End-points utilized in breast cancer - PowerPoint Presentation

1. End-points utilized in breast cancer clinical trials - DiscussionNovartis OncologyDiscussant - Arunava ChakravarttyStat4Onc Symposium , Hartford, CT Apr-26, 2019

2. DisclaimerAll views expressed in the slide represent the individual views of the presenter and should not be seen as representing the views of Novartis, its board or its shareholders.

3. OverviewDiscussion to be focused on the two presentationsNCI-Breast Cancer Steering Committee Working Group Report on Meaningful and Appropriate Endpoints for Clinical Trials in Metastatic Breast Cancer , Andrew D Seidman, MDTailoring endpoints in Later-phase Oncology Trials, Meredith M Regan , ScDNext few discussion slides will focus on some of the key messages and my perspectives on those as a statistician in the biopharmaceutical industry

4. Presentation 1 ( Seidman)Recommendations on mBC endpoints from NCI BC Steering Committee WGOS as a 1ry endpoint is preferred in more aggressive disease settings eg . Later lines of treatment(≥3L) , Triple Negative BC In less aggressive settings eg : early lines (1-2L), HER2+ , HR+ where there are existing therapies , PFS as primary or co-primary with OS is recommended.Discussion PointsWhile OS is the gold standard to assess the clinical benefit of any therapy , it may not be feasible due to multiple factors Long post progression survival requiring very large trials or very long follow-upA large gap between OS and PFS increases the risks of confounding and bias due to cross over to other therapies (even in cases when it is not planned in a study). Under certain assumptions OS benefit can be assessed after adjusting for such cross overs. Examples : IPCW based : Letrozole in Adjuvant BIG 1-98 (Colleoni et al 2011) , RPSFT based : Evorolimus in RCC (Korhonen et al 2012)Even in situations where PFS is acceptable as a Primary endpoint, additional considerations are requiredInterpretation of results can be impacted by measurement error and multiple sources of bias . Eg : Evaluation time bias, Reader evaluation bias and Attrition bias.Therefore additional supportive analyses required based on independent central Review. Audit based approaches are also accepted in metastatic setting (Amit et al 2011)Systemic differences in parallelism of PFS assessments can impact the final results. FDA guidance of Onco Endpoints recommends additional sensitivity analyses - Actual event and Backdating PFS event analyses to assess the impact of such biases.4

5. Presentation 1 ( Seidman)Recommendations on mBC endpoints from NCI BC Steering Committee WGHigher imbalances in toxicities may outweigh marginal gains in PFS or even OS.Additional endpoints may be explored that incorporate both eg Theraputic IndexDiscussion PointsCurrent regulatory environment has a strong focus on Benefit-risk assessment. ICH M4E(R2) now requires identification of key benefits and risks in the Clinical overview of any filing , including magnitude of effects.Examples : FDA benefit-risk grid , EMA effects table. Some of the challenges include evaluating non-overlapping toxicities across therapies on a common scale for any quantitative comparison.Some efforts in this direction include patient solicited utility functions – EQ-5D. ( Used in many QALY analyses for Reimbursement Discussions)Time to Treatment Failure (TTF) has also been used 2ry or exploratory endpoint to compare drugs that are expected differ in terms of discontinuation due to Adverse Events. Not recommended as an endpoint for regulatory approval due to its inability to distinguish efficacy from other variables. Can be useful for reimbursement discussions.5

6. Presentation 2 (Regan)Use of Surrogate endpoints (SE) in OncologyEvaluation of surrogacy requires correlation with true endpoint at patient level also at the level of treatment effect over a range of trials. Surrogate Threshold Effect (STE) : A minimum effect on SE to predict any benefit for the True Endpoint(TE).Discussion PointsSurrogacy between PFS and OS has been indication and setting dependent. Has been established in Advanced CRC (Buyse et al 2007) and Adv Ovarian (Buyse et al 2010) for cytotoxics. Yet to be established in Adv BC.However in most regulatory standpoint it is accepted that PFS as a precursor to death in cancer patients is likely to predict survival (Chakravarty, Sridhara 2008)Reliability of PFS as a surrogate is dependent on its magnitude of benefit (Fleming 2009). Similar thought process as STE.STE is an interesting idea. Not seen much use in clinical trials. May useful as a decision making thresholds at interim based on short term endpoints to predict final outcome on long term clinical endpoint in adaptive designs.As newer class and MoA of drugs get developed historical results of surrogacy may not be always applicable in newer therapies even in the same indication. Recent examples in I/O have shown that due to pseudo-progression trials may show positive OS despite a not so strong RECIST based PFS HR (Buyse et al 2018), indicating some other intermediate endpoint may turn out to be a better surrogate in the future (i-RECIST) 6

7. Presentation 2(Regan)Alternative endpointsMilestone Survival, RMSTDuration of ResponseMethods based on Partitioning of overall survival : QTWiST , Treatment Free SurvivalDiscussion PointsMilestone Survival and RMST have been increasingly discussed in the context of Delayed treatment effect seen under Immuno Oncology (IO) compoundsUnder scenarios of delayed effect they serve as alternative metrics to the Hazard Ratio to assess treatment benefit as the proportional hazards assumption no longer holds.Some Challenges : (1) De-links estimation of treatment benefit from the log-rank test of hypothesis. (2) It has been shown that RMST based tests have even lower power compared Log rank based tests under delayed effect (Tian et al 2016)Endpoints like DoR , Trt. Free Survival are clinically appealing and intuitive endpoints. Can still be used as 2ry endpoints as descritptive measures of treatment benefit. DoR has often been used in registration of single arm trials that under a fast track Subpart (H) scenario eg (initial approval of some ICI’s in PD-1+ patients based on ORR)One of the caveats is that these suffer from non-randomized comparisons which can be biased and confounded under imbalances of response rates/ PFS between the armsPFS-2 : Time from randomization to progression on subsequent therapy . In the absence of OS , PFS2 has been considered as a important endpoint to assess the clinical benefit on subsequent therapies. Recommended in EMA Oncology Guidance. 7

8. Final PFSInterimSome additional thoughts – Multiple EndpointsAn example of OS/PFS Testing Strategy in HR+/HER2- Adv BCPFS – 1ry endpoint, OS – key 2ry endpointHierarhical Testing Strategy to control Type-1 error on bothMultiple endpoints – Multiple Interims (Glimm et al 2011)Group Sequential design with pre-specified alpha spending on both endpointsBusiness Use Only8PFSSigNonsigOSPFSSigOSOSFinal OS

9. Some additional thoughts – Alternative endpoints/analysesAn unifying framework between the objective and endpoint in a trial is the estimand of interest.An Estimand defines in detail what needs to be estimated to address a specific scientific question of interest (ICH E9 R1 addendum, CHMP 2017)Includes four key attributes : (1) Population under investigation (2) endpoint to be obtained for each patient to address the scientific question (3) specification of handling intercurrent events (4) a population based summary as a basis of treatment comparison(estimator)Many of the alternative endpoints can been seen different options of handling the intercurrent events in relation to OS eg : Post treatment toxicities, Subsequent therapies , Disease RelapseThese could assessed under alternate composite and/or prinicipal stratum (not same as subgroup analyses) estimand strategiesActive area of research to develop causal inference methods under such scenarios.9

10. ReferencesColleoni M, Giobbie-Hurder A, Regan MM, Thürlimann B, Mouridsen H, Mauriac L, Forbes JF, Paridaens R, Láng I, Smith I, Chirgwin J, Pienkowski T, Wardley A, Price KN, Gelber RD, Coates AS, Goldhirsch A. Analyses adjusting for selective crossover show improved overall survival with adjuvant letrozole compared with tamoxifen in the BIG 1-98 study. J Clin Oncol. 2011 Mar 20;29(9):1117-24.Korhonen P, Zuber E, Branson M, Hollaender N, Yateman N, Katiskalahti T, Lebwohl D, Haas T. Correcting overall survival for the impact of crossover via a rank-preserving structural failure time (RPSFT) model in the RECORD-1 trial of everolimus in metastatic renal-cell carcinoma. J Biopharm Stat. 2012;22(6):1258-71Fleming TR, Rothmann MD, Lu HL. Issues in using progression-free survival when evaluating oncology products. J Clin Oncol. 2009 Jun 10;27(17):2874-80. doi: 10.1200/JCO.2008.20.4107. Epub 2009 May 4ICH E9 working group (2017). ICH E9 (R1): addendum on estimands and sensitivity analysis in clinical trials to the guideline on statistical principles for clinical trialsAmit O et al. Blinded independent central review of progression in cancer clinical trials: results from a meta-analysis. European Journal of Cancer 2011; 47: 1772-1778.Guidance for Industry - Clinical Trial Endpoints for the Approval of Cancer Drugs and Biologics. FDA (2007). CHMP (2012). Appendix 1 to the guideline on the evaluation of anticancer medicinal products in man. Methodological consideration for using progression-free survival (PFS) or disease-free survival (DFS) in confirmatory trials Health Canada: Issue analysis summary: The use of PFS as the efficay endpoint for approval of targeted and chemotherapy agents for advanced cancer (2007)Chakravarty A + Sridhara R (2008). Use of PFS as a surrogate marker in oncology trials: some regulatory issues. Statistical Methods in Medical Research.Buyse M, Burzykowski T, Saad ED. The search for surrogate endpoints for immunotherapy trials. Ann Transl Med. 2018;6(11):231. doi:10.21037/atm.2018.05.16Buyse M, Burzykowski T, Carroll K, Michiels S, Sargent DJ, Miller LL, et al. Progression-free survival is a surrogate for survival in advanced colorectal cancer. J Clin Oncol. 2007;25(33):5218–24Buyse M, Burzykowski T, Carroll K, Michiels S, Sargent DJ, Miller LL, et al. Progression-free survival is a surrogate for survival in advanced colorectal cancer. J Clin Oncol. 2007;25(33):5218–24.Shi Q, de Gramont A, Grothey A, Zalcberg J, Chibaudel B, Schmoll HJ, et al. Individual patient data analysis of progression-free survival versus overall survival as a first-line end point for metastatic colorectal cancer in modern randomized trials: findings from the analysis and research in cancers of the digestive system database. J Clin Oncol. 2015;33(1):22–8.10