10èmes Journées Scientifiques du Cancéropôle Nord-Ouest 10-12 mai 2017, Deauville Session thématisée Les Innovations diagnostiques en cancérologie Les signatures multigéniques pronostiques dans le cancer du sein Martine J. Piccart-Gebhart, MD, PhD Institut Jules Bordet, Brussels, Belgium Université Libre de Bruxelles Breast International Group (BIG aisbl), Chair Disclosures No conflict of interest with any of the companies which are marketing gene expression signatures BREAST CANCER TODAY Significant Progress Achieved Incidence Mortality Ferlay J. Globocan 2012 BREAST CANCER TODAY 5 year survival rates (SEER) ADJUVANT SYSTEMIC THERAPY FOR BREAST CANCER TODAY The Age of Escalation Targeted therapy SX Chemo RT Targeted therapy Endocrine therapy Cure a substantial proportion of women Treatment duration : anywhere from 1 day to 15 years ! ADJUVANT SYSTEMIC THERAPY FOR BREAST CANCER TODAY The Age of Escalation More Patients Treated Longer Drug Exposure Therapeutic Escalation More Drugs TREATMENT ESCALATION Consequences The cost threatens to blow up our health care systems – the ones that actually pay for treatments! Many women live longer but bear the consequences of our aggressive treatments – and we can never ignore this! BREAST CANCER « Adjuvant » medical therapies Lung Liver Bone Localized disease Curable But risk of : •overtreatment • undertreatment • wrong treatment • suboptimal treatment Generalized disease Very difficult to cure ADJUVANT SYSTEMIC THERAPY FOR EARLY BREAST CANCER Benefit / Risk Balance Lessons learned from 3 decades of clinical trials BENEFIT Survival (2 to 12%) POTENTIAL “HARM” • Mood alterations • Thrombo-embolic events • Arthralgia • Osteoporosis LONG-TERM RISKS • Secondary cancers • Cardiac toxicity • Early menopause • Cognitive function … AND SOCIO-ECONOMIC BURDEN ST. GALLEN DEFINITIONS OF RISK Low Intermediate R I S K High Node – G1 HER2 – T2 LVI absent AGE < 35 G2-3 T>2 Only 20% of patients! Most difficult group for CT decision! Node –, HER2+ or LVI present Node + (1-3) and HER2 - Node + (1-3) and HER2 + Node + 4 Other similar guidelines exist : NCCN, ESMO,… R I S K IMPROVED RISK ASSESSMENT OF EARLY BREAST CANCER THROUGH GENE EXPRESSION PROFILING L. van ‘t Veer R. Bernards 78 untreated N ─ primary tumors 295 partially treated N─ / N+ tumors microarray 44 w/o relapse at 8 y follow-up Gene-expression profile 34 with a relapse within 5 y 5000 genes 231 genes 70 genes Poor prognosis signature van ‘t Veer L., Nature 2002; 415 (31) : 530-536 Van de Vijver MJ, N Engl J Med 2002; 347 (24): 1999-2009 B.C. CLINICAL OUTCOME PREDICTION 70-gene profiler outperforms St Gallen criteria Van de Vijver MJ, N Engl J Med 2002; 347 (24): 1999-2009 Validation of the MammaPrint® signature : 15 years of intensive collaborative work! 2002 2003 Signature External validation published of the signature by the TransBIG network 2007-2011 2016 Conduct of MINDACT MINDACT Results EORTC-BIG Mindact Trial Design 6.694 node negative & 1-3 node positive women 70-gene signature (Mammaprint®) risk AND Clinical-Pathological risk Both low risk Discordant cases Both high risk No chemotherapy Randomization Chemotherapy Supported by the EU 6th framework grant (7 million euros) Total cost of trial ≈ 45 million euros! Tumor biology MammaPrint® Low vs High genomic risk Tumor anatomy VERSUS (+ a few biological features) Adj.! Online Low vs High clinical risk Hypothesis : the Genomic assay will outperform the Clinical criteria by reducing the prescription of adjuvant chemotherapy WITHOUT IMPAIRING OUTCOME More specifically: clinically high risk patients with a low risk gene signature randomized not to receive CTX should have a 5 year DMFS of ≥ 92% The MINDACT study: Patient demographics N = 6693 Median age = Node Node + T1 tumours Grade 2 HR positive HER2+ 55y 79% 21% 72% 49% 88% 10% Discordant N=2745 N=1806 clinical Low/ genomic Low clinical High/ genomic High N=592 N=1550 clinical Low/ genomic High clinical High/ genomic Low Clinical « low risk » (50%) Clinical « high risk » (50%) ER status HER2 status Grade Nodal status N─ HER2 negative ER positive well differentiated 1-3 positive nodes moderately differentiated poorly differentiated OR undifferentiated N─ 1-3 positive nodes N─ 1-3 positive nodes Tumor size Clinical Risk in MINDACT ≤ 3cm C-low 3.1-5 cm C-high ≤ 2 cm C-low 2.1-5 cm C-high ≤ 2 cm C-low 2.1-5 cm C-high Any size C-high ≤ 1 cm C-low 1.1-5 cm C-high Any size C-high C-high risk : expected 10 y OS < 92% with endocrine therapy alone (as per Adjuvant! Online) The MINDACT study at a median follow-up of 5 years N = 6693 women N = 672 relapses 208 deaths N = 362 distant relapses Clinical outcome of the MINDACT population at 5y median follow-up DMFS IN ALL 4 RISK GROUPS Clinical outcome of the MINDACT population at 5y median follow-up DISCORDANT RISK GROUPS: PRIMARY TEST The primary analysis population The primary statistical test (DMFS at 5Y) Discordant risks c-Low /g-High c-High/g-Low RANDOMIZATION No chemotherapy N = 748 No change in risk post enrollement and no CT received N = 644 CT Null Hypothesis: set at 92% Observed 5Y DMFS = 94.7% 95% CI ≈ 92.5 – 96.2% excludes 92% !!! MINDACT: whole population vs “key subgroup” MINDACT High clin risk/low genomic risk (N=1550) MINDACT Entire population (N=6693) T1 T2 T3 27% 1% T3 79% 52% 3+ nodes 22% G2 G1 49% 29% Er‒ PR‒ 12% HER2+ 9,5% 4% N0 2+ nodes G3 54% 1+ node 21% G1 42% T2 N0 N1-3 T1 72% 33% 10% 5% 6% G2 64% G3 29% Er‒ PR‒ 2% HER2+ 8% • • Low proportion Less confidence in the results? Can we be sure that modern pathology is not also able to distinguish “indolent” luminal cancers from “aggressive” ones? Discordances between central immunohistochemical and molecular breast cancer subtyping in the MINDACT trial “luminal” tumors (N=4718) Molecular subtyping based on BluePrint and MammaPrint® Pathological subtyping Luminal A Luminal B ER+ and PgR ≥ 20% and HER2and Ki67 < 20% ER+ and PgR < 20% and/or Ki67 ≥ 20% and HER2- Can we be sure that adjuvant CT does not provide a benefit in the “discordant” groups? Efficacy: CT vs no CT in discordant risk groups Intent-to-treat analysis Distant Metastasis Free Survival c-Low/g-High Distant Metastasis Free Survival c-High/g-Low Allocated Treatment strategy CT no CT % at 5 Year(s) (95% CI) Hazard Ratio p-value (adjusted (adjusted Cox model) logrank) (95% CI) 95.9 0.78 (94.0, 97.2) (0.50,1.21) 94.4 (92.3, 95.9) 1.00 Allocated Treatment strategy CT 0.267 no CT % at 5 Year(s) (95% CI) 95.8 (92.9, 97.6) 95.0 (91.8, 97.0) Hazard Ratio p-value (adjusted Cox (adjusted model) logrank) (95% CI) 1.17 (0.59,2.28) 1.00 Allocated to: Allocated to: Small CTX benefit not ruled out… but ≤ 2% 0.657 No suggestion of any CT benefit Adjuvant therapy decision-making NEVER FORGET THAT… The « low-enough » risk that justifies treatment de-escalation… is a patient’s decision! « High risk » does not mean that the treatment will work! Practical use of MammaPrint® in the clinic Start by evaluating the “Clinical Risk” * Clinical risk “LOW” Clinical risk “HIGH” Treatment according to guidelines Discuss with patient if she would value a < 2% gain in DMFS with adjuvant CTX *Exclude women with T3 or N2─ 3+ ? NO YES Order MammaPrint® test Proceed with CTX In 48% of patients it will come back “low risk” allowing for adjuvant CTX sparing Are there ways to forgo GEPs??? Dr M de Block – Minister of Health More than 15 years of research… and still a lot of controversy regarding who can be spared adjuvant chemotherapy and who does not need extended adjuvant endocrine therapy!!! SYSTEMIC THERAPY FOR EARLY BC Gene Signatures assisting with decision making OnCoG GI oncoDNA /Myriad Most useful in Node negative disease Ribinikar D - ASCO Educational Book 2016 Proliferation genes drive the prognostic power of GEP signatures Original signature Proliferation genes Original signature Non-Proliferation genes Sotiriou C, et al. N Engl J Med. 2009; Wirapati P, et al. Breast Cancer Res. 2008; Sotiriou C, et al. Nat Rev Cancer. 2007 LOE 1A (Prospective Validation) MINDACT (MammaPrint) TAILORX (Oncotype Dx) RxPONDER (Oncotype Dx) ASTER70 (Genomic Grade) Reported Partial results reported Expected final results 2017 Expected results 2022 Expected results 2018 Adjuvant therapy decision-making NEVER FORGET THAT… The « history » of luminal BC extends beyond 5 years, with late relapses seen particularly if large T a/o N+ SABCS 2016 New data regarding MammaPrint® An ultra low risk cut-off identifies N- postmemopausal women with excellent 20-year outcomes Molecular definition of “Indolent” 70 gene Prognosis Signature: “Ultra-low Threshold” 70 significant prognosis genes Ultralow Threshold van´t Veer et al., Nature ,2002 Threshold derived from TRANSBIG with 25-year follow-up and no metastatic events In women with breast cancer and WITHOUT ANY SYSTEMIC THERAPY This presentation is the intellectual property of the author. Contact them at [email protected] permission to reprint and/or distribute. Validation in the STO 3 Trial postmenopausal women, <3cm, N0 tumors Stockholm: 1976-1991 This presentation is the intellectual property of the author. Contact them at [email protected] permission to reprint and/or distribute. All Patients by 70 Gene, Ultralow, low≠ultralow, high This presentation is the intellectual property of the author. Contact them at [email protected] permission to reprint and/or distribute. Intergroup TAILORx Trial (N=10,253) Trial Assigning IndividuaLized Options for Treatment Node Negative ER+ and/or PgR+ HER2 negative (IHC 0-1+ or FISH [-]) 21-Gene Recurrence Score Assay RS <11 Hormonal Therapy* 16% Results released in 2015 RS 11 - 25 Randomize Hormonal Rx* vs Chemotherapy* + Hormonal Rx* 67% RS >25 Chemotherapy* + Hormonal Rx* 17% *Choice of therapy at investigator discretion including option of adjuvant chemotherapy trials TAILORx: RS 0-10 N=1626 / 88 events at median follow-up of 69 months Rate of freedom from recurrence at distant site : 99.3% (95% ci: 98.7-99.6) Distant recurrences N=10 Locoregional recurrences N=8 Contralateral BC (invasive) N=15 Other primary cancers N=43 30 deaths (12 w/o cancer) Sparano J.A. et al, N Engl J Med, Nov 19, 2015 Stage (N) still matters Prognostic Signatures are not good enough if high tumor burden! Trans ATAC (Oncotype DX) NKI Validation (MammaPrint) LN- Dowsett et al., J Clin Oncol. 2010 LN+ Van de Vijver et al., NEJM 2002 SABCS 2016 (Sestak et al) Head to head comparison of several signatures and their added value beyond a clinical treatment score in Node ─ or Node + patients from ATAC treated with endocrine therapy only • Clinical treatment score: N status, T size, grade, age, therapy • Signatures: RS (Oncotype-DX), BCI, ROR score (Prosigna)*, EndoPredict Clinical** *incorporates T size ** incorporates T size and nodal status Head to head comparison of multigene prognostic signatures in the ATAC trial N = 818 women (591 N─, 227 N+) 10 year distant recurrence rate 5-10 year distant recurrence rate 29% 14% 10% 6% N─ N+ N─ N+ San Antonio Breast Cancer Symposium – December 6-10, 2016 Prognostic value years 0-10 – node-negative CTS 31,8 IHC4 30,6 43,8 BCI RS % Improvement 22,8 ROR EPclin 50,8 40,6 Likelihood Ratio χ2 IHC4 31.8 BCI 31.8 RS 31.8 ROR 31.8 EPclin 31.8 17,1 22,5 53.8% 70.8% 33.3% 10,6 23,7 15,2 74.5% 47.8% Likelihood Ratio ∆χ2 This presentation is the intellectual property of the author/presenter. Contact [email protected] for permission to reprint and/or distribute. San Antonio Breast Cancer Symposium – December 6-10, 2016 Prognostic value years 0-10 – node-positive IHC4 BCI RS ROR EPclin % Improvement 40,2 CTS 6,3 9,6 6,4 15,5 35,6 Likelihood Ratio χ2 IHC4 40.2 4,8 11.9% BCI 40.2 5,2 12.9% RS 40.2 5.0 12.3% ROR 40.2 6.0 14.9% EPclin 40.2 8,5 21.1% Likelihood Ratio ∆χ2 This presentation is the intellectual property of the author/presenter. Contact [email protected] for permission to reprint and/or distribute. Prognostic (± predictive) value of multigene signatures in years 0-10 • Node negative disease: all signatures show added value beyond the Clinical Treatment Score and identify 55 to 70% of patients as having a 10y risk of distant metastases ranging between 3 and 6.6% • Node positive disease: the added value of the signatures is inexistent or modest (ROR, EndoPredict Clinical) EndoPredict Clinical identifies 19% of patients with a 10y risk of distant mets below 6% San Antonio Breast Cancer Symposium – December 6-10, 2016 Prognostic value years 5-10 – node-negative IHC4 6,6 19,5 BCI RS ROR EPclin % Improvement 16,6 CTS 3,4 31,3 24.0 Likelihood Ratio χ2 IHC4 16.6 BCI 16.6 RS 16.6 ROR 16.6 EPclin 16.6 20.0% 3,3 11,2 67.5% 11.4% 1,9 18,4 10,3 111.0% 62.0% Likelihood Ratio ∆χ2 This presentation is the intellectual property of the author/presenter. Contact [email protected] for permission to reprint and/or distribute. San Antonio Breast Cancer Symposium – December 6-10, 2016 Prognostic value years 5-10 – node-positive % Improvement 16.0 CTS IHC4 1.0 IHC4 16.0 1,2 7.5% BCI 3,1 BCI 16.0 1,8 11.3% RS 1,1 RS 16.0 1.1 6.9% ROR 16.0 4.1 25.6% EPclin 16.0 ROR EPclin 7,3 14,9 Likelihood Ratio χ2 4,4 27.5% Likelihood Ratio ∆χ2 This presentation is the intellectual property of the author/presenter. Contact [email protected] for permission to reprint and/or distribute. Prognostic (± predictive) value of multigene signatures in years 5-10 • Node negative disease: added value is seen for BCI, ROR and EndoPredict Clinical. EndoPredict Clinical identifies 73.5% of patients with a risk of distant relapse of 4.3% • Node positive disease: a modest added value is seen for signatures that combine genomic with anatomic information. EndoPredict Clinical, for examples, identifies 22% of patients with a risk of distant relapse of 3.3% What about the “predictive” power of the gene signatures for: - CTX benefit - Endocrine therapy benefit ? High proliferative breast cancers may benefit from adjuvant chemotherapy CMF regimen LN- Paik et al., JCO 2006 CAF regimen LN+ Albain et al., Lancet Oncology 2009 Breast Cancer Index Algorithmic combination of proliferation related gene signature (Molecular Grade Index) and an estrogen signaling pathway signature (Hox B13/IL17 BR, H/I) Summary of BCI Predictive (H/I) Validation Data ▪ H/I shown to be a significant predictor of endocrine benefit in 3 randomized trial cohorts Study Cohort Treatment Stockholm (n=600)1 Adjuvant tamoxifen vs H/I High HR: 0.35 (0.19-0.65); p=0.0005 untreated H/I Low HR: 0.67 (0.36-1.24), p=0.2 0.003 TransATAC (n=665)2 Adjuvant anastrozole vs tamoxifen 0.004 MA.17 (n=249)3 Extended letrozole vs H/I High OR: 0.33 (0.15-0.73); p=0.006 placebo H/I Low OR: 0.58 (0.25-1.36), p=0.21 Predictive analysis Interaction P value H/I High HR: 0.51 (0.27-0.97); p=0.04 H/I Low HR: 1.33 (0.65-2.71), p=0.4 0.03 ▪ Results suggest generalizability as an endocrine response biomarker 1. Zhang Y, et al. Clin Cancer Res. 2013;19(15):4196-205. 2. Sgroi D, et al. Lancet Oncol. 2013 Oct;14(11):1067-76. 3. Sgroi et al, J Natl Cancer Inst. 2013;105:1036-1042 Which signature for which patient? Luminal BC patients, aged ≥ 40y, No or N1-3+ at diagnosis Low clinical risk as per MINDACT SIGNATURE High clinical risk as per MINDACT Patient willing to have CTX for very small benefit Patient unwilling to have CTX for small benefit SIGNATURE SIGNATURE - Prognosis within 5y most relevant (N─) Prognosis within 10y most relevant (N+) MammaPrint® Oncotype EndoPredict, BCI, ROR EndoPredict, BCI, ROR Not accepted by guidelines Which signature for which patient? Luminal BC patients, aged ≥ 40y, No or N1-3+ after 5y of endocrine therapy (and no CTX!) Node ─ Node + BCI, ROR, EndoPredict Clinical all useful Only ROR and EndoPredict Clinical add value but in a modest way EndoPredict Clinical identifies 73% of patients with a distant relapse risk of 4.3% EndoPredict Clinical identifies 22% of patients with a distant relapse risk of 3.3% Not accepted currently by guidelines TAILORING OF SYSTEMIC THERAPY FOR EARLY BC Multi Gene Signatures – Time to move to the next chapter Bardelli A et al Cancer Cell 10.1016/j.ccell.2017.01.002 BREAST Data Center Team BIG HeadquartersTeam Institut Jules Bordet Team BIG Executive Board 2014-2018 Backups EBCTCG, submitted (2017) EBCTCG, submitted (2017) EBCTCG, submitted (2017) EBCTCG, submitted (2017) EBCTCG, submitted (2017) Reclassification with molecular subtyping • Molecular Subtyping classified 54% as Luminal A among the Luminal B by Pathological Subtyping (PS). • MS classified 38% as Luminal (A and B) and 5% as Basaltype among the HER2+ by PS. • MS classified 5% as Luminal (A and B) among the TN cases by PS. This presentation is the intellectual property of the author/presenter. Contact them at [email protected] for permission to reprint and/or distribute. Adjuvant! Online for breast cancer (Updated version) used for a standardized approach to “Clinical Risk” P. Ravdin Interpretation of the ATAC multigene signature comparison study • All patients received endocrine treatment! • Node negative disease: all signatures show added value and perform well in years 0-10 and in years 5-10 • Node positive disease: the performance of the signatures is more modest and only the ones that integrate tumor burden are useful; more research here is needed…