Hormone therapyduration: Can weselectthosepatientswho benefitfromtreatmentextension?

Hormone therapyduration: Can weselectthosepatientswho benefitfromtreatmentextension? Ivana Sestak, PhD Centre for Cancer Prevention Wolfson Institute of Preventive Medicine Queen Mary University London London, United Kingdom

ER-positive breast cancer Lancet 2011; 378: 771 84, Early Breast Cancer Trialist Collaborative Group

Pan et al., J ClinOncol34, 2016 (suppl; abstr505) ER-positive breast cancer II

Time patterns Subtrinsic subtypes MammaPrint Oncotype Dx Jatoi et al. JCO 2011;29:2301-2304

Extended endocrine therapy trials Trial Inclusion Extended treatment Median follow-up Event rate per arm (%) ATLAS Pre-and postmenopausal, any ER-status, any nodal status TAM 5 years vs. no treatment >10 years 17.9% vs. 20.8% significant benefit (DFS andos) attom Pre-and postmenopausal, any ER-status, any nodal status TAM 5 years vs. no treatment 8.6 years 12.9% vs. 13.0% NOsignificant benefit (RFS) MA.17 Postmenopausal, ER+, T1-T3, any nodal status LET 5 years vs. placebo 2.5 years 6.3% vs. 9.1% significant benefit (DFS andos) ABCSG-6a Postmenopausal, ER+, T1-T3, any nodal status ANA3 years vs. placebo 5 years 7.8% vs. 12.2% significant benefit (DFS) NSABP-B42 Postmenopausal, ER+, any nodal status LET 5 years vs. placebo 5 years 15.3% vs. 18.7% NOsignificant benefit (DFS and OS) IDEAL Postmenopausal, ER+, any nodal status LET 5 years Vs. LET 2.5 years 5 years 11.6% vs. 12.1% NOsignificant benefit (DFS and OS) DATA Postmenopausal, ER+, any nodal status ANA 6 years vs. ANA3 years 5 years 9.3% vs. 11.1% NOsignificant benefit (DFS and OS)

The issue in ER+ breast cancer Substantial continuing risk of recurrence for at least 15 years after 5 years of endocrine therapy The risk may be reduced by extending therapy (toxicity) Which patients have sufficient residual risk to potentially benefit? Which patients have very low risk of late recurrence where extended therapy can be omitted?

Clinicopathological parameters

Battle of the genomic assays Clinical Treatment Score/IHC4 Mammaprint Rotterdam Signatures Genomic Grade Index Oncotype Dx RS Breast Cancer Index (BCI) Prosigna PAM50 EndoPredict(EPclin).. PAM50 EPclin BCI RS Adapted from E. Winer, ASCO 2013

RS late relapse B-28 B-14 Wolmark et al., JCO, 2016; 34(20): 2350-2358.

BCI late relapse 1 Sgroi DC, et al. Lancet Oncol. 2013;14:1067-76. 2 Zhang Y, et al. Clin Cancer Res. 2013;19:4196-4205.

ROR late relapse Distant recurrence-free survival (%) 100 90 80 70 0 low intermediate high 10-yr DRFS (95% CI) 98.7 (96.9-99.5) 95.2 (92.3-97.0) 91.5 (87.8-94.1) 15-yr DRFS (95% CI) 97.6 (94.7-98.9) 90.9 (85.9-94.2) 82.5 (74.8-88.1) 5 6 7 8 9 10 11 12 13 14 15 Follow-up time (years) Patients at risk low 460 447 439 412 331 250 188 125 81 50 25 intermediate high 416 370 400 347 387 330 370 301 289 238 220 198 161 153 109 119 71 82 41 43 14 24 Filipitset al., 2014, Clin. Cancer Res; 20(5):1298-1305

EPclinlate relapse 1.8% (0.15-3.5) 12.3 % (7.5-17.1) Dubsky et al., 2013, Br J Cancer; 109(12): 2959-2964.

Individual studies late relapse Signature RS BCI ROR EPclin Early relapse Late relapse Validation studies NSABP-B20 transatac SWOG S8814 transatac Stockholm TAM transatac ABCSG-8 transatac ABCSG-8 ABCSG-6 Validation in different patient population, no head-to-head comparison in same dataset!

Multigeneassays Need to show accurate prediction of late recurrence Rate of late recurrence can be reduced by extended therapy Need to demonstrate independent prognostic value above and beyond clinical parameters Combined model (molecular and clinical) might be most helpful in the decision making process

Comparison of prognostic signatures in transatac Aims: 1. Prognostic performance of six signatures for distant recurrence in N- and N+ separately in transatac in absence of chemotherapy: In years 5-10 (extended endocrine therapy) 2. Added prognostic value of signatures to clinical variables Sestak et al., 2017, Cancer Res; 77(4): S6-05

transatac: Prognostic signatures Signature Clinical Treatment Score (CTS) Information included Nodal status, grade, tumour size, age, treatment Immunohistochemical markers (IHC4) ER, PgR, Ki67, HER2 Oncotype Recurrence Score (RS) 21 genes (oestrogen, proliferation,invasion, HER2 genes) Breast Cancer Index(BCI) H/I and 5 proliferation genes (MolecularGrade Index) Prosigna(ROR) 46 genes, proliferation score, tumour size (EU cut-offs from transatacfor N-and N+) EndoPredict(EPclin) 12 genes (proliferation, differentiation,oestrogen); nodal status and tumour size Sestak et al., 2017, Cancer Res; 77(4): S6-05

5-10 years: Node-negative CTS 16,6 % Improvement IHC4 6,6 IHC4 16.6 3,3 20.0% BCI 19,5 BCI 16.6 11,2 67.5% RS 3,4 RS 16.6 1,9 11.4% ROR 31,3 ROR 16.6 18,4 111.0% EPclin 24.0 EPclin 16.6 10,3 62.0% Likelihood Ratio χ 2 Likelihood Ratio χ 2 Sestak et al., 2017, Cancer Res; 77(4): S6-05

DR risk (%): Node-negative Distant recurrence free (%) 80 90 100 80 90 100 80 90 100 80 90 100 BCI RS ROR EPclin 5 6 7 8 9 10 Follow-up time [years] DR risk % 2.5% 14.4% 15.9% 1.4% 10.0% 23.2% 4.3% 14.6% Patients % in risk groups 63.6% patients 23.6% patients 12.9% patients 4.8% 65.6% patients 9.6% 25.1% patients 16.1% 9.4% patients 54.6% patients 30.8% patients 14.6% patients 73.5% patients 26.5% patients Sestak et al., 2017, Cancer Res; 77(4): S6-05

5-10 years: Node-positive CTS 16.0 % Improvement 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 7,3 ROR 16.0 4.1 25.6% EPclin 14,9 EPclin 16.0 4,4 27.5% Likelihood Ratio χ 2 Likelihood Ratio χ 2 Sestak et al., 2017, Cancer Res; 77(4): S6-05

Distant recurrence free (%) 60 80 100 60 80 100 60 80 100 60 80 100 DR risk (%): Node-positive BCI RS ROR EPclin 5 6 7 8 9 10 Follow-up time [years] DR risk % 14.3% 19.7% 36.5% 0% 13.0% 25.0% 3.3% 23.6% Patients % in risk groups 51.7% patients 32.4% patients 15.9% patients 17.9% 61.0% patients 19.5% 30.2% patients 27.5% 8.8% patients 8.2% patients 28.0% patients 63.7% patients 22.0% patients 78.0% patients Sestak et al., 2017, Cancer Res; 77(4): S6-05

transataccohort Strengths: Mature clinical trial with over 10 years of follow-up Well annotated samples AI as well as tamoxifen Comprehensive molecular profiles using commercial tests Weaknesses: A single study For molecular studies beyond 5 years modest sample size CTS trained in transatac

Clinical risk groups Grade Nodal status Tumour size (mm) Clinical risk group 1 N- <=30 Low 30-50 High N1-3 <=20 Low 20-50 High 2 N- <=20 Low 20-50 High N1-3 Any size High 3 N- <=10 Low 689 transatac patients with HER2-negative disease 395 (57%) low risk 294 (43%) high risk 10-50 High N1-3 Any size High

Likelihood Ratio χ2 Prognostic performance by clinical risk: 5-10 years Clinical low risk (N=395) Clinical high risk (N=294) 18 16,8 16 14 14 12 10 9 8 6 6 4,2 4 5,3 2 1,5 0,7 0 BCI RS ROR Epclin BCI RS ROR Epclin HRs 3.6 1.5 3.8 2.6 1.4 1.1 1.5 1.7

BCI Clinical low risk (N=395) #(DR) 10-year DR risk (%) Low 291 (5) 1.5% (0.6-4.0) Intermediate 79 (6) 6.8% (2.9-15.5) High 25 (3) 13.0% (4.4-35.2) RS Low 293 (5) 1.9% (0.8-4.4) Intermediate 83 (6) 6.4% (2.7-14.8) High 19 (3) 11.1%(2.9-37.6) ROR Low 260 (3) 0.8% (0.2-3.1) Intermediate 113 (7) 6.0% (2.7-12.8) High 22 (4) 22.2% (9.0-48.9) EPclin DR risk by clinical groups Low 337 (10) 2.6% (1.3-5.1) High 58 (4) 7.5% (2.9-18.8)

DR risk by clinical groups BCI Clinical low risk (N=395) #(DR) 10-year DR risk (%) Low 291 (5) 1.5% (0.6-4.0) Intermediate 79 (6) 6.8% (2.9-15.5) High 25 (3) 13.0% (4.4-35.2) RS Low 293 (5) 1.9% (0.8-4.4) Intermediate 83 (6) 6.4% (2.7-14.8) High 19 (3) 11.1%(2.9-37.6) ROR Low 260 (3) 0.8% (0.2-3.1) Intermediate 113 (7) 6.0% (2.7-12.8) High 22 (4) 22.2% (9.0-48.9) EPclin Low 337 (10) 2.6% (1.3-5.1) High 58 (4) 7.5% (2.9-18.8) Clinical high risk (N=294) # (DR) 10-year DR risk (%) 133 (12) 9.3% (5.3-16.2) 97 (20) 24.1%(16.2-35.0) 64 (12) 20.6% (12.2-33.5) 152 (23) 17.3% (11.7-25.3) 96 (13) 14.2% (8.5-23.2) 46 (8) 20.1% (10.5-36.4) 47 (2) 4.8% (1.2-17.7) 103 (14) 15.9%(9.6-25.6) 144 (28) 21.4% (15.2-29.8) 96 (8) 10.7% (5.4-20.7) 198 (36) 19.8% (14.6-26.5)

Summary I Clinical parameters (nodal status and size) important for prediction of late recurrence ROR, BCI, EPclin: Provide additional prognostic information concerning the likelihood of recurrence >5 years after surgery Node-negative only ROR/EPclinscore show highest level of evidence for late recurrence as validated in large phase III biomarker samples/studies (Level I, Simon et al. JNCI, 2009; 101(21):1446-1452)

Summary II Low risk groups by any assay are very unlikely to benefit from extended endocrine therapy In clinical low risk group High risk groups by ROR and EPclingood candidates for extended endocrine therapy Irrespective of clinical risk group Incorporation of certain clinical variables important Combined model best for prediction of late recurrence Guidelines do not support use of multigeneassays for predication of late distant recurrence

Future directions Specific signatures for predicting late recurrence needed Individual gene analysis for late recurrence (PAM50 and EP) transatac Development of CTSP5 transatac/big1-98 Circulating tumour cells, ctdna, microrna In development and clinical utility needs to be established Outstanding issues: TNBC Premenopausal women

Acknowledgment ATAC patients TransATAC investigators and pathologists LATTE Steering Committee