Breast Specimen Processing and Reporting With an Emphasis on Margin Evaluation. A College of American Pathologists Survey of 866 Laboratories

Transcription

1 Breast Specimen Processing and Reporting With an Emphasis on Margin Evaluation A College of American Pathologists Survey of 866 Laboratories Anthony J. Guidi, MD; Joseph A. Tworek, MD; Daniel D. Mais, MD; Rhona J. Souers, MS; Barbara J. Blond, MT(ASCP), MBA; Richard W. Brown, MD Context. The College of American Pathologists (CAP) developed protocols for reporting pathologic characteristics of breast cancer specimens, including margin status. The Society of Surgical Oncology (SSO) and the American Society for Radiation Oncology (ASTRO) published treatment guidelines regarding margins in patients with invasive cancer; and SSO, ASTRO, and the American Society of Clinical Oncology (ASCO) recently published guidelines for patients with ductal carcinoma in situ. Objective. To assess current practices among pathologists with regard to the processing/reporting of breast specimens, assess compliance with CAP cancer protocols, and assess alignment with SSO/ASTRO and SSO/ASTRO/ ASCO guidelines. Design. A survey concerning breast specimen processing/reporting was distributed to pathologists enrolled in the CAP Performance Improvement Program in Surgical Pathology. Results. Ninety-four percent (716 of 764 respondents) and 91% (699 of 769 respondents) define positive margins as tumor on ink for invasive cancer and ductal carcinoma in situ, respectively, in compliance with CAP cancer protocols and with SSO/ASTRO and SSO/ASTRO/ ASCO guidelines. Of 791 respondents who provided details regarding methods for margin evaluation, 608 (77%) exclusively examine perpendicular margins, facilitating guideline compliance. However, 183 of 791 respondents (23%) examine en face margins in at least a subset of specimens, which may preclude guideline compliance in some cases. When separate cavity (shave) margins are examined, while 517 of 586 respondents (88%) ink these specimens, 69 of 586 (12%) do not, and this may also preclude guideline compliance in some cases. Conclusions. A substantial proportion of survey participants report margin status for breast cancer specimens in a manner consistent with CAP cancer protocols, and in alignment with SSO/ASTRO and SSO/ASTRO/ASCO guidelines. However, there are opportunities for some laboratories to modify procedures in order to facilitate more complete adherence to guidelines. (Arch Pathol Lab Med. 2018;142: ; doi: / arpa cp) The College of American Pathologists (CAP) convened multidisciplinary consensus panels including physicians with expertise in breast cancer, who subsequently developed comprehensive protocols regarding the examination and Accepted for publication July 11, Published as an Early Online Release January 12, From the Department of Pathology, Newton-Wellesley Hospital, Newton, Massachusetts (Dr Guidi); the Department of Pathology, St. Joseph Mercy Hospital, Ann Arbor, Michigan (Dr Tworek); the Department of Pathology, University of Texas Health Sciences Center University Hospital, San Antonio (Dr Mais); Biostatistics (Ms Souers) and Surveys Cytopathology (Ms Blond), College of American Pathologists, Northfield, Illinois; and the Department of Pathology, Memorial Hermann Southwest Hospital, Houston, Texas (Dr Brown). All authors of this study are members of the College of American Pathologists Quality Practices Committee, and this manuscript is a product of our committee. No financial support other than routine committee resources were required to complete this study and manuscript. The authors have no other relevant financial interest in the products or companies described in this article. Corresponding author: Anthony J. Guidi, MD, Department of Pathology, Newton-Wellesley Hospital, 2014 Washington St, Newton, MA ( aguidi@partners.org). reporting of surgical specimens from patients with invasive breast cancer (IC) 1 and ductal carcinoma in situ (DCIS). 2 The protocols were developed to assist laboratories in providing cancer reports with clinically relevant information in order to inform optimal treatment decisions. The CAP breast cancer protocols include both required and nonrequired data elements, and state that the latter may be clinically relevant, but not yet validated or regularly used in patient management. Of the many data elements detailed in both protocols, one required element with major treatment implications for patients is margin status. For both IC and DCIS, the CAP breast cancer protocols require laboratories to report margins as positive or uninvolved (negative) whenever possible, and define a positive margin as ink on tumor, characterized by a distance between the leading edge of tumor and ink of 0 mm. For uninvolved/negative margins, the protocols require that the distance between the leading edge of the tumor and the closest inked margin be provided, as well as the specific location of the closest margin, when possible. Nonrequired data elements with regard to specimen margins include (1) the extent of positive margin involvement (eg, focal, minimal/ 496 Arch Pathol Lab Med Vol 142, April 2018 Breast Specimen Processing and Reporting Guidi et al

2 moderate, or extensive), and (2) the distance between tumor and ink for negative margins in addition to the closest margin. With regard to specimen processing and margin assessment, the explanatory notes in the protocols state that specimens should be processed in a manner that allows identification of specific margins (eg, anterior, inferior, lateral, posterior, superior, medial) and also stipulate that all relevant margins should be evaluated grossly and histologically in a manner that allows measurement of the closest margin, when possible. This may involve assessing margins submitted by the surgeon on a single excision specimen, and/or additional separately submitted cavity (shave) margins. 1,2 The CAP breast cancer protocols detail many factors in addition to margin status that impact patient management. First, with regard to the extent of tissue sampling, the CAP protocol for IC states: If the specimen consists predominantly of DCIS with microinvasion, complete submission of the entire specimen, or at a minimum the entire grossly involved area, is recommended to identify additional areas of invasion and/or lymph-vascular invasion. 1 The CAP protocol for DCIS states: When practical, the entire specimen should be submitted in a sequential fashion for histologic examination. If this is not possible, at least the entire region of the targeted lesion should be examined microscopically. If DCIS, lobular carcinoma in situ, or atypical hyperplasia is identified, all fibrous tissue should be examined. 2 Further, the DCIS protocol states: For specimens with a known diagnosis of DCIS (eg, by prior core needle biopsy) it is highly recommended that the entire specimen is examined using serial sequential sampling to exclude the possibility of invasion, to completely evaluate the margins, and to aid in determining extent. 2 Second, in the DCIS protocol there is a requirement to provide an assessment of the size/extent of DCIS, which can be determined by a number of methods detailed in the protocol. The extent of DCIS in excision specimens is clinically relevant because it correlates with the presence of residual disease after re-excision, 3 6 and the risk of ipsilateral breast tumor recurrence (IBTR) In addition, larger DCIS lesions are also associated with a higher likelihood of finding areas of invasion. 11,12 In the CAP protocol for IC, the extent of associated DCIS may be provided, but is not required. Third, both protocols state that if there has been a prior biopsy (core or excision), the biopsy site should be sampled and documented in the report. 1,2 In addition to CAP, others have developed consensus guidelines that have implications with regard to the processing and reporting of breast cancer margins. In 2014, the Society of Surgical Oncology (SSO) and the American Society for Radiation Oncology (ASTRO) published multidisciplinary consensus guidelines regarding margin status in patients with IC. 13 The guidelines were subsequently endorsed by the American Society of Clinical Oncology (ASCO), 14 the 2015 St. Gallen Conference, 15 and the American Society of Breast Surgeons, 16 and have also been incorporated into the National Comprehensive Cancer Network (NCCN) guidelines. 17 Subsequently, the SSO/ ASTRO/ASCO published multidisciplinary consensus guidelines regarding margin status in patients with DCIS. 18 The consensus panels used as their primary evidence base metaanalyses of margin width and IBTR from numerous published clinical outcome studies. Similar to the CAP breast cancer protocols, the SSO/ASTRO IC guideline and the SSO/ ASTRO/ASCO DCIS guideline define positive margins as tumor on ink. In patients with IC, margins were defined as positive if either IC or DCIS extended to ink, and the panel concluded that positive margins were associated with a 2-fold increase in the risk of IBTR as compared to negative margins. This increased risk was not mitigated by favorable biology, endocrine therapy, or administration of a radiation boost. Moreover, obtaining more widely negative margins than no ink on tumor did not significantly decrease the rate of IBTR. The panel concluded that the use of no tumor on ink should be the standard for an adequate margin in patients with IC. 13 In patients with DCIS unassociated with IC, the consensus panel reported that negative margins decreased the risk of IBTR by 50% as compared to positive margins. In contrast to patients with IC, however, patients with pure DCIS experienced a significantly reduced risk of IBTR with wider negative margins a 2-mm negative margin significantly reduced the risk of IBTR, compared to margins less than 2 mm. However, achieving margins greater than 2 mm did not significantly reduce the risk of IBTR, compared to 2- mm margins. The panel concluded that 2-mm margins should be the standard for adequate margins in patients with pure DCIS treated with breast conserving therapy (BCT). 18 Given the very recent publication of the SSO/ASTRO/ ASCO consensus guideline for patients with DCIS, it is too early to gauge its impact on clinical practices. However, multiple retrospective and prospective reports suggest a major impact resulting from the adoption of the SSO/ ASTRO consensus guideline for patients with IC, including (1) a reduction in the number of unnecessary re-excisions, (2) a reduction in the number of mastectomies performed on women who are appropriate candidates for BCT, and (3) a significant decrease in health care costs In view of the evidence-based nature of the SSO/ASTRO and SSO/ ASTRO/ASCO consensus guidelines, and given the impact of the SSO/ASTRO guidelines with regard to more costeffective patient care, it is critically important that pathology laboratories process and report breast cancer specimens in a manner that facilitates our clinical colleagues ability to follow the guidelines. The purpose of this study was to use a survey tool to assess current practices among pathologists with regard to the processing of breast cancer specimens as well as the reporting of selected elements detailed in the CAP breast cancer protocols. The survey was primarily focused on the elements of breast cancer processing and reporting with the greatest impact on evaluating the risk of IBTR in patients treated with BCT, with a particular emphasis on specimen margins. Moreover, we assess the level of compliance among participants in the survey with CAP breast cancer protocols, as well as alignment of their practices with the recent SSO/ ASTRO and SSO/ASTRO/ASCO consensus guidelines. MATERIALS AND METHODS In March 2016, a survey was distributed with the CAP Performance Improvement Program in Surgical Pathology slide set (PIP-A) to laboratories enrolled in this program, and recipients were instructed to forward the survey to the medical director of surgical pathology at their institution, or the pathologist in the department most knowledgeable about breast specimen processing and reporting. The survey included 35 questions, of which 9 (26%) were associated with follow-up questions depending on the responses to initial questions. Most questions had predetermined, multiple-choice answers (single or multiselect); 9 questions (26%) allowed entry of numerical answers. Two questions (6%) addressed demographic features of the participating laboratories. Three questions (9%) addressed the intraoperative assessment of margins. Five questions (14%) addressed specific inking protocols Arch Pathol Lab Med Vol 142, April 2018 Breast Specimen Processing and Reporting Guidi et al 497

3 Table 1. Demographics of Survey Participants Institution Type n ¼ 781 Nonprofit hospital 291 (37) For-profit hospital 136 (17) Municipal or state hospital 103 (13) University hospital/academic medical center 83 (11) Regional/local reference laboratory 74 (9) National or corporate reference laboratory 29 (4) Veteran s Administration hospital 29 (4) Clinic, group, or physician office laboratory 20 (3) Active duty military hospital 15 (2) Other 1 (,1) for breast excision specimens. Thirteen questions (37%) addressed additional details regarding breast specimen grossing methodology. Twelve questions (34%) addressed elements included in written reports. Chi-square tests were used to compare compliance among survey participant subgroups. The survey results were summarized and statistical tests performed by using Statistical Analysis Software (SAS), version 9.3 (SAS Institute, Cary, North Carolina). RESULTS Of the 2242 surveys sent, 866 (39%) were completed and returned to the CAP. Not all institutions provided responses to all questions. The demographic characteristics of the 781 respondents who provided this information are listed in Table 1; 657 (73%) were hospital laboratories, and of these, 83 (11%) were academic medical centers. Of 866 respondents, 750 (87%) were located in the United States. Among the 640 respondents who provided 2015 volumes for breast cancer resection specimens, the median yearly volume was 100 cases (range, cases; 5th 95th percentiles). Preliminary/Intraoperative Margin Assessment Survey results with regard to the intraoperative margin assessment of breast specimens are summarized in Table 2. Of 797 respondents, 532 (67%) reported that they do not assess breast specimen margins intraoperatively. Of the 265 respondents (33%) who do assess margins intraoperatively, 172 (65%) assess margins for IC and DCIS, 86 (32%) assess margins for IC only, and 7 (3%) assess margins for DCIS only. Of the 264 respondents who provided information regarding the relative frequency of performing intraoperative examinations, 101 (38%) reported performing evaluations in more than half of cases; 88 (33%) in 10% to 50% of cases; and Table 2. Intraoperative Margin Assessment Intraoperative assessment performed n ¼ 797 Yes 265 (33) No 532 (67) Diagnoses for which intraoperative n ¼ 265 assessment performed IC and DCIS 172 (65) IC only 86 (32) DCIS only 7 (3) Proportion of cases in which intraoperative n ¼ 264 assessment performed.50% 101 (38) 10% 50% 88 (33),10% 75 (28) Abbreviations: DCIS, ductal carcinoma in situ; IC, invasive carcinoma. Table 3. Inking Protocols: Personnel Specimen ink application n ¼ 840 Laboratory personnel in all cases 600 (71) Surgeons in at least a subset of cases 240 (29) Proportion of cases in which surgeons apply ink n ¼ % 68 (28) 10% 50% 57 (24),10% 115 (48) 75 (28%) in fewer than 10% of cases. Of the 259 respondents who provided sufficient detail regarding the type of intraoperative margin assessment used, 171 (66%) reported performing gross examinations only, 73 (28%) perform frozen sections only, 3 (1%) perform imprint cytology only, and 12 (5%) perform a combination of various methods. Final Margin Assessment: Tissue Processing Considerations Survey results with regard to personnel who ink breast specimens are summarized in Table 3. Of 840 respondents, 600 (71%) reported that only laboratory personnel perform this function, while the remaining 240 respondents (29%) reported that surgeons perform this function in at least a subset of cases. Of the 240 institutions in which surgeons ink at least a subset of specimens, fewer than 10% of specimens are inked by surgeons in 115 (48%), 10% to 50% in 57 (24%), and more than half of specimens in 68 (28%). Of 846 respondents, 771 (91%) reported inking breast excision specimens by using multicolored inks, while 75 (9%) reported that they did not use multicolored inks. Only 384 of 758 respondents (51%) reported having a written procedure regarding the inking of oriented breast specimens, while 374 (49%) did not have a written procedure. Of 379 respondents, 296 (78%) reported that they detail a standardized inking schema (eg, anterior ¼ green, inferior ¼ blue, lateral ¼ orange, posterior¼ black, superior¼ red, medial¼yellow) in a written procedure, while 83 (22%) reported not standardizing an inking schema in a written procedure. Table 4 summarizes survey results with regard to surgical techniques and related processing issues used for margin sampling in patients undergoing BCT. Of 799 respondents, 173 (22%) reported that surgeons in their institution submit Table 4. Utilization and Processing of Cavity (Shave) Margins Surgical techniques n ¼ 799 One specimen only 173 (22) Cavity (shave) margins in at least 626 (78) a subset of cases Proportion of cases using cavity (shave) n ¼ 626 margins.50% 70 (11) 10% 50% 246 (39),10% 310 (50) Cavity (shave) margins inked n ¼ 586 Yes 517 (88) No 69 (12) Cavity (shave) margins entirely submitted n ¼ 581 for histologic evaluation Yes 424 (73) No 157 (27) 498 Arch Pathol Lab Med Vol 142, April 2018 Breast Specimen Processing and Reporting Guidi et al

4 Table 5. Methods of Margin Evaluation Perpendicular versus en face (parallel) n ¼ 791 margins used Perpendicular margins in all cases 608 (77) En face margins in all cases 64 (8) Both perpendicular and en face margins 119 (15) a single excision specimen for margin evaluation in all cases. The remaining 626 respondents (78%) reported that surgeons submit separate cavity (shave) margins in at least a subset of cases of these, 310 (50%) examine additional cavity margins in fewer than 10% of cases, 246 (39%) in 10% to 50% of cases, and 70 (11%) in more than half of cases. When separate cavity margins are submitted, 517 of 586 respondents (88%) reported that they ink the new margins to assess final margin status, while the remaining 69 (12%) reported that they do not ink separate cavity (shave) margins. With regard to how separately submitted cavity margins are processed, 424 of 581 respondents (73%) reported that they submit the margins for histologic examination in their entirety, while 157 (27%) perform a more limited sampling. We also queried institutions with regard to the formal techniques used to sample margins in the laboratory, and the results are summarized in Table 5. Of 791 respondents, 608 (77%) reported exclusively examining perpendicular margins, 64 (8%) exclusively examine en face (parallel) margins, and 119 (15%) use a combination of both techniques. Therefore, 183 respondents (23%) examine en face margins in at least a subset of cases. Respondents who perform both margin sampling techniques frequently noted that they examine perpendicular margins when grossly evident tumor is in close proximity to margins, and en face margins when tumor is grossly distant from margins. Table 6 summarizes how institutions sample excision specimen margins in cases with a grossly identifiable mass that was previously diagnosed by core biopsy as IC or DCIS. Of 805 respondents, 571 (71%) reported submitting all tissue from margins, or at least sampling all margins in these circumstances. The remaining 234 respondents (29%) reported sampling a subset of margins in these circumstances, with 124 of these (15% of the total) sampling the closest margin, and 101 (13% of the total) sampling all margins within a certain specified distance of the mass. The remaining 9 respondents (1% of the total) did not provide details regarding their margin sampling strategy. Only 234 Table 6. Extent of Margin Sampling in Cancer Excisions With Grossly Evident Mass Extent of margin sampling n ¼ 805 All margins either submitted entirely 571 (71) or sampled Subset of margins sampled 234 (29) Excisions in which a subset of margins n ¼ 234 sampled Only the closest margins sampled 124 (53) All margins within a specified distance 101 (43) sampled Details not provided 9 (4) Table 7. Characterization of Margin Status in Pathology Reports IC with or without DCIS n ¼ 764 Positive margin ¼ tumor on ink 716 (94) Positive margins include tumor close to ink 48 (6) DCIS without IC n ¼ 769 Positive margin ¼ tumor on ink 699 (91) Positive margins include tumor close to ink 70 (9) Abbreviations: DCIS, ductal carcinoma in situ; IC, invasive carcinoma. of 802 respondents (29%) reported that they specify a margin sampling strategy in a formal written procedure. Final Margin Interpretation and Reporting Table 7 summarizes how institutions characterize margin status in pathology reports. With regard to reporting margins in specimens with IC with or without associated DCIS, 716 of 764 respondents (94%) reported that they define positive margins as tumor on ink. However, 48 (6%) reported that they also characterize cases with tumor close to ink as positive. Of these, 43 provided further details: 27 (63%) characterize tumor 1 mm or less from ink as positive, 10 (23%) categorize tumor 2 mm or less as positive, and 6 (14%) even include tumor greater than 2 mm as positive. Similarly, for specimens with pure DCIS, 699 of 769 respondents (91%) reported that they define positive margins as tumor on ink. However, 70 (9%) reported that they also characterize cases with tumor close to ink as positive. Of these, 58 provided further details: 37 (64%) characterize DCIS 1 mm or less from ink as positive, 16 (27%) categorize DCIS 2 mm or less as positive, and 5 (9%) even include tumor greater than 2 mm as positive. With regard to reporting negative margins in cases of pure DCIS, 770 of 786 respondents (98%) reported providing the distance between the leading edge of tumor and ink in reports, while 16 (2%) do not provide this information. Of the respondents who provide the measurement in specimen reports, 445 of 770 (58%) report only the distance between the leading edge of the tumor and the nearest inked margin, 164 (21%) also report margins within a specified distance to the margin, 151 (20%) report distances to all margins, and 10 (1%) did not provide details regarding their procedure. We also queried institutions with regard to which specific lesions they report margin status for, and the results are summarized in Table 8: 778 of 786 respondents (99%) report margins for IC, 783 of 792 (99%) report margins for DCIS, 455 of 723 (63%) report margins for classical lobular carcinoma in situ (LCIS), 635 of 723 (88%) report margins for pleomorphic LCIS, 423 of 723 (59%) report margins for Table 8. Diagnoses for Which Margins Are Reported IC 778/786 (99) DCIS 783/792 (99) LCIS, classical type 455/723 (63) LCIS, pleomorphic type 635/723 (88) LCIS, other variant types 423/723 (59) ADH 302/723 (42) ADH bordering on DCIS 487/723 (67) Abbreviations: ADH, atypical ductal hyperplasia; DCIS, ductal carcinoma in situ; IC, invasive carcinoma; LCIS, lobular carcinoma in situ. Arch Pathol Lab Med Vol 142, April 2018 Breast Specimen Processing and Reporting Guidi et al 499

5 Table 9. Extent of Breast Specimen Sampling Wire localization specimen without mass, previously diagnosed with DCIS by core biopsy n ¼ 804 Submit entire area near localizing wire, biopsy clip, or targeted abnormality 434 (54) Submit entire specimen, regardless of specimen size 232 (29) Submit all grossly identifiable fibroglandular tissue 73 (9) Submit to a maximum number of blocks 34 (4) Other 31 (4) Specimen with identifiable mass, previously diagnosed by core biopsy with IC or DCIS n ¼ 805 Submit entire mass 362 (45) Submit 1 section per centimeter 226 (28) Submit up to a maximum number of blocks 73 (9) Other 144 (18) Abbreviations: DCIS, ductal carcinoma in situ; IC, invasive carcinoma. other types of variant LCIS, 302 of 723 (42%) report margins for atypical ductal hyperplasia (ADH), and 487 of 723 (67%) report margins for ADH bordering on DCIS. The final margin-related topic we addressed in this survey concerned tissue artifacts. Cautery effect (thermal injury from a surgical bovie device) and specimen fragmentation may render the assessment of specimen margins challenging. We asked participants to consider a hypothetical scenario that occurs not uncommonly in practice: if ink has run down along a crack/cleft of the adipose tissue deep into the specimen and is touching DCIS or IC, how do you categorize the margin? Of 784 respondents, 487 (62%) reported that they would categorize the margin as negative owing to iatrogenic ink spillage, 231 (30%) would categorize the margin as positive if cautery effect was also present within the cleft, 47 (6%) would categorize the margin as indeterminate, and 19 (2%) would categorize the margin as positive. Other Specimen Processing and Reporting Factors Institutions differ with regard to personnel who primarily perform the gross examination of breast excision specimens. Of 836 institutions that provided sufficient detail, pathologists primarily perform this function in 413 (49%), pathologists assistants in 348 (42%), pathology residents in 68 (8%), and 7 (1%) report that grossing duties are equally split between pathologists assistants and residents, or between pathologists and pathologists assistants. We also queried institutions regarding the extent of specimen sampling performed in various circumstances, and the responses are summarized in Table 9. The first section of this table summarizes the sampling of breast excision specimens in which DCIS was previously diagnosed on core biopsy, with no grossly identifiable mass. Only 232 of 804 respondents (29%) reported submitting all tissue for histologic evaluation in these circumstances, regardless of specimen size. The remaining 572 (71%), report performing a more limited sampling, with the largest proportion of these (434, 76%) submitting the entire area near the localization wire, biopsy clip, or targeted abnormality. Of note, only 73 respondents who perform a limited sampling (13%) reported submitting all grossly identifiable fibroglandular tissue. Only 222 of 808 respondents (28%) reported that they detail their tissue sampling process in a formal written procedure. Table 9 also summarizes survey results regarding the sampling of breast excision specimens in which IC and/or DCIS was previously diagnosed on core biopsy, and a grossly identifiable mass lesion is present. Of 805 respondents, 362 (45%) reported that they submit the entire mass for histologic evaluation in these circumstances. The remaining 443 (55%) reported that they perform a more limited sampling of the mass, with the largest proportion of these (226, 51%) submitting 1 section per centimeter of mass. Only 226 of 807 respondents (28%) reported that they detail their tissue sampling process in a formal written procedure. Survey participants were also asked whether and how they define the extent of DCIS in cases without IC. Of 786 respondents, 748 (95%) reported providing an estimate of the extent of DCIS, while 38 (5%) provide no estimate. Institutions that report the extent of DCIS use multiple methods, summarized in Table 10. Of note, when DCIS is Table 10. Extent of Ductal Carcinoma In Situ (DCIS) in Cases Without Invasive Carcinoma Methods used to define the extent of DCIS (multiple responses allowed) n ¼ 748 Direct microscopic measurement from the slide, when limited to 1 slide 457 (61) No. of slides with DCIS as a proportion of total slides 354 (47) Use of gross assessment, when DCIS is limited to a grossly identifiable lesion 324 (43) If DCIS on multiple slides, multiply the number of blocks involved by DCIS by a factor estimating the 286 (38) thickness of the tissue in the block If DCIS on multiple slides, use a 3-dimensional gross microscopic mapping strategy to estimate the total span 135 (18) Other 36 (5) Reconciling the extent of DCIS when multiple methods are used n ¼ 657 Favor the largest size by any measurement 213 (32) Favor direct measurement from the slide 209 (32) Favor the number of blocks involved multiplied by a factor 85 (13) Provide a range based on the various methods 72 (11) Favor estimate by 3-dimensional mapping strategy 58 (9) Other 20 (3) 500 Arch Pathol Lab Med Vol 142, April 2018 Breast Specimen Processing and Reporting Guidi et al

6 Table 11. Documentation of Marker Clips and Biopsy Site Changes Marker clip documentation in pathology reports n ¼ 802 Yes 703 (88) No 99 (12) Location of marker clip documentation in n ¼ 703 report Gross description 587 (83) Gross description and diagnostic portion 104 (15) of report Diagnostic portion of report only 12 (2) Documentation of biopsy site changes n ¼ 801 in pathology reports Yes 769 (96) No 32 (4) Location of biopsy site changes n ¼ 769 documentation in report Gross description 130 (17) Gross description and diagnostic 480 (62) portion of report Diagnostic portion of report only 159 (21) limited to a single slide, and 2 widely spaced foci of DCIS are present, 479 of 737 respondents (65%) reportedly consider each focus separately and provide independent measurements, 219 (30%) include the distance between both foci in the size determination, and 39 (5%) use an alternative method. When multiple methods for defining the extent of DCIS are used in a case, differences in the size determinations using alternative methods are reconciled in different ways, and these responses are also summarized in Table 10. Of note, 213 of 657 respondents (32%) reported favoring the largest size estimate in these circumstances. Only 134 of 731 respondents (18%) reported that they address the quantitation of the extent of DCIS in a formal written procedure. When breast core biopsies are performed, a marker clip is typically placed in the breast to help localize the subsequent excision, if required. Although not specifically addressed in the CAP breast cancer protocols, many institutions document the presence of marker clips in subsequent excision specimens and mastectomies. Survey results with regard to the marker clip documentation are summarized in Table 11. Of 802 respondents, 703 (88%) reported that they document the presence or absence of marker clips in specimen reports, while 99 of 802 (12%) do not provide this documentation. Of the 703 institutions that document specimen clips, 587 (83%) reported that they provide this information in the gross description, 104 (15%) in the gross description and the diagnostic portion of the report, and 12 (2%) in the diagnostic portion of the report only. Only 168 of 700 respondents (24%) reported that they address the documentation of biopsy clips in a formal written procedure. Many institutions also comment on the presence or absence of biopsy site changes in excision specimens and mastectomies, following core biopsy or excision. Survey results with regard to the documentation of biopsy site changes are also summarized in Table 11. Of 801 respondents, 769 (96%) reported that they document the presence or absence of biopsy site changes in the specimen report, while 32 (4%) do not provide this documentation. Of the 769 respondents who provide this information, 480 (62%) reported that they do so in both the gross description and diagnostic portion of the report; 159 (21%) in the diagnostic portion of the report only; and 130 (17%) in the gross description only. Only 159 of 760 respondents (21%) reported that they address the documentation of biopsy site changes in a formal written procedure. Compliance With Margin Guidelines: Breakdown by Institution Type We also examined the relationships among institution type and compliance with the major components of the CAP, SSO/ASTRO, and SSO/ASTRO/ASCO margin guidelines, and the results of these comparisons are provided in Table 12. For the purposes of statistical analysis, institutions were stratified into academic institutions, for-profit institutions, non-profit institutions, and other categories (listed in Table 1). No statistically significant differences among these groups were identified with regard to the definition of positive margins for IC or DCIS, the evaluation of perpendicular versus en face margins, or the evaluation of inked margins when surgeons submit separate cavity (shave) margins. DISCUSSION This study examines current pathology practices with regard to processing and reporting breast cancer specimens. Table 12. Compliance With Margin Guidelines: Breakdown by Institution Type Academic, For-Profit, Nonprofit, Other, IC with or without DCIS n ¼ 83 n ¼ 131 n ¼ 287 n ¼ Positive margin ¼ tumor on ink 81 (98) 120 (92) 272 (95) 233 (92) Positive margins include tumor close to ink 2 (2) 11 (8) 15 (5) 20 (8) DCIS without IC n ¼ 82 n ¼ 129 n ¼ 279 n ¼ Positive margin ¼ tumor on ink 76 (93) 118 (91) 260 (93) 227 (88) Positive margins include tumor close to ink 6 (7) 11 (9) 19 (7) 30 (12) Perpendicular versus en face margins n ¼ 81 n ¼ 134 n ¼ 281 n ¼ Perpendicular margins in all cases 57 (70) 104 (78) 221 (79) 199 (77) En face margins examined in at least a subset of cases 24 (30) 30 (22) 60 (21) 61 (23) Margins inked when cavity (shave) margins used n ¼ 66 n ¼ 100 n ¼ 225 n ¼ Yes 60 (91) 84 (84) 200 (89) 154 (89) No 6 (9) 16 (16) 25 (11) 19 (11) Abbreviations: DCIS, ductal carcinoma in situ; IC, invasive carcinoma. Arch Pathol Lab Med Vol 142, April 2018 Breast Specimen Processing and Reporting Guidi et al 501 P

7 In particular, we focus on the aspects of tissue processing and reporting that most impact clinical decision making with regard to BCT. While it has been more than 2 decades since the publication of 6 randomized clinical trials reporting that patients with early-stage breast cancer experience equivalent survival with BCT compared to mastectomy, it has been only recently that multidisciplinary consensus panels under the auspices of SSO, ASTRO, and ASCO have reported on the results of meta-analyses, and have promulgated robust evidence-based treatment guidelines that have important implications for pathologists. Together with the CAP breast cancer protocol recommendations, 1,2 the SSO/ASTRO consensus guideline for patients with IC 13 and the SSO/ASTRO/ASCO consensus guideline for patients with DCIS 18 provide important resources that pathologists should consider to help ensure their reports result in the most appropriate patient treatment decisions. The results of our survey indicate that a substantial proportion of survey participants do in fact process breast specimens and issue breast cancer reports in a manner that complies with CAP breast cancer protocols and is in alignment with the SSO/ASTRO and SSO/ASTRO/ASCO consensus guidelines. However, the survey results also highlight some variability in practices that warrant further discussion, spanning the entire process from the intraoperative examination of specimens to the generation of final reports. Breast specimen processing begins in the operating room. Pathologists may be asked to perform intraoperative evaluation of breast specimens for a number of reasons, including margin assessment. The results of our survey indicate that two-thirds of institutions currently do not perform intraoperative margin assessment. Of the institutions that perform intraoperative margin assessment, approximately two-thirds perform gross examination only, while one-third use frozen section examination, imprint cytology, or some combination of methods. While published studies examining imprint cytology for intraoperative margin evaluation have reported encouraging results, this procedure requires interpretation by pathologists with specialty training in cytology, limiting its broad application, and this may at least in part explain its limited use by institutions in our survey. With regard to the use of frozen section evaluation of margins, while some have reported that this procedure reduces the frequency of subsequent reexcisions, others 34 have questioned the utility of this approach. Of note however, the studies reporting decreased rates of re-excision with frozen section margin evaluation were published before widespread adoption of the SSO/ ASTRO guidelines for patients with IC, which indicate a lack of benefit of obtaining margins greater than no tumor on ink. 13 Therefore, the benefit of frozen sections in this context may no longer be as significant. Regardless, similar to imprint cytology there has been limited adoption of frozen section margin evaluation among participants in this survey. We also surveyed institutions with regard to 2 additional intraoperative considerations: (1) the role surgeons may play in the inking of breast excision specimens, and (2) the use of cavity (shave) margins by surgeons. There are theoretical benefits of specimen inking by surgeons, including (1) a better appreciation by surgeons of patient-specific anatomic considerations and specimen orientation, compared to laboratory personnel, who may be limited not only by a lack of visualization of the anatomy during specimen removal but also by artifacts secondary to specimen compression (the pancake phenomenon ), 35 and (2) eliminating the time required to place orienting sutures and provide suture notation information on specimen requisitions. Nevertheless, 71% of respondents in our survey reported that specimen inking is performed exclusively by laboratory personnel. With regard to the use of cavity (shave) margins, multiple investigators have reported a reduced rate of final positive margins and a reduced need for subsequent re-excisions compared to simple excision, in both retrospective and prospective studies. Our survey results confirm the growing popularity of this technique, as 78% of respondents report that cavity margins are used in at least a subset of cases at their institution. While 88% of respondents in our survey report that they apply ink to the new true cavity margin, it is notable that 12% do not ink cavity margins. Not inking cavity margins may in some cases preclude adherence to CAP breast cancer protocols and to SSO/ASTRO and SSO/ASTRO/ASCO guidelines defining a positive margin as tumor on ink, and in other cases may preclude measuring the distance between tumor and inked margin. The SSO/ASTRO and SSO/ASTRO/ASCO consensus panel conclusions are consistent with CAP breast cancer protocols with respect to how positive margins should be defined ( tumor on ink ), and how negative margins should be reported. In our survey, most respondents comply with CAP recommendations and SSO/ASTRO and SSO/ASTRO/ ASCO guidelines defining positive margins as tumor on ink in IC (94%) and DCIS (91%). However, for the minority of institutions that include cases without tumor on ink as positive margins, there is significant concern based on the results of the SSO/ASTRO and SSO/ASTRO/ ASCO meta-analyses that unnecessary re-excisions could be performed in some patients, or mastectomies could be performed in patients who are appropriate candidates for BCT. One intriguing finding regarding margin-reporting practices in this survey concerns which diagnostic entities are included in margin assessments. As expected, most institutions (99%) report margins for IC and DCIS. However, an unexpected finding in this survey was the number of respondents who report margins for entities for which margin evaluation does not typically impact treatment decisions. In particular, 63% of respondents report margins for classical LCIS, and 42% report margin status for ADH. Lobular carcinoma in situ represents both a risk factor and a nonobligate precursor for the development of invasive carcinoma in either breast. 41,42 The presence of LCIS at margins in association with IC is not associated with an increased risk for IBTR. 43,44 For practical purposes, classical LCIS should be considered a benign neoplasm, and in fact LCIS has been removed from the Tis classification in the 8th edition of the AJCC (American Joint Committee on Cancer) Cancer Staging Manual. 45 For these reasons, reporting margin status for classical LCIS is not recommended. 41 Similarly, ADH is regarded as a risk factor for the development of invasive cancer in either breast. Despite published studies raising a possible role for reporting margins in ADH, this practice is currently not supported in the NCCN guidelines. 17 In view of current guidelines for reporting and treating patients with breast cancer, and in the absence of compelling literature to support the practice, the routine reporting of margins for classical LCIS and ADH should be discouraged. In fact, reporting this information 502 Arch Pathol Lab Med Vol 142, April 2018 Breast Specimen Processing and Reporting Guidi et al

8 may cause confusion among caregivers, raising the risk of unnecessary additional surgical procedures. Moreover, reporting this information may also raise unnecessary consternation and confusion among patients, which is a particular concern given the increasing frequency with which patients have access to their pathology reports via electronic patient portals. In contrast to reporting margin status in classical LCIS and ADH, a reasonable justification can be made to report margin status in cases of pleomorphic LCIS, other forms of variant LCIS and ADH with severe atypia, bordering on DCIS. Pleomorphic and other variant forms of LCIS (including LCIS with comedo-type necrosis) are entities with aggressive morphologic features that are often treated in a similar fashion to DCIS even though the natural history of these lesions is poorly understood The NCCN guidelines state that clinicians may consider excision to negative margins in patients with variant LCIS, despite the fact that outcome data supporting the efficacy of this approach are lacking. 17 In our survey, 88% of respondents report margin status for pleomorphic LCIS and 59% report margin status for other types of variant LCIS. Similarly, reporting positive margins in excision specimens with ADH bordering on DCIS may be justified in order to more fully evaluate the lesion to see if it fulfills the criteria for DCIS. 57 In our survey, 67% of respondents report margin status for ADH bordering on DCIS. The CAP breast cancer protocols and the SSO/ASTRO and SSO/ASTRO/ASCO guidelines have major implications with regard to how specimen margins are sampled. There are 2 methods of margin sampling in general use: (1) perpendicular margins, which allow visualization of ink on tumor with positive margins and measurement of the distance between the leading edge of tumor and ink with negative margins, and (2) en face, or parallel margins, in which margins are peeled from the outer aspect of the specimen in a manner analogous to peeling the skin from an orange. 58,59 The en face technique impacts how margins are defined and reported, because a positive margin with this technique is no longer defined as tumor on ink, but rather tumor present anywhere on the slide. Moreover, the en face technique does not allow a measurement between the leading edge of a tumor and ink for close but negative margins, in contrast to perpendicular margins. In our survey, 76% of respondents reported examining perpendicular margins exclusively, while 23% reported examining en face margins in at least some circumstances. Respondents who reported using both methods often stated that they evaluate perpendicular margins when grossly evident tumor is close to margins, and en face margins when grossly evident tumor is away from margins. However, the major concern with this approach is that many lesions (particularly DCIS and invasive lobular carcinomas) are often not grossly evident. Optimal compliance with the CAP breast cancer protocols and the SSO/ASTRO and SSO/ASTRO/ASCO guidelines therefore favors the exclusive use of the perpendicular margin technique, because this is the only method that allows visualization of tumor on ink and a measurement of the distance between tumor and inked margins in all cases. In a study directly comparing both the perpendicular and the en face margin techniques, an important disadvantage of the en face method was the higher frequency of margins categorized as positive that would have been categorized as negative by the perpendicular margin method. 60 As noted by the SSO/ASTRO and SSO/ASTRO/ ASCO consensus panels, if pathologists use the en face margin technique and subsequently incorrectly categorize negative margins as positive, this may result in unnecessary re-excisions, or may lead to mastectomies for patients when BCT is an acceptable alternative. 13,18 Another important margin-related issue addressed in this study concerns the extent of margin sampling. Institutions were queried with regard to how many margins are sampled for excisions performed for a macroscopically identifiable mass that was previously diagnosed by core biopsy as DCIS or IC. While 71% of respondents reported either submitting all tissue from margins or at least sampling all margins, 29% sample a subset of margins; for those that perform limited sampling, 15% sample the closest margin, and 13% sample all margins within a certain specified distance of the mass. Most institutions are therefore in compliance with the CAP breast cancer protocols, which state that all margins should be evaluated grossly and histologically in a manner that allows measurement of the closest margin, when possible. In addition to margin status, other important issues related to the processing and reporting of breast cancer specimens were addressed in this survey, including the extent of specimen sampling. The survey specifically queried institutions regarding the extent of sampling for wire localization excision specimens when DCIS was previously diagnosed on core biopsy. While 29% of institutions reported submitting the entire specimen for histologic evaluation regardless of specimen size, 9% reported submitting all grossly evident fibroglandular tissue. The remaining institutions sample the specimens in a more limited way, with the largest proportion (54%) submitting the entire area near the localization wire, biopsy clip, or targeted abnormality. The results are interesting in view of the CAP protocol for DCIS, which highly recommends that in this circumstance the entire specimen should be submitted in a sequential fashion for histologic examination, 2 although the protocol does stipulate when practical, which may be interpreted differently by different laboratories. The purpose of complete or at least extensive sampling in cases of DCIS is to evaluate for possible invasion, to evaluate all margins, and to aid in determining the extent of DCIS. The DCIS protocol further states that if complete histologic evaluation is not possible, at least the entire region of the targeted lesion should be examined microscopically. 2 In this respect, it appears that most respondents in our survey are in substantial compliance with the CAP breast cancer protocols. Another important element regarding breast processing and reporting concerns defining the extent of DCIS. This is a required element in the CAP DCIS protocol 2 and a nonrequired (but suggested) element in the CAP IC protocol. 1 The extent of DCIS in excision specimens is clinically relevant because it correlates with the presence of residual disease following re-excision, 3 6 and perhaps more importantly because it correlates with the risk of IBTR In addition, larger DCIS lesions are associated with a higher probability of finding areas of invasion. 11,12 In our survey, 95% of respondents provide an estimate of the extent of DCIS. This is a significantly higher rate of compliance than noted in a prior study in which only 21% of breast cancer reports reviewed in a formal audit provided an estimate of the size of DCIS, although it is important to note the rate of compliance in our study is based on self-reporting rather than an actual audit. 61 Institutions that report the extent of DCIS use multiple methods, which are discussed in detail in Arch Pathol Lab Med Vol 142, April 2018 Breast Specimen Processing and Reporting Guidi et al 503

9 the CAP DCIS protocol 2 and elsewhere. 59,62,63 When multiple methods are used to measure the size/extent of DCIS in the same case, they may provide different estimates, and in this circumstance the explanatory notes section of the CAP DCIS protocol suggests that the largest estimate obtained by the various methods should be used to report the estimated size (extent) of DCIS. 2 In our survey, respondents differ in their approach to reconciling varying estimates, with only 32% of respondents favoring the method yielding the largest measurement. In addition, the CAP DCIS protocol explanatory notes suggest that when DCIS is limited to 1 slide, and multiple foci of DCIS are present, the largest distance between foci should be reported. 2 However, in our survey, 65% of respondents consider each focus separately and provide independent measurements, while 30% include the distance between foci in the size determination. While the variability of practices in determining and reporting the extent of DCIS does not represent noncompliance with required data elements in the CAP breast cancer protocols, it may represent an opportunity to work toward greater standardization among laboratories. An additional important consideration addressed in this study involves the documentation of prior marker clips and biopsy site changes in breast excision specimens. Marker clips are typically placed at the time of core biopsy to help guide subsequent excisions, if necessary. To document that the clip was removed in the procedure, institutions commonly re-image excision specimens in the operating room, radiology department, or in some cases the laboratory. Although not specifically addressed in the CAP breast cancer protocols, many laboratories comment onthepresenceorabsenceofthebiopsyclipsinexcision and mastectomy specimens in order to provide additional documentation of their removal. In our survey, 88% of respondents document the presence or absence of specimen marker clips in patient reports in the gross description, the diagnostic portion of the report, or both. In addition to documenting the presence or absence of marker clips, many institutions also comment on the presence or absence of histologic evidence of a prior biopsy ( biopsy site changes ) in both excision specimens and mastectomies. The CAP breast cancer protocols state that if there has been a prior core biopsy or excision, the biopsy site should be sampled and documented in the report. 1,2 In our survey, 96% of respondents comment on the presence or absence of biopsy site changes somewhere in the diagnostic report. We believe that documentation of marker clips and biopsy site changes is useful from a patient safety and risk management perspective, to help confirm that the lesion of interest is completely excised. This documentation can be difficult in some cases, particularly when multiple prior biopsies have been performed, and the surgeon s intent may be to remove all or just a subset of the prior biopsy sites. Effective communication between surgeons, radiologists, and pathologists is therefore imperative in these circumstances. One final consideration of interest in our study concerns whether key aspects of specimen processing and reporting are detailed in formal written procedures. With the exception of specimen inking, where just over half of respondents report developing formal written procedures, only 20% to 30% of respondents reported having formal written procedures in place, detailing most elements of specimen processing and reporting. We believe that creating such procedures may have potential benefits for laboratories, including (1) creating a first step toward ensuring standardization and compliance with relevant guidelines, (2) serving as a useful tool for training pathologists, residents, and pathologists assistants new to a practice setting, and (3) serving as a useful reference for part-time pathologists/ pathologists assistants, those who work in multiple practice settings, and those who infrequently process and report breast specimens. In conclusion, the survey results indicate that a substantial proportion of institutions process and report breast specimens in a manner that is compliant with CAP breast cancer protocols. Moreover, although we did not specifically ask survey participants if they were familiar with the SSO/ ASTRO guidelines for IC and the SSO/ASTRO/ASCO guidelines for DCIS were published after our survey was conducted the survey results indicate that a substantial proportion of institutions also process breast specimens and issue reports in a manner that is in alignment with both guidelines. While we do highlight potential opportunities for process improvement for a subset of laboratories, we also acknowledge that in some instances deviation from published guidelines may be necessary owing to the specific circumstances of a case, and is not necessarily an indication of substandard practice. There may be other reasons underlying the variability demonstrated in this survey. In some instances perceived variability may reflect the complexities inherent in examining gross specimens and reporting findings that cannot be captured with a general survey tool like the one used in this study. In other circumstances variability may reflect pathology departments deferring to institutionally defined clinical practices and protocols. In fact, in the preface to the CAP breast cancer protocols, the CAP specifically states that...the manner in which these elements are reported is at the discretion of each specific pathologist, taking into account clinician preferences, institutional policies, and individual practice. 1,2 Nevertheless, the closer pathology departments move toward standardization in the processing and reporting of breast specimens in a manner that is most compliant with CAP breast cancer protocols and with SSO/ ASTRO and SSO/ASTRO/ASCO guidelines, the greater chance we have to assure that patients are treated in the most cost-effective manner possible, with optimal clinical outcomes. The authors would like to formally acknowledge Stuart J. Schnitt, MD, from the Beth Israel Deaconess Medical Center, Boston, Massachusetts, and Susan C. Lester, MD, PhD, from the Brigham and Women s Hospital and Dana Farber Cancer Institute, Boston, Massachusetts, for helpful comments they provided in the preparation of this manuscript. References 1. Lester SC, Bose S, Chen Y, et al. Protocol for the examination of specimens from patients with invasive carcinoma of the breast. Version cap.org/web/home/resources/cancer-reporting-tools/cancer-protocol-templates?_ afrloop¼ #!%40%40%3f_afrloop%3d % 26_adf.ctrl-state%3D19nhx42b0v_17. Accessed June 13, Lester SC, Bose S, Chen Y-Y, et al. Protocol for the examination of specimens from patients with ductal carcinoma in situ of the breast. Version D %26_adf.ctrl-state%3D19nhx42b0v_17. Accessed June 13, Cheng L, Al-Kaisi NK, Gordon NH, Liu AY, Gebrail F, Shenk RR. Relationship between the size and margin status of ductal carcinoma in situ of the breast and residual disease. J Natl Cancer Inst. 1997;89(18): Arch Pathol Lab Med Vol 142, April 2018 Breast Specimen Processing and Reporting Guidi et al