PrECOG 0102: A Randomized, Double-Blind Phase II Trial of Fulvestrant plus Everolimus or Placebo in Post-Menopausal Women with Hormone-Receptor Positive, HER2-Negative Metastatic Breast Cancer Resistant to Aromatase Inhibitor Therapy Noah S Kornblum, MD1, Judith Manola, MS2, Paula Klein, MD3, Bhuvaneswari Ramaswamy, MD4,Adam Brufsky, MD PhD5, Phillip J Stella, MD6, Brian Burnette, MD7, Melinda Telli, MD8, Della F Makower, MD1,Joseph Leach, MD9, Cristina I Truica, MD10, Antonio C Wolff, MD11, Gamini S Soori, MD12, Barbara Haley, MD13, Arun Nagarajan, MD14, Timothy R Wassenaar, MD15, Lori Goldstein, MD16, Kathy D Miller, MD17,and Joseph A Sparano, MD1 Institutions: 1Montefiore-Einstein Center for Cancer Care, Bronx, New York, United States, 10461; 2Dana-Farber Cancer Institute, Boston, MA, United States, 02284-9168; 3Mount Sinai Beth Israel Comprehensive Cancer Center, New York, New York, United States, 10011; 4Ohio State University Comprehensive Cancer Center, Columbus, OH, United States, 43212; 5University of Pittsburgh, Pittsburgh, PA, United States, 15213; 6Saint Joseph Mercy (Michigan Cancer Consortium), Ann Arbor, MI, United States, 48106; 7Saint Vincent Hospital, Green Bay, WI, United States, 54301; 8Stanford University Medical Center, Stanford, CA, United States, 95304; 9Metro-Minnesota Community Oncology Research Consortium, Saint Louis Park, MN, United States, 55416; 10Penn State Hershey Cancer Institute, Hershey, PA, United States, 17033; 11Johns Hopkins Sidney Kimmel Comprehensive Cancer Center, Baltimore, MD, United States, 21287; 12Missouri Valley Cancer Consortium, Omaha, NE, United States, 68106; 13UT Southwestern Medical Center, Dallas, TX, United States, 75390; 14CAMC Health System, Charleston, WV, United States, 25304; 15Pro Health Care, Waukesha, WI, United States, 53188, 16Fox Chase Cancer Center, Philidelphia, PA, United States, 19111 and 17Indiana University School of Medicine, Indianapolis, IN, United States, 46202-5689. I am grateful for the opportunity to present our results on behalf of all the investigators and individuals involved in PreCOOG 0102. This is a randomized, double blind, Phase II trial of Fulvestrant with everolimus or placebo for the treatment of hormone-receptor positive, HER2 negative Metastatic breast cancer resistant to aromatase inhibitor therapy. This presentation is the intellectual property of the author. Contact nkornblu@montefiore.org for permission to reprint and/or distribute.

San Antonio Breast Cancer Symposium, December 6-10, 2016 Background: Resistance to Aromatase Inhibitor (AI) Therapy is Inevitable Strategies to address resistance are emerging Targeting PI3K-AKT-mTOR pathway 1 Addition of the mTOR inhibitor Everolimus to the steroidal AI exemestane improved median PFS (3.2 vs. 7.8 mo., p<0.0001) in the phase III BOLERO-2 trial Use of a selective estrogen receptor down regulator (SERD) More complete blockade of ER signaling than other anti-estrogens Hypothesis: The combination of everoliumus/fulvestrant would be more effective than fulvestrant alone in AI resistant disease While aromatase inhibitor therapy is highly effective treatment for hormone receptor positive metastatic breast cancer, resistance to AI therapy ultimately develops in most One, strategy for addressing AI resistance is by targeting the PI3K-AKT-mTOR pathway For example, combining the mTOR inhibitor Everolimus with the steroidal AI exemestane significantly improved median PFS from 2.8 to 6.9 months compared with exemestane alone in the BOLERO 2 trial, although overall survival was not improved The addition of everolimus to tamoxifen in AI resistant breast cancer also improved clinical outcome in the randomized phase II TAMRAD study Another strategy for overcoming AI resistance is by more completely blocking ER signaling through use of an selective estrogen receptor downreguator, which may result in more complete blockade of the ER signaling pathway than a a steroidal AI such as exemesstne For example, although fulvestrant given at a dose of 250 mg monthly was not more effective than exemestane in the EFECT trial, the 500 mg monthly dose proved to be more effective than the 250 mg dose, suggesting that the higher fulvestrant dose may be more effective antiestrogen therapy than exemestane in AI resistant disease We therefore hypothesized that the combination of everoliumus and fulvestrant would be a more effective combination than fulvestrant alone in AI resistant disease 1Ma, C, et al. Nature Reviews Cancer 15, 261–275 (2015) ; 2 Baselga, J. et al, New Engl. J. Med. 366, 520–529 (2012) ; 3Yardley, D. et al, Adv Ther (2013) 30:870–884; 4 Piccart, M, et al. Ann Oncol 25, 12, 2357-2362 (2014) ; 5Bachelot, T. et al. J Clin Oncol. 30, 2718–2724 (2012); 6 Chia, S., et al. J Clin Oncol 26:1664-70 (2009).; 7Di Leo, A et al. J Clin Oncol. 28, 4594-4600 (2010) This presentation is the intellectual property of the author. Contact nkornblu@montefiore.org for permission to reprint and/or distribute.

Methods: Key Eligibility Criteria San Antonio Breast Cancer Symposium, December 6-10, 2016 Methods: Key Eligibility Criteria Post-menopausal women HR-positive, HER2-negative (ASCO-CAP) Inoperable locally advanced or metastatic breast cancer AI resistant disease: Relapse while receiving adjuvant AI therapy Progression after one or more AIs for metastatic disease ECOG PS 0-1 Normal organ function < 1 prior chemotherapy regimen for metastasis Measurable and/or non-measurable disease (RECIST 1.1) 2 doses of fulvestrant permitted within 28d prior to randomization Key eligibility requirements for study participation are listed here and include: Post-menopausal women HR-positive, HER2-negative, Inoperable locally advanced or metastatic breast cancer AI resistant disease: which was defined as Relapse while receiving adjuvant AI therapy Progression after one or more AIs for metastatic disease ECOG PS 0-1 Normal organ function No more than 1 prior chemotherapy regimen for metastatic disease was allowed Measurable and/or non-measurable disease (RECIST 1.1) criteria 2 doses of fulvestrant were permitted within 28d prior to randomization This presentation is the intellectual property of the author. Contact nkornblu@montefiore.org for permission to reprint and/or distribute.

Methods: Study Schema Induction Phase Continuation Phase San Antonio Breast Cancer Symposium, December 6-10, 2016 Methods: Study Schema Prior Chemo for Mets (y/n) Measurable Disease (y/n) ECOG PS: 0 vs. 1 1:1 Randomization Stratify: Fulvestrant 500 mg day 1 and 15 of cycle 1, then day 1 of cycles 2-12 (28d cycles) Everolimus 10 mg PO QD Arm A Arm A & B (week 48) Unblind and continue Fulvestrant +/- Everolimus Placebo PO QD Arm B Induction Phase Continuation Phase Induction Phase: Treat until evidence of progressive disease or unacceptable toxicity for a maximum of 12 cycles (48 weeks) Continuation Phase: If no disease progression or unacceptable toxicity after 12 cycles, unblind and continue fulvestrant +/- everolimus The study schema is shown here: Patients were stratified by ECOG performance status, presence of measurable disease, and prior chemotherapy status They were then randomized 1:1 for treatment with high dose fulvestrant with either everolimus or matching placebo In the induction phase patients received their assigned treatments until evidence of POD or unacceptable toxicity, for a maximum of 12 cycles (48 weeks) Patients who did not experience Progression or unacceptable toxicity were un-blinded and allowed to proceed to the continuation phase of the study and remain on fulvestrant +/- everolimus Tumor measurements were performed every 12 weeks (+/- 1 week) Of note, Corticosteroid mouthwash prophylaxis was not used Treatment Plan: Tumor measurements every 12 weeks (+/- 1 week) by local treating physician Supportive Care: Corticosteroid mouthwash prophylaxis was not used This presentation is the intellectual property of the author. Contact nkornblu@montefiore.org for permission to reprint and/or distribute.

Methods: Statistical Design San Antonio Breast Cancer Symposium, December 6-10, 2016 Methods: Statistical Design Primary endpoint: progression-free survival (by investigator assessment) 130 total patients (120 eligible assuming 10% ineligible) 90% power 1-sided Type I error 10% 70% improvement in median PFS - 5.4 mo. (CONFIRM)1 to 9.2 mo. Stratified log-rank test ITT analysis Full information = 98 events (death or progression) The statistical design employed is described here The primary endpoint of the trial is progression free survival as assessed by the treating physician We calculated a sample size of 130 total patients (120 eligible assuming ~ 10% ineligible rate) would be required for the study To have a 90% power using a 1-sided Type 1 error of 10% to detect a 70% improvement in median PFS from 5.4 months (the historical median from the CONFIRM trial) to 9.2 months Comparison was made by stratified log-rank test And an intention to treat analysis including all randomized patients was performed We estimate that full information would be available after = 98 events (death or progression) 1Di Leo, A et al. J Clin Oncol. 28, 4594-4600 (2010) This presentation is the intellectual property of the author. Contact nkornblu@montefiore.org for permission to reprint and/or distribute.

San Antonio Breast Cancer Symposium, December 6-10, 2016 Progression Free Survival (by investigator assessment – primary study endpoint) Here we see the Kaplan-Meier curve of progression free survival as assessed by the treating physician, the primary study endpoint There were 101 events at the time of the analysis, which exceed the minimum of 98 events required to trigger the analysis. The addition of everolimus improved median PFS from 5.1 months to 10.4 months The hazard ratio for PFS was 0.60 favoring the everolimus arm, with 95% confidence intervals of 0.40 and 0.92 The stratified long rank p value was 0.02, indicating that the primary trial endpoint was met This presentation is the intellectual property of the author. Contact nkornblu@montefiore.org for permission to reprint and/or distribute.

Conclusions Addition of everolimus to fulvestrant improved PFS San Antonio Breast Cancer Symposium, December 6-10, 2016 Conclusions Addition of everolimus to fulvestrant improved PFS median PFS 5.1 vs. 10.4 months HR 0.60, p=0.02 Associated with more toxicity, including Grade 3 adverse events: 48% (F/E) vs. 14% (F/P) Most common grade 3 A.E.s occurring in > 5% included stomatitis (9%), pneumonitis (6%), fatigue (5%), & hyperglycemia (6%) Saftey profile consistent with everolimus in BOLERO-21 Prophylactic corticosteroid mouthwash was not used, which has been shown to reduce risk of grade 1-2 stomatitis from about 65% to 20%2 Provides additional evidence that adding everolimus to anti-estrogen therapy in AI resistant disease improves clinical outcomes Addition of everolimus to fulvestrant improved PFS median PFS 5.1 vs. 10.4 months HR 0.60, p=0.02 Associated with more toxicity, including Grade 3 adverse events: 48% (F/E) vs. 14% (F/P) Most common grade 3 A.E.s occurring in > 5% included stomatitis (9%), pneumonitis (6%), fatigue (5%), & hyperglycemia (6%) Toxicity profile consistent with everolimus in BOLERO-21 Prophylactic corticosteroid mouthwash was not used, which has been shown to reduce risk of grade 1-2 stomatitis from about 65% to 20%2 This study provides additional evidence that adding everolimus to anti-estrogen therapy in AI resistant disease improves clinical outcomes 1 Baselga, J. et al, New Engl. J. Med. 366, 520–529 (2012) 2 Rugo, H, et al. J Clin Oncol 34, 2016 (suppl; abstr 525) This presentation is the intellectual property of the author. Contact nkornblu@montefiore.org for permission to reprint and/or distribute.