % AS PER KIT DAKO DAB+ 1 Dako DAB+ Liquid (K3468) 1 1 Dako EnVision Plus kits Dako FLEX DAB Leica Bond Polymer Refine kit (DS98) Other Ventana DAB Ventana Ultraview DAB HNPCC Run: 16 MLH1 PMS2 Automation N % N % BioGenex GenoMX 6i 1 1 Dako Autostainer Link Dako Autostainer plus Dako Autostainer Plus Link Dako Omnis LabVision Autostainer Leica Bond Max Leica Bond-III Menarini - Intellipath FLX None (Manual) 1 Ventana Benchmark 1 Ventana Benchmark ULTRA Ventana Benchmark XT BEST METHODS - Gold Standard Antibody A selection from just a few of the best methods employed by participants MLH1 - Method 1 Participant scored 17/2 (UK NEQAS Slide) and 2/2 (In House slide) using this method. Primary Antibody: Biocare medical CM/PM 22 (G168-15), 24 Mins, 36 ºC Dilution 1: 15 Automation: Ventana Benchmark ULTRA Ventana Optiview Main Buffer: Ventana reaction buffer (95-3) Ventana CC1 64mins NOT APPLICABLE Chromogen: Other, 36 ºC., Time 1: 12 Mins, Time 2: 12 Mins Detection: Ventana OptiView Kit (76-7), 16 Mins, 36 ºC Prediluted MLH1 - Method 2 Participant scored 18/2 (UK NEQAS Slide) and 18/2 (In House slide) using this method. Primary Antibody: Novocastra NCL-L-MLH1 (ES5), 3 Mins, 2 ºC Dilution 1: 1 Automation: Dako Autostainer Link 48 Dako FLEX+ kit Main Buffer: Dako FLEX wash buffer, PH: 7.6 Dako PTLink, Buffer: HIGH PH TRS, PH: 9 NOT APPLICABLE Chromogen: Dako FLEX DAB, 2 ºC., Time 1: 1 Mins, Time 2: 1 Mins Detection: Dako EnVision FLEX+ ( K82/12), 15 Mins, 2 ºC Prediluted 82

85 Run 16 HNPCC Module MLH1 - Method 3 Participant scored 2/2 (UK NEQAS Slide) and 16/2 (In House slide) using this method. Primary Antibody: Dako Flex RTU IR79/IS79 (ES5), 3 Mins, 21 ºC Prediluted Automation: Leica Bond Max Leica BondMAx Refine KIT Main Buffer: Bond Wash Buffer (AR959) Leica ER2 2 mins Chromogen: Leica Bond Polymer Refine kit (DS98), 21 ºC., Time 1: 1 Mins, Time 2: 5 Mins Detection: Leica Bond Polymer Refine (DS98), 8 Mins, 21 ºC Prediluted MLH1 - Method 4 Participant scored 2/2 (UK NEQAS Slide) and 15/2 (In House slide) using this method. Primary Antibody: Novocastra NCL-L-MLH1 (ES5), 3 Mins, RT ºC Dilution 1: 8 Automation: Leica Bond-III Other Main Buffer: Bond Wash Buffer (AR959) Leica ER2 3 mins NOT APPLICABLE Chromogen: Leica Bond Polymer Refine kit (DS98) Detection: Leica Bond Polymer Refine (DS98) Prediluted BEST METHODS - Secondary Antibody A selection from just a few of the best methods employed by participants PMS2 - Method 1 Participant scored 17/2 (UK NEQAS Slide) and 17/2 (In House slide) using this method. Primary Antibody: Leica/Novoca NCL-L-PMS2 (MOR4G), 32 Mins, 37 ºC Dilution 1: 1/2 Automation: Leica Bond Max Leica BondMAx Refine KIT Main Buffer: Bond Wash Buffer (AR959) Leica ER2 2 mins NOT APPLICABLE Chromogen: AS PER KIT Detection: AS PER KIT PMS2 - Method 2 Participant scored 17/2 (UK NEQAS Slide) and 16/2 (In House slide) using this method. Primary Antibody: Ventana (EPR3947) Automation: BioGenex GenoMX 6i Ventana Optiview Main Buffer: Optimax Wash Buffer Pressure Cooker Chromogen: Ventana DAB Detection: Ventana OptiView Kit (76-7) 83

86 Run 16 HNPCC Module PMS2 - Method 3 Participant scored 2/2 (UK NEQAS Slide) and 17/2 (In House slide) using this method. Primary Antibody: Dako RTU FLEX IR87 (EP51), 3 Mins, 21 ºC Prediluted Automation: Leica Bond Max Leica BondMAx Refine KIT Main Buffer: Bond Wash Buffer (AR959) Leica ER2 2 mins Chromogen: Leica Bond Polymer Refine kit (DS98), 21 ºC., Time 1: 1 Mins, Time 2: 5 Mins Detection: Leica Bond Polymer Refine (DS98), 8 Mins, 21 ºC Prediluted PMS2 - Method 4 Participant scored 16/2 (UK NEQAS Slide) and 18/2 (In House slide) using this method. Primary Antibody: BD Bio/Pharmingen (A16-4), 6 Mins, 37 ºC Dilution 1: 1 Automation: Leica Bond Max Leica BondMAx Refine KIT Main Buffer: Bond Wash Buffer (AR959) Leica ER2 3 mins Chromogen: Leica Bond Polymer Refine kit (DS98) Detection: Leica Bond Polymer Refine (DS98) 84

87 The HER2 ISH Module - Interpretive Run 35 Merdol Ibrahim Gene Assessed: Sections Circulated: HER2 Number of Registered Participants: 231 Number of Participants This Run 114 (49%) Composite slide consisting of 4 breast carcinoma sections (see table below) The Summary box below gives details of the HER2 phenotype and genotype of the tissues circulated for this assessment, which comprised of formalin fixed and paraffin processed sections from a composite block of four human breast carcinomas. The sections were positioned on the microscope slides as indicated. Tissue Section HER2 status by IHC HER2 status by FISH A 2+ Non-Amplified B 2+ Amplified C 2+ Non-Amplified D 2+ Equivocal Tissue Section Positioning: Tissue sections were positioned on microscope slides as illustrated below. Slide Label A B C D Introduction After the successful conclusion and publication of the first pilot results with a three year follow-up [1,2], the interpretive part of the ISH scheme is now a full UK NEQAS module. Furthermore, the scheme has also allowed multiple ring studies to be undertaken, which have enabled newer methodologies, for example the Ventana silver ISH (SISH) [3], Kreatech FISH probes [4] and the Ventana Dual ISH (BDISH and DDISH) [5] to be analysed. With the growing importance of colourimetric assays, both chromogenic and silver based in situ detection systems, the scheme changed from a FISH to an ISH based scheme, where all molecular in situ assays for HER2 gene copy/amplification determination are included. This move reflects an on-going development in chromogenic assay systems which is leading to a wider uptake of these assays within the diagnostic community. For UK laboratories carrying out diagnostic HER2 ISH testing participation is therefore mandatory under Clinical Pathology Accreditation (UK), and National Quality Assurance Advisory Panel (NQAAP) regulations. It should also be emphasised that the scheme remains very-much open to non-uk laboratories as an aid to monitoring and improving performance. Recommendations Diagnostic testing recommendations for HER2 ISH was originally published by Walker et at [6] but a more recent publication covering recommendations for both breast and gastric in-situ hybridisation methods is now available [7]. Assessment of Slides (rationale behind the adopted scoring system) The cell lines initially used by UK NEQAS were selected to cover diagnostic intervals for the HER2 FISH test, where gene amplification is recognised if the ratio of HER2 signals to Chromosome 17 signals exceeds 2.. Previous evidence suggests that inter-observer error in scoring FISH signals can be controlled within 1% for both HER2 and other solid tissue analyses (Bartlett et al., 21, Bartlett, 25 for review). This results in a small population of cases (<2.% of cases, unpublished observations; FISH used as frontline test) where the result reported lies within a 1% range of the diagnostic interval, i.e. between 1.8 and 2.2. Whilst this compares favourably with the approximately 15% of cases reported as borderline by immunohistochemistry it highlights the need to rigorously control scoring approaches. Data suggest that allowing variation to increase by 5% (to 15%) would double the number of equivocal cases (to 4%) and allowing an increase to 2% variation at the diagnostic interval would lead to a fourfold increase in equivocal cases (Bartlett unpublished observations). These observations underpin the approach taken in assessing participants results. For each of the tissue sections, participants scored 3 points if the result reported fell within the range observed by the reference laboratories. These reference laboratory results already reflect the acceptable observer variation of 1% for inter-observer variation in scoring discussed above since they UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 85

88 The HER2 ISH Module - Interpretive Run 35 How the current scoring system works Schematic representation of the scoring system; the example illustrated uses the Reference Centre set of HER2/Ch17 ratios obtained for the SK-BR3 (3+ Amplified) cell line at run 4. In this case the lowest ratio obtained by a Reference Centre was 3.19, and the highest 4.1; participants submitting ratios within this range were judged to have achieved an appropriate result (score = 3). The lower cut-off for acceptable ratios (score = 2) was calculated as 3.19 minus 1% of 3.19, that is 2.87; and the upper cut-off was calculated as 4.1 plus 1% of 4.1, that is Participants who submitted ratios out-with these 1% cut-offs were judged to have achieved an inappropriate and received a score of 1. Except in the case of the MDA-MB-453 cell line, misdiagnosis (amplified reported as non-amplified, and visa versa) resulted in a score of. Superscript notation and abbreviation used in figure: * = does not apply to results obtained for MDA-MB-453 (2+ Borderline: Amplified to Non-Amplified) cell line; RC = Reference Centre. are collated from seven different laboratories, each of which presents two scores (up to 14 observations per cell line per assessment). Thus results must fall within this range to be regarded as appropriate. For each tissue section, participants scored 2 points if results lay within a further 1% of the range of results observed by the reference centres. These results reflect a band approximating to the 2% variation. Such results reflect acceptable performance with the potential for improvement. For results, which lay out with this second band of variation one point was given, representing inappropriate performance. For such specimens it is important to note that despite the correct diagnosis being given in many cases, a degree of variation >2% from the accepted result is indicative of a high degree of potential for error in diagnosis. In cases where results were not only deviating to this degree from the accepted result but were also misdiagnosed (e.g. where a non-amplified case is reported as amplified) a score of was recorded in recognition of the negative impact this has on the laboratories likely performance on diagnostic samples. References 1. Bartlett JMS, Ibrahim M, Jasani B, et al. (27) External quality assurance of HER2 FISH testing: Results of a UK NEQAS pilot scheme. J Clin Pathol; 6: Bartlett JM, Ibrahim M, Jasani B, Morgan JM, Ellis I, Kay E, Connolly Y, Campbell F, O'Grady A, Barnett S, Miller K. (29) External quality Assurance of HER2 FISH and ISH testing: three years of the UK national external quality assurance scheme. Am J Clin Pathol; 131: Bartlett JM, Campbell FM, Ibrahim M, Wencyk P, Ellis I, Kay E, Connolly Y, O'Grady A, Di Palma S, Starczynski J, Morgan JM, Jasani B, Miller K. (29) Chromogenic in situ hybridization: a multicenter study comparing silver in situ hybridization with FISH. Am J Clin Pathol; 132: Bartlett JM, Campbell FM, Ibrahim M, Thomas J, Wencyk P, Ellis I, Kay E, Connolly Y, O'Grady A, Barnett S, Starczynski J, Cunningham P, Miller K. (21) A UK NEQAS ICC and ISH multicentre study using the Kreatech Poseidon HER2 FISH probe: intersite variation can be rigorously controlled using FISH. Histopathology; 56: Bartlett JM, Campbell FM, Ibrahim M, O'Grady A, Kay E, Faulkes C, Collins N, Starczynski J, Morgan JM, Jasani B, Miller K. (211) A UK NEQAS ISH multicenter ring study using the Ventana HER2 dual-color ISH assay. Am J Clin Pathol; 135: Walker RA, Bartlett JM, Dowsett M, Ellis IO, Hanby AM, Jasani B, Miller K. Pinder SE. (28) HER2 testing in the UK: further update recommendations. J Clin Pathol; Bartlett JM, Starczynski J, Atkey N, Kay E, O'Grady A, Gandy M, Ibrahi M, Jasani B, Ellis IO, Pinder SE, Walker RA. (211) HER2 testing in the UK: recommendations for breast and gastric in-situ hybridisation methods. J Clin Pathol; 64: UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 86

89 ÑÒÓ ÔÕ ISH Module 8 Ö Ø ÙÚ ÛÜÖÝÞ ßàÛ Ý áââ ãäåæçèçéäêæë Individual 4 = 2 (2%) 32 Ö Ø ÙÚ ÛÜÖÝÞ ßàÛ Ý ì âäíë îêâï Individual 4 = (%) 7 5 = 1 (1%) 6 = (%) 28 5 = (%) 6 = (%) 6 7 = 2 (2%) 8 = 4 (3%) 24 7 = 1 (2%) 8 = 1 (2%) no. of returns = 5 (4%) 1 = 23 (17%) 11 = 2 (15%) 12 = 75 (57%) no. of returns = (%) 1 = 7 (16%) 11 = 5 (11%) 12 = 3 (68%) 2 Summary = 75 (57%) 8 Summary = 3 (68%) = 48 (36%) 4-8 = 9 (7%) = 12 (27%) 4-8 = 2 (5%) Median = Median = 12. Ö Ø ÙÚ ðñæòîó Üôéâîïñó Ý áââ ãäåæçèçéäêæë íï ãñåèñêæäõñ îö ãäåæçèçéäêæë ùøants % of Partic % Pass Rate Copy No.: Other 9 Ratio: Dako Pharm Dx 2 Ratio: Kreatech Probes 5 Ratio: Leica HER2 FISH TA Ratio: Other - CISH Ratio: Other - FISH Ratio: Pathvysion Vysis Kit 7 Ratio: Ventana Dual ISH (BDISH) Ratio: Ventana Inform Dual ISH (DDISH) 3 Ratio: Ventana Inform SISH 1 Ratio: ytovision ytodot 2C 5 Ratio: ytovision ytolight 1 Ö Ø ÙÚ ðñæòîó Üôéâîïñó Ý ì úäíë îêâï íï ãñåèñêæäõñ îö ãäåæçèçéäêæë ùøants % of Partic % Pass Rate Copy No.: Other 5 Ratio: Kreatech Probes 7 Ratio: Leica HER2 FISH TA Ratio: Other - FISH Ratio: Pathvysion Vysis Kit 2 Ratio: Ventana Dual ISH (BDISH) Ratio: Ventana Inform Dual ISH (DDISH) 5 Ratio: Ventana Inform SISH 2 Ratio: ytovision ytolight 1 87

90 ûüý þÿ ISH Module R iants Number of Partic % Pass Rate Copy No.: Other 12 Ratio: Dako Pharm Dx 3 Ratio: Kreatech Probes 7 Ratio: Leica HER2 FISH TA Ratio: Other - CISH 3 Ratio: Other - FISH Ratio: Pathvysion Vysis Kit 9 Ratio: Ventana Dual ISH (BDISH) Ratio: Ventana Inform Dual ISH (DDISH) 4 Ratio: Ventana Inform SISH 1 Ratio: Zytovision ZytoDot 2C 6 Ratio: Zytovision ZytoLight 2 1 R iants 7 7 Number of Partic % Pass Rate Copy No.: Other Ratio: Kreatech Probes Ratio: Leica HER2 FISH TA Ratio: Other - FISH Ratio: Pathvysion Vysis Kit 1 Ratio: Ventana Dual ISH (BDISH) Ratio: Ventana Inform Dual ISH (DDISH) 2 1 Ratio: Ventana Inform SISH Ratio: Zytovision ZytoLight 1 88

91 The HER2 ISH Module - Technical (Pilot) Run 35 Merdol Ibrahim, Suzanne Parry Gene Assessed: HER2 Sections circulated: Number of Registered Participants: 231 Composite slide consisting of 4 breast carcinoma sections (see table below) Number of Participants Taking Part This Run (%) (Chromogenic & Fluorescent) The Summary Box below gives details of the HER2 phenotype and genotype of the tissue sections circulated for this assessment, which comprised of formalin fixed and paraffin processed sections from a composite block of four human breast carcinomas Tissue Section HER2 status by IHC HER2 status by FISH A 2+ Non-Amplified B 2+ Amplified C 2+ Non-Amplified D 2+ Equivocal Tissue Section Positioning: Tissue sections were positioned on microscope slides as illustrated below. Slide Label A B C D Assessment Procedure Results Summary Chromogen ISH (CISH / SISH / BDISH / DDISH etc.) was assessed around a multi-header microscope with each slide being assessed by 4 independent assessors each providing overall scores out of 5 for the slide as a whole (not the individual tissue sections). The scores are added together to give a score out of 2, along with an overall pass rate comment. Fluorescent ISH (FISH) was assessed by a single assessor from one of the UK NEQAS reference centres and given a score out of 2 for the slide as a whole (not the individual tissue section), along with an overall pass rate comment. A summary of the assessment scoring criteria and it s interpretation is shown in the table on the next page Each of the UK NEQAS tissue sections (A-D) (as well as inhouse samples where submitted) are individually assessed for the quality of ISH staining. Assessors do not count the HER2/ Ch17 signals. The accuracy of signal enumeration is assessed in the 'interpretive' section of the HER2 ISH module. Important: If two separate slides are submitted with separate Ch17 and HER 2 copy numbers (e.g. Ventana/Roche SISH), then the assessors regard this as a single result, and as such, combined comments are placed in a single results column. Note that, Kit control slides are not requested for submission for this module. CISH Results Selected images showing the acceptable and unacceptable levels of staining for the different methods are illustrated in figures 1-6. The overall results from the CISH technical assessment were similar to the previous assessment, with 6% of labs achieving an acceptable pass, and a further 23% receiving a borderline result. Likewise, the unacceptable rate remained very similar at 19%. The main reason for unacceptable results was due to weak or no Ch17 signals. Several labs also showed morphology damage due to excessive pre-treatment. The main CISH method of choice is still the Ventana Inform Dual ISH (DDISH) used by 69% of CISH labs. This showed an acceptable pass rate of 63%, with a further 19% receiving a borderline pass. The Ventana Dual ISH (BDISH) method was also popular, and is used by 15% of labs. This showed an acceptable pass rate of 56%, and a further 33% of labs receiving borderline. Of the participants who submitted slides for the technical assessment, 41 labs (66%) also submitted their in-house control slides for assessment, of which 68% achieved an acceptable level of staining and a further 1% received a borderline pass. 22% of labs failed on their in-house slides, which indicates that participants still need to optimise their methodology to allow for more accurate clinical interpretation. It is also important to emphasise that when changing methodology, e.g. from FISH to CISH, the new technique must been fully validated, and staff must undergo extra training to gain expertise with the new UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 89

92 The HER2 ISH Module - Technical (pilot) Run 35 Table below shows a summary of the assessment scoring criteria and interpretation. Scoring Interpretation Acceptable Individual Assessor 4-5/5 or Overall score >13/2 Borderline Individual Assessor 3/5 or Overall Score = 1-12/2 Unacceptable Individual Assessor 1-2/5 or Overall score <9/2 = Score = The NEQAS distributed samples and/or in-house tissue show a good/very good standard of staining and are suitable for interpretive analyses. The NEQAS distributed samples and/or in-house tissue is borderline interpretable with 2 out of the 4 assessors indicating that improvements can be made with respect for example: Excessive / Weak Nuclear staining Level of probe hybridisation Excessive background staining Also see assessor comments on your report The NEQAS distributed samples and/or in-house tissue is unacceptable for interpretation with at least 3 out of the 4 assessors indicating that there are hybridisation issues, which could be due: Excessive / Weak Nuclear staining Poor probe hybridisation Missing Her2 copy no. / CEP 17 Excessive background staining Also see assessor comments on your report Slide Not submitted for assessment method employed. Any laboratory experiencing staining problems with their CISH method should contact the relevant company for further support. FISH Results Images of acceptable and unacceptable levels of staining are illustrated in figures Overall the FISH results also showed similar pass rates for this assessment compared to the previous run (34), with 55% of labs achieving an acceptable pass, and a further 18% receiving a borderline. There was also still a high fail rate of 27%. The main reason for unacceptable results was due to weak or no signals. Again, the Pathvysion Vysis kit was the most popular FISH method, used by 55% of labs for this assessment, and had an acceptable pass rate of 43%, with a further 23% receiving a borderline pass. Results for the labs in-house control material showed a better acceptable pass rate than the NEQAS material at 79%, with a further 15% receiving a borderline score. Only 2 labs (6%) failed on their in house material, which was also lower than the previous assessment (Run 34). Validating ISH The UK NEQAS data indicates that there is a definite move uptake of the Ventana DDISH method and laboratories introducing either this method or another ISH method should make sure that: Staff perform a minimum of 1 ISH evaluations in parallel with an experienced member of staff to ensure scoring consistency before working independently. A minimum concordance of 95% is recommended for both diagnostic results (amplified and non-amplified status) and numerical results (for both HER2 and Ch17). Clinical and UK NEQAS samples should be assessed for quality of hybridisation prior to interpretation. You should make sure that probe signal is easily discernable, normal cells show the expected HER2 and/or Ch17 signal, there is minimal background and leaching of signal Further validation guidelines can be found in: Bartlett et al., (211) HER2 testing in the UK: recommendations for breast and gastric in-situ hybridisation methods. J Clin Pathol.64(8): Recommendations for Returning FISH Slides for NEQAS Assessments a. Section should be mounted using a fluorescence antifade mounting media, which prevents loss/quenching of fluorescence e.g. Vectashield (Vector labs), Dako Fluorescence Mounting Medium (Dako), Fluoromount (SIGMA/Aldrich, VWR). Note: The Abbott Vysis kit already contains DAPI in phenylenediamine dihydrochloride, which is an anti fading reagent, but we have found that some laboratories also used the above mentioned mounting media. b. Seal the coverslip edges onto the slide using clear varnish or another sealing agent, which will stop the slide from drying out. c. Send back FISH slides as soon as you have finished your own interpretation. d. There is no need to send back slides packed in ice/dry ice. Please return in the slide mailer that is provided. UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 9

93 The HER2 ISH Module - Technical (Pilot) Run 35 How to Use this Report to Compare NEQAS Cell Line Technical & Interpretive Results: Once you have your results for both the Interpretive and Technical HER2 ISH modules, use both results to troubleshoot whether there are interpretive and/or technical issues with your ISH methodology. The table below shows a brief troubleshooting guide, which we hope will assistant you further. UK NEQAS ICC & ISH recommends that laboratories should continually monitor their performance to make sure that both technical and interpretive results are frequently audited. Interpretive & Technical Troubleshooting Guide Interpretive Result Appropriate or Acceptable Unacceptable Appropriate or Acceptable Unacceptable Appropriate or Acceptable Unacceptable Technical Result (pre-pilot module) Acceptable Acceptable Borderline Borderline Unacceptable Unacceptable Overall Feedback Acceptable The NEQAS samples show a good standard of staining and have been interpreted correctly Unacceptable The NEQAS samples show a good standard of staining BUT there appears to be issues with interpretation i.e. HER2 copy number and/or Ch17 overestimated/underestimated. Recommend that scoring/counting criteria is reviewed Borderline Acceptable The NEQAS samples are of borderline acceptability for staining quality. Recommend that technical method (kit/assay) is further optimised. Unacceptable The technical staining can be improved as this may be effecting interpretation. Recommend that technical method (kit/assay) is optimised (or where relevant a standardised kit is used as per instructions) as interpretation may be effected by the borderline quality of staining. Unacceptable The NEQAS samples are unacceptable for interpretation and caution should be taken when interpreting from these slides. Recommend that technical method (kit/assay) is optimised (or where relevant a standardised kit is used as per instructions) as interpretation, although appears correct, may lead to incorrect interpretation of clinical cases Unacceptable The NEQAS samples are unacceptable for technical staining and interpretation. Reporting from such cases is very likely to lead to incorrect interpretation of clinical cases. If there is persistent underperformance: seek assistance from kit/assay manufacturer seek assistance from UK NEQAS or colleagues re-validate protocol (retrospectively and prospectively) review scoring criteria send clinical cases to a reference centre to confirm your results UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent 91

94 !" BREAST HER2 ISH Module Selected Images showing Optimal and Sub-optimal Immunostaining Fig 1. Acceptable Ventana DDISH in the UK NEQAS samples. (A) non-amplified tissue sample A and (B) amplified tissue sample B. In both cases distinct HER2 signals (black) and Ch17 signals (red) are clearly demonstrated with the expected level of copies per cell. Fig 2. Acceptable Ventana DDISH in the UK NEQAS samples: (A) shows the non-amplified tissue sample C and (B) equivocal tissue sample D. In both examples the HER2 (black) and Ch17 (red) signals are strong and clear and show the expected level of copies per nuclei. Fig 3. Two examples of slides that received a borderline pass; UK NEQAS amplified tissue sample B. In (A) the Ch17 signals are weak (but readable). (B) morphology damage, most likely due to over-digestion. Both slides stained with Ventana DDISH. Fig 4. Two examples of slides with unacceptable level of staining. (A), NEQAS sample C, shows very weak or no HER2 signals, and (B), NEQAS sample D, has no HER2 signals. Both sections stained with Ventana DDISH. Fig 5. Two examples of suboptimal level of staining. (A) NEQAS sample C, shows leaching of the Cep17 probe. (B) Participants in-house sample, which shows excessive silver deposit and signals outside of the nuclei, which makes the slide difficult to interpret. Fig 6. Good example of an in house composite control consisting of (A) non-amplified and (B) amplified tissue samples. Both cases show strong, clear and distinct HER2 signals (black) and Ch17 signals (red). 92

95 #$% &' BREAST HER2 ISH Module Selected Images showing Optimal and Sub-optimal Immunostaining Fig 7. Acceptable FISH examples from UK NEQAS distributed samples (A) sample A - non-amplified and (B) sample B, amplified. Both samples shown without dapi, but clearly show distinct HER2 signals (red) and Ch17 signals (green). Stained using (A) Dako pharm DX and (B) Abbott Pathvysion Vysis. Fig 8. Acceptable FISH examples from UK NEQAS distributed samples stained using the Zytovision ZytoLight (A) sample C, non-amplified and (B) sample D, equivocal. Note that this method stains the HER2 signals in green and the Ch17 signals in red. Fig 9. Unacceptable FISH from the UK NEQAS distributed sample B amplified case, showing single images for (A) HER2 and (B) Ch17. (A) HER2 signals are clearly demonstrated and readable but (B) Ch17 signals are uniterpretable. This is probably due to insufficient digestion of the sample. Stained using the Pathvysion Vysis kit. Fig 1. Unacceptable FISH from the UK NEQAS distributed sample B amplified case, showing single images for (A) HER2 and (B) Ch17. There is a lack of signal for both (A) HER2 or (B) Ch17. This is probably due to insufficient digestion of the sample. Stained using the Pathvysion Vysis kit. Fig 11. Excellent in-house example of an amplified case showing very clear HER2 (red) and Ch17 (green) signals. This example also illustrates the stability of fluorescence signal that can be obtained even after samples have been transported for external quality assessment. Stained using the Pathvysion Vysis kit. Fig 12. Unacceptable FISH, with non-specific background staining. We have noticed that this is not necessarily a methodological issue but relates to the slides that are used, in this case the lab used the Leica Xtra slides, which we do not recommend for HER2 FISH. 93

96 Run 35 Breast HER2 ISH Technical Module Chromogen ISH: Pass Rates and Methods Overall Pass Rates NEQAS Slide (n=62) In-house slide (n=41) % % % 18% % 22% Acceptable (n=37) Borderline (n=14) Unacceptable (n=11) Acceptable (n=28) Borderline (n=4) Unacceptable (n=9) 1% 9% 8% 7% 2 NEQAS Slides: Methods Used and Percentage Pass Rates % 5% 4% 3% 2% Unacceptable Borderline Acceptable 1% % Not Stated (n=5) Other - CISH (n=1) Ventana Dual ISH (BDISH) (n=9) Ventana Inform Dual ISH (DDISH) (n=42) Ventana Inform SISH (n=4) Zytovision ZytoDot 2C (n=1) 1% 9% In-house Slides: Methods Used and Percentage Pass Rates % 7% % 5% 4% 3% 2% Unacceptable Borderline Acceptable 1% % Other - CISH (n=1) Ventana Dual ISH (BDISH) (n=6) Ventana Inform Dual ISH (DDISH) (n=31) Ventana Inform SISH (n=2) Zytovision ZytoDot 2C (n=1) 94

97 Run 35 Breast HER2 ISH Technical Module Fluorscent ISH: Pass Rates and Methods FISH: Overall Pass Rates NEQAS slide (n=55) In-house slide (n=34) % Acceptable (n=3) 18% Borderline (n=1) 27% Unacceptable (n=15) % Acceptable (n=27) 15% Borderline (n=5) 6% Unacceptable (n=2) 1% 9% 8% 7% 6% 5% 4% 3% 2% NEQAS Slides: Methods Used and Percentage Pass Rates Unacceptable Borderline Acceptable 1% % Dako Pharm Dx (n=7) Kreatech Probes (n=2) Leica HER2 FISH TA9217 (n=7) Not Stated (n=2) Other - FISH (n=1) Pathvysion Vysis Kit (n=3) Zytovision ZytoLight (n=6) 1% 9% 8% 7% 5 In-house Slides: Methods Used and Percentage Pass Rates % 5% 4% Unacceptable Borderline Acceptable 3% 2% 5 1% % Dako Pharm Dx (n=4) Kreatech Probes (n=1) Leica HER2 FISH TA9217 (n=2) Pathvysion Vysis Kit (n=23) Zytovision ZytoLight (n=4) 95

98 UK NEQAS ICC & ISH Improving Immunocytochemistry for over 25 Years For advertising opportunities contact: Neil Bilbe: UK NEQAS ICC & ISH. No part of this document can be copied or used without prior written consent