A new way of looking at breast cancer tumour biology

A new way of looking at breast cancer tumour biology

Contents Intrinsic subtypes of breast cancer 3 Gene expression assays 3 Basis of the Prosigna test 4 Information provided by Prosigna 5 The accuracy and utility of Prosigna 7 Unselected cohorts (no systemic therapy) 7 Post-menopausal breast cancer 8 Pre-menopausal breast cancer 8 Neo-adjuvant chemotherapy 9 Advanced breast cancer 9 Other tests for intrinsic subtype 10 Other indicators of recurrence risk 12 Other gene expression assays 13 Uses of Prosigna 14 Ordering Prosigna 14 References 15 Prosigna is a test of gene activity in breast cancer. It assesses a patient s pattern of genetic activity to provide a more accurate, personalised prediction about the risk of recurrence, and informs decisions about cancer treatment. Fifteen years ago, researchers found that each breast cancer has one of four different patterns of gene activity. Each pattern can be associated with different risks of cancer recurrence, and with different responses to chemotherapy. The significance of these patterns of gene activity has now been described in thousands of medical publications. Prosigna uses this information to help you and your patient decide whether certain treatments, such as chemotherapy, are warranted and, if so, what type of chemotherapy is likely to be the most effective. The accuracy of the Prosigna prediction has been proven in numerous independent studies involving thousands of women, and has been shown to be more reliable in determining the pattern of gene activity, and more accurate in predicting the risk of recurrence, than other genetic tests. This means that Prosigna provides a more secure basis for your decisions about cancer treatment than other tests. The use of Prosigna has also been endorsed by independent international professional and regulatory organisations. Intrinsic subtypes of breast cancer Breast cancers can be characterised by a multitude of clinical and pathological features. These may provide prognostic information (risk of recurrence) and predictive information (response to therapy), but the accuracy of this information is often limited. A molecular classification of breast cancer based on patterns of activity in 1,753 genes was first proposed in 2000 1. This study described four distinct patterns of gene activity, or intrinsic subtypes, called Luminal A, Luminal B, HER2-enriched and Basal-like. (A fifth category, normal, was subsequently removed from the classification as it reflects a failure to sample malignant tissue rather than there being a category of cancer which exhibits normal gene expression.) Since then, this classification has been repeatedly utilised in breast cancer research and management in 2,500 PubMed articles and in more than 80 prospective clinical trials. Gene expression assays The same four patterns of gene activity can be reliably identified with a subset of 50 genes (called PAM50 ), thereby simplifying the molecular analysis of breast cancer. Various methods and analyses of the PAM50 set of genes have been described 2, with more than 120 publications describing the use of this particular subset *. Assays of this type typically involve using a set of reference samples to train the assay, i.e. to adjust the algorithm to provide the most accurate answer, and then assessment of the trained assay with a second independent reference set to validate the training. Once these statistical requirements have been met, the assay should be stable and reproducible when used by that laboratory for the patients represented in the training and validation cohorts. 3

Basis of the Prosigna test Information provided by Prosigna Prosigna is a new implementation of the PAM50 assay which, in contrast to earlier PCR assays, makes a digital count of molecules. This makes the assay robust in a wide variety of laboratory settings 3. Prosigna was developed using a broad range of patient samples (as FFPE tissue blocks) from multiple cohorts and sites, and is not restricted by the specific type of breast cancer, grade, histopathology, tumour size or age of patients. The patients were North American (40% African American, 60% Caucasian). Samples are FFPE and can be up to 10 or more years old. Assay results are consistent in multiple samples from the one tumour, and are robust in the presence of blood, necrotic tissue, DCIS, and up to 70% normal tissue 4. Gene expression is measured by digitally counting the number of RNA molecules in the tissue sample from each gene without requiring enzymatic amplification. The assay shows high correlation with a benchmark study using the PAM50 assay (see proliferation subscore in figure at right). The sample analysis and interpretation are locked and remain stable within and between runs, and between laboratories 4. Prosigna has regulatory approval in Europe (CE Mark), Canada, the USA (FDA) and Australia (TGA); our implementation of the Prosigna test has also been accredited by NATA/RCPA. Previously Published Proliferation Score -2-1 0 1 2 Training Samples Proliferation Score R-squared = 0.978 Slope = 1 Y-intercept = 0.02-2 -1 0 1 2 Prosigna Proliferation Score Prosigna analyses the activity of 50 genes in the patient sample. On the basis of the pattern of gene activity, it assigns the patient s cancer to the most likely intrinsic subtype. This assignment of the patient to a diagnostic category is the basis for providing prognostic and predictive information. Prosigna goes a step further and determines a Risk of Recurrence (or ROR) score for the individual patient. The score is a weighted sum of the correlation coefficients of gene expression in the sample with each of the subtype references, the correlation coefficient of the sample with a subset of genes (across all subtypes) involved in proliferation, and tumour size. 10-Year Predicted Risk of Distant Recurrence (%) 100 80 60 40 20 Node negative Node positives (1-3 positive nodes) Rather than provide information on the basis of the patient being in a diagnostic category, the test is able to provide information for the individual patient. The risk of recurrence for a given ROR Score varies with the number of involved lymph nodes as shown in the figure below. The distinction between category-based versus individual information is important, as the evidence for the performance of Prosigna may be based on cohort information, e.g. outcomes for a subtype category, or on the individual information, e.g. prognosis indicated by the score, or both. 0 0 20 40 60 80 100 ROR 4 5

The accuracy and utility of Prosigna The performance of the ROR score and of intrinsic subtyping by PAM50 and, more recently, the Prosigna assay, has been assessed in a number of different settings. The equivalent performance of PAM50 and Prosigna is best documented in post-menopausal women with early-stage hormone-sensitive breast cancer; this is the basis for FDA approval of Prosigna in this setting. Given the stability and reproducibility of Prosigna, it is anticipated that the test will have equivalent (or better) performance than the PAM50 assay in other settings. No adjuvant systemic therapy Unselected cohorts (no systemic therapy) Parker et al 2 reported a set of 761 women derived from five different cohorts. The cancers varied in histology, grade, and size, and included pre- and post-menopausal women. There were major differences in risk of recurrence (prognosis determined by the PAM50 assay) that were highly significant. Relapse-Free Survival (probability) 1.0 0.8 0.6 0.4 0.2 Basal-like HER2-enriched Luminal A Luminal B Log-rank P = 2.26e-12 0 2 4 6 8 10 Time (years) 1.0 In the same data set, the patients were categorised by PAM50 as being at low, intermediate and high risk on the basis of the ROR Score. The observed rates of recurrence were substantially different and highly significant (Figure at right). Relapse-Free Survival (probability) 0.8 0.6 0.4 High Low Medium Log-rank P = 9.52e-11 0 2 4 6 8 10 Time (years) 6 7

Post-menopausal breast cancer This is the setting for which there is the most evidence regarding Prosigna test performance currently available; other trials are ongoing and being reported. The accuracy of Prosigna has been assessed in a number of trials of women with early-stage, hormone-sensitive post-menopausal breast cancer, who were taking adjuvant hormonal therapy. Intrinsic subtype and ROR score have been demonstrated to provide clinically and statistically significant information regarding prognosis in a Vancouver cohort 3,6, the ATAC trial (Figure at right) 7, and the ABCSG8 trial 8,9. Patients (%) 100 90 80 70 HR, 4.78; 95% CI, 2.97 to 7.70 P <.001 Luminal A, N0 Luminal B, N0 60 0 1 2 3 4 5 6 7 8 9 10 Follow-Up Time (years) Neo-adjuvant chemotherapy In a sequential cohort of patients requiring neo-adjuvant chemotherapy, patients with Luminal B cancers determined by PAM50 subtyping had a 9-fold relative risk of recurrence versus patients with Luminal A cancers 14. In a study of combination neo-adjuvant chemotherapy in invasive ductal cancer 2, the intrinsic subtype was able to stratify groups according to responsiveness. Prosigna subtype Odds ratio of response P-value Luminal A 1 reference Luminal B 3.0 0.02 HER2-enriched 1.9 0.13 Basal-like 2.4 0.04 The prognostic information extends to accurately estimating the risk of recurrence between five and 10 years after the initial diagnosis (Figure at right) 10,11. This information may be useful in determining the duration of hormonal therapy in this setting. Distant Recurrence (%) 20 15 10 5 0 HR (95% CI) Low (n = 1,183; 55.4%) - Intermediate (n = 538; 25.2%) 3.26 (2.07 to 5.13) High (n = 416; 19.5%) 6.90 (4.54 to 10.47) Low (0-26) Intermediate (26-68) High (> 68) In a study of neo-adjuvant chemotherapy in ER+/HER2- breast cancer 16, intrinsic subtyping using Prosigna demonstrated that Luminal A cancers had a lower probability of responding than other subtypes. Prosigna subtype Odds ratio of response P-value Luminal A 0.34 0.04 Other subtypes reference 5 6 7 8 9 10 Pre-menopausal breast cancer Follow-Up Time (years) Three studies have examined the utility of the PAM50 assay in managing adjuvant therapy in pre-menopausal breast cancer. The ROR score provided prognostic information in this setting. In addition, the intrinsic subtype determined by PAM50 provided predictive information, identifying subtypes which responded to adjuvant hormone therapy (Figure at right) 12. The intrinsic subtype determined by immunohistochemistry was unable to identify this group (also see section below, Other tests for intrinsic subtype). In a separate study, the intrinsic subtype provided predictive information, identifying a subtype (HER2-enriched) which was comparatively sensitive to adjuvant anthracycline 13. PAM50 subtype and HER2 status by IHC provided independent information predicting responsiveness. (see also 15 ). IHC Classifier Luminal subtypes Other subtypes PAM50 Classifier Luminal subtypes Other subtypes 0 0.5 1 1.5 Tamoxifen Better Placebo Better 0.75 (0.51, 1.09) 0.71 (0.40, 1.26) 0.52 (0.32, 0.86) 0.80 (0.50, 1.29) Advanced breast cancer In patients with advanced disease or post-relapse, the PAM50 intrinsic subtype provided prognostic information regarding time to progression 17 and disease-specific survival 18. The subtype could also indicate the probability of response to certain chemotherapy protocols (Figure at right) 17. Group n HR 95% CI PAM50 subtypes Luminal A Luminal B Basal like HER2 enriched P=.50 Luminal A Non Luminal A P=.47 Basal like Non Basal like P=.22 84 97 43 46 84 186 43 227 0.71 0.55 0.39 0.82 0.72 0.56 0.39 0.66 (0.39 to 1.29) (0.32 to 0.94) (0.19 to 0.82) (0.39 to 1.75) (0.40 to 1.30) (0.38 to 0.81) (0.19 to 0.82) (0.46 to 0.94) Favors G+D Favors D 0.15 0.3 0.6 1 2 4 8 9

Other tests for intrinsic subtype A combination of immunohistochemical stains (ER, PR, HER2, and Ki67) has been proposed as an alternative means of defining intrinsic subtype 19. This surrogate definition of subtypes has the advantage of being more readily available than molecular testing, but there is a lack of concordance between the IHC and molecular methods of defining subtype. In a recent analysis 20 of three trials in which cancers where subtyped by IHC and PAM50 (Table below), there was frequent disagreement in the subtyping by these two methods. IHC Luminal A Luminal B HER2-enriched Basal-like IHC disagrees with PAM50 Luminal A 477 16 3 6 5% Lumina B 346 39 8 15 6% HER2-enriched 102 69 87 48 72% Basal-like 8 0 6 207 6% PAM50 disagrees with IHC 12% 69% 16% 25% 1,437 This Table excludes 65 samples in which ER/PR staining was borderline. The comparison did not include Ki67, and the Ki67-dependent distinction between Luminal A and B is not considered in this comparison. Immunohistochemical methods show less within-lab consistency in subtyping 21 than Prosigna, even when Prosigna is compared between labs 4. 10 11

Other indicators of recurrence risk The ability of a continuous variable (recurrence score) to predict a binary outcome (distant recurrence within 10 years) is typically assessed using an ROC curve. The area under the curve (c-statistic) is a measure of the analytical power of the test. Other gene expression assays Two other gene expression assays for prognosis or prediction in breast cancer using FFPE samples are available in Australia. The following Table summarises their performance relative to Prosigna. The upper Figure shows the c-statistics for Prosigna; a combination of ER, PR, HER2, and Ki67 (IHC4); a combination of nodal involvement, grade, size, age (CTS); and OncotypeDX (ODx) for node-negative women in the TransATAC trial 7. The lower figure shows the same statistics for node-negative/her2- women in this trial. The prognostic ability of Prosigna exceeds that of the other indicators in these node-negative women. ATAC node Prosigna IHC4 CTS ODx 0.5 0.55 0.6 0.65 0.7 0.75 0.8 c-statistic ATAC node-/her2- Comparator Prosigna Oncotype Dx Endopredict Methodology NanoString (digital) PCR (analog) PCR (analog) Intrinsic subtype Yes No No Prediction by subtype Yes No No Precision of estimate at 10% risk * +/- 1.5% (node -) +/- 2.5% (node -) +/- 2.0% (node +/-) Accuracy versus IHC/clinical scores Better 7 Worse/same 7 Better 22 Accuracy of risk prediction vs Oncotype DX Better 7 Reference Unknown % patients at intermediate risk (ATAC trial) 7 24% 33% Not studied Australian laboratory Yes No Yes Regulatory approval in Australia TGA/NATA Not applicable TGA * Data from companies literature PAM50 has also been tested against immunohistochemical and clinical indicators in other trials, and has demonstrated better accuracy in pre- and post-menopausal breast cancers 2,3,12. Prosigna IHC4 CTS ODx 0.5 0.55 0.6 0.65 0.7 0.75 0.8 c-statistic There are multiple studies documenting the performance of Prosigna in predicting response to therapy (see above). There is one study demonstrating the predictive performance of Endopredict in patients requiring neo-adjuvant chemotherapy 23. Studies regarding the predictive performance of Oncotype DX are compromised by reliance on a study on the NSABP-20 cohort 24. This study used the same data set to develop and then validate the claims for Oncotype Dx s predictive ability. As noted above, the training and validation sets used in developing a gene expression assay must be independent if an assay is to have any validity. A critique of this study 25 demonstrated that the data presented actually showed that Oncotype DX had no predictive value in the validation cohort. Other reports of the predictive performance of Oncotype DX have used a variety of clinical factors in conjunction with the Oncotype DX result. The applicability of such combinations beyond these studies remains uncertain. 12 13

Uses of Prosigna Ordering Prosigna References Prosigna has demonstrated utility in decision-making for neoadjuvant chemotherapy, adjuvant chemotherapy, adjuvant hormone therapy and chemotherapy after relapse. The studies involved both pre-menopausal and post-menopausal women. Prosigna provides more information, and is more accurate, than other genetic and immunohistochemical tests available in Australia. The utility of Prosigna extends to women at apparently low risk of recurrence by clinical criteria. In one study, one third of postmenopausal women at low risk of recurrence on clinical criteria were at high risk based on the more accurate estimates provided by Prosigna 11. Prosigna should be ordered by the clinician responsible for decision-making regarding adjuvant therapy for a woman with breast cancer. We recognise that Prosigna is not necessary for every woman facing significant decisions in breast cancer management, but it can provide important information in many instances. To order Prosigna, please complete the special request form which is available at www.sonicgenetics.com.au/tests/ prosigna. Please provide the clinical information requested and attach a copy of the histology report and of any immunohistochemical studies. The patient is requested to make prepayment for the test. This can be done via our website or by telephoning 1800 010 447. The request form (and associated reports) are then faxed or emailed to our laboratory. We will contact the laboratory holding the FFPE sample to arrange transfer of the tissue sections necessary for the test. On receipt of the sections, the test takes 10 working days. 1. Perou, C. M. et al. Molecular portraits of human breast tumours. Nature 406, 747 752 (2000). 2. Parker, J. S. et al. Supervised risk predictor of breast cancer based on intrinsic subtypes. J. Clin. Oncol. 27, 1160 7 (2009). 3. Wallden, B. et al. Development and verification of the PAM50-based Prosigna breast cancer gene signature assay. BMC Med. Genomics 8, 54 (2015). 4. Nielsen, T. et al. Analytical validation of the PAM50-based Prosigna Breast Cancer Prognostic Gene Signature Assay and ncounter Analysis System using formalin-fixed paraffin-embedded breast tumor specimens. BMC Cancer 14, 1 14 (2014). 5. ProSigna Product Insert (2013). 6. Nielsen, T. O. et al. A comparison of PAM50 intrinsic subtyping with immunohistochemistry and clinical prognostic factors in tamoxifen-treated estrogen receptor-positive breast cancer. Clin. Cancer Res. 16, 5222 32 (2010). 7. Dowsett, M. et al. Comparison of PAM50 risk of recurrence score with oncotype DX and IHC4 for predicting risk of distant recurrence after endocrine therapy. J. Clin. Oncol. 31, 2783 90 (2013). 8. Gnant, M. et al. Predicting distant recurrence in receptor-positive breast cancer patients with limited clinicopathological risk: Using the PAM50 Risk of Recurrence score in 1478 postmenopausal patients of the ABCSG-8 trial treated with adjuvant endocrine therapy alone. Ann. Oncol. 25, 339 345 (2014). 9. Gnant, M. et al. Identifying clinically relevant prognostic subgroups of postmenopausal women with node-positive hormone receptor-positive early-stage breast cancer treated with endocrine therapy: a combined analysis of ABCSG-8 and ATAC using the PAM50 risk of recurrence. Ann. Oncol. 1 7 (2015). doi:10.1093/annonc/mdv215 10. Sestak, I. et al. Factors predicting late recurrence for estrogen receptor-positive breast cancer. J. Natl. Cancer Inst. 105, 1504 11 (2013). 11. Sestak, I. et al. Prediction of Late Distant Recurrence After 5 Years of Endocrine Treatment: A Combined Analysis of Patients From the Austrian Breast and Colorectal Cancer Study Group 8 and Arimidex, Tamoxifen Alone or in Combination Randomized Trials Using the PAM50 Risk. J. Clin. Oncol. 1 10 (2014). doi:10.1200/jco.2014.55.6894 12. Chia, S. K. et al. A 50-Gene Intrinsic Subtype Classifier for Prognosis and Prediction of Benefit from Adjuvant Tamoxifen. 18, 4465 4472 (2013). 13. Cheang, M. C. U. et al. Responsiveness of intrinsic subtypes to adjuvant anthracycline substitution in the NCIC.CTG MA.5 randomized trial. Clin. Cancer Res. 18, 2402 12 (2012). 14. Lips, E. H. et al. Breast cancer subtyping by immunohistochemistry and histological grade outperforms breast cancer intrinsic subtypes in predicting neoadjuvant chemotherapy response. Breast Cancer Res. Treat. 140, 63 71 (2013). 15. Liu, S. et al. Prognostic and predictive investigation of PAM50 intrinsic subtypes in the NCIC CTG MA.21 phase III chemotherapy trial. Breast Cancer Res. Treat. 149, 439 48 (2015). 16. Prat, A. et al. Prediction of Response to Neoadjuvant Chemotherapy Using Core Needle Biopsy Samples with the Prosigna Assay. Clin. Cancer Res. 119 129 (2015). doi:10.1158/1078-0432.ccr-15-0630 17. Jørgensen, C. L. T. et al. PAM50 breast cancer intrinsic subtypes and effect of gemcitabine in advanced breast cancer patients. Acta Oncol. 53, 776 87 (2014). 18. Tobin, N. P. et al. Molecular subtype and tumor characteristics of breast cancer metastases as assessed by gene expression significantly influence patient post-relapse survival. Ann. Oncol. 26, 81 88 (2014). 19. Goldhirsch, a et al. Strategies for subtypes--dealing with the diversity of breast cancer: highlights of the St. Gallen International Expert Consensus on the Primary Therapy of Early Breast Cancer 2011. Ann. Oncol. 22, 1736 47 (2011). 20. Cheang, M. C. U. et al. Defining breast cancer intrinsic subtypes by quantitative receptor expression. Oncologist 20, 474 82 (2015). 21. Ekholm, M. et al. Highly reproducible results of breast cancer biomarkers when analysed in accordance with national guidelines a Swedish survey with central re-assessment. Acta Oncol. (Madr). 1 9 (2015). doi:10.3109/0284186x.2015.1037012 22. Filipits, M. et al. A new molecular predictor of distant recurrence in ER-positive, HER2-negative breast cancer adds independent information to conventional clinical risk factors. Clin. Cancer Res. 17, 6012 6020 (2011). 23. Bertucci, F., Finetti, P., Viens, P. & Birnbaum, D. EndoPredict predicts for the response to neoadjuvant chemotherapy in ER-positive, HER2-negative breast cancer. Cancer Lett. 355, 70 5 (2014). 24. Paik, S. et al. Gene expression and benefit of chemotherapy in women with node-negative, estrogen receptor-positive breast cancer. J. Clin. Oncol. 24, 3726 34 (2006). 25. Ioannidis, J. P. a. Is molecular profiling ready for use in clinical decision making? Oncologist 12, 301 311 (2007). 26. Ward, R. L. et al. Cost of cancer care for patients undergoing chemotherapy : The Elements of Cancer Care study. Asia. Pac. J. Clin. Oncol. 11, 178 186 (2015). Prosigna costs $2,900 * ; there is no Medicare rebate and the test must be prepaid. This is likely to be a significant expense for patients. To place this cost in context, adjuvant chemotherapy for breast cancer costs an average of $3,600 per month, with the patient being out of pocket for $199 per month 26. The costs to an individual patient are hard to predict. From a financial perspective, the decision to utilise adjuvant chemotherapy should not be taken lightly, and Prosigna can provide important information to assist in making that judgement. If you have any further questions, please do not hesitate to contact your local Business Development Manager or call Sonic Genetics on 1800 010 447. * Current at December 2015 14 15

Further questions? Information for patients and their doctors is available on the Sonic Genetics website (www.sonicgenetics.com.au/tests/ prosigna). For further information, please refer to our website, www.sonicgenetics.com.au/tests/prosigna or call us on 1800 010 447 14 Giffnock Avenue, Macquarie Park, NSW 2113, Australia T (02) 9855 5369 F (02) 9855 5446 E: info@sonicgenetics.com.au www.sonicgenetics.com.au The published evidence that is the basis for this leaflet is detailed in other leaflets about Prosigna available on the Sonic Genetics website. Printed December 2015