Restricted Mean Survival Time and Cure-Rate Modeling in Estimating Relapse-Free Survival Benefit With Adjuvant Dabrafenib + Trametinib Treatment in Melanoma John M. Kirkwood,1 Reinhard Dummer,2 Axel Hauschild,3 Mario Santinami,4 Victoria Atkinson,5 Vanna Chiarion Sileni,6 James Larkin,7 Marta Nyakas,8 Andrew Haydon,9 Caroline Dutriaux,10 Jacob Schachter,11 Caroline Robert,12 Laurent Mortier,13 Hiya Banerjee,14 Tomas Haas,15 Monique Tan,16 Mike Lau,17 Dirk Schadendorf,18 Georgina V. Long,19 Mario Mandalà20 1Melanoma Program, UPMC Hillman Cancer Center, University of Pittsburgh, Pittsburgh, PA, USA; 2Department of Dermatology, University Hospital Zürich Skin Cancer Center, Zürich, Switzerland; 3Department of Dermatology, University Hospital Schleswig-Holstein, Kiel, Germany; 4Melanoma and Sarcoma Unit, Department of Surgery, Fondazione IRCCS Istituto Nazionale dei Tumori, Milano, Italy; 5Division of Cancer Services, Princess Alexandra Hospital, Gallipoli Medical Research Foundation, University of Queensland, Greenslopes, QLD, Australia; 6Melanoma Cancer Unit, Department of Experimental and Clinical Oncology, Veneto Oncology Institute-IRCCS, Padova, Italy; 7Department of Medical Oncology, Royal Marsden NHS Foundation Trust, London, UK; 8Department of Cancer, Oslo University Hospital, The Norwegian Radium Hospital, Oslo, Norway; 9Melanoma Service, Department of Medical Oncology, The Alfred Hospital, Melbourne, VIC, Australia; 10Service de Dermatologie, Centre Hospitalier Universitaire de Bordeaux, Hôpital Saint-André, Bordeaux, France; 11Division of Oncology, Ella Lemelbaum Institute for Immuno-Oncology and Melanoma, Sheba Medical Center, Tel Hashomer, and Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel; 12Dermatology Service and Melanoma Research Unit, Gustave Roussy and Paris-Sud-Paris-Saclay University, Villejuif, France; 13Service de Dermatologie, Université de Lille, INSERM U 1189, Lille, France; 14Clinical Development and Analytics, Novartis Pharmaceuticals Corporation, East Hanover, NJ, USA; 15Clinical Development and Analytics, Novartis Pharma AG, Basel, Switzerland; 16Oncology Clinical Development, Novartis Pharmaceuticals Corporation, East Hanover, NJ, USA; 17Global Medical Affairs, Novartis Pharma AG, Basel, Switzerland; 18Department of Dermatology, Comprehensive Cancer Center (Westdeutsches Tumorzentrum), University Hospital Essen, Essen, and German Cancer Consortium, Heidelberg, Germany; 19Department of Medical Oncology, Melanoma Institute Australia, The University of Sydney, and Royal North Shore and Mater Hospitals, Sydney, NSW, Australia; 20Department of Oncology and Hematology, Papa Giovanni XXIII Cancer Center Hospital, Bergamo, Italy Background • In COMBI-AD analysis, 5-year relapse-free survival (RFS) rates were 52% in patients with stage III BRAF V600– mutant melanoma who received ≤ 12 months of dabrafenib + trametinib compared with 36% in those who received placebo1 • Kaplan-Meier and Cox regression analyses have been used to assess adjuvant treatment effects based on time-to-event analyses1-3 • Unfortunately, these statistical methods do not account for nonproportional hazards and the fact that some patients never experience relapse4-8 • To overcome these limitations, we evaluated treatment effects in COMBI-AD using: – Restricted mean survival time (RMST): population average for the length of event-free survival time estimated by the area under a survival curve up to a specified time point that accounts for nonproportional hazards4-7 – Cure-rate modeling: a statistical approach to model time- to-event data that estimates the proportion of patients in each arm who may never experience an event of interest (eg, relapse)8,9 Methods • COMBI-AD (NCT01682083) is a double-blind, randomized, Phase III trial that compared 12 months of adjuvant dabrafenib 150 mg twice daily + trametinib 2 mg once daily vs 2 matched placebos in patients with resected stage III BRAF V600E/K–mutant melanoma (Figure 1) – Patients were stratified by BRAF V600E or V600K status and disease stage (by AJCC 7 criteria) • RMST analysis – The length of event-free survival time (ie, RFS) was estimated by assessing the area under the Kaplan-Meier curve up to 60 months of follow-up time in each treatment arm • Cure-rate modeling – The use of a cure-rate model is considered appropriate because it is reasonable to assume there is a subset of patients in each disease substage who are “cured” by resection alone and will remain relapse free. This assumption is supported by the appearance of a plateau in RFS Kaplan-Meier curves – A mixed Weibull cure-rate model was used to estimate the proportion of patients who might never experience disease relapse Figure 1. COMBI-AD Study Design Key eligibility criteria • Completely resected cutaneous melanoma • BRAF V600E/K mutant • Stage IIIA, IIIB, or IIIC (AJCC 7) • Resection ≤ 12 weeks before randomization • No prior systemic anticancer therapy • ECOG PS 0-1 N = 870 Stratified by: • BRAF mutation (V600E or V600K) • Disease stage (IIIA, IIIB, or IIIC) R 1:1 n = 438 Dabrafenib 150 mg BID + Trametinib 2 mg QD n = 432 2 matched placebos Treatment (12 months)a Primary endpoint: RFS Secondary endpoints: OS, DMFS, FFR, safety Follow-upb until the end of study AJCC, American Joint Committee on Cancer; AJCC 7, AJCC Cancer Staging Manual, 7th edition; BID, twice daily; DMFS, distant metastasis–free survival; ECOG PS, Eastern Cooperative Oncology Group performance status; FFR, freedom from relapse; OS, overall survival; QD, once daily; R, randomization; RFS, relapse-free survival. a Or until disease recurrence, death, unacceptable toxicity, or withdrawal of consent. b Patients were followed up for disease recurrence until the first recurrence and thereafter for survival. Results RMST • Median duration of follow-up was 60 months in the dabrafenib + trametinib arm and 58 months in the placebo arm (data cutoff, November 8, 2019) • RMST across the stage III patient population was improved in the dabrafenib + trametinib arm (41.5 months [95% CI, 39.4-43.6 months]) vs the placebo arm (28.7 months [95% CI, 26.3-31.2 months]) (Figure 2; Table 1) – These results suggest that on average, over a 60-month period, patients treated with dabrafenib + trametinib gain an additional 12.8 months of remaining relapse free vs placebo Figure 2. Overall RMST at 60 Months 0 3 6 9 12 15 18 21 24 27 30 33 36 39 Months Since Randomization S ur vi va l P ro ba bi lit y 12.8-month RMST difference 42 45 48 51 54 57 60 63 66 69 72 75 78 81 438 411 391 376 354 329 298 278 262 254 242 236 229 226 217 211 204 201 195 164 133 100 80 41 17 7 2 0 432 Dab + tram No. at risk Pbo Dab + tram Pbo 335 280 250 219 200 185 176 168 166 158 155 147 144 140 139 136 133 132 116 99 74 56 29 13 1 0 0 0.0 0.2 0.1 0.3 0.5 0.4 0.6 0.7 0.9 1.0 0.8 dab, dabrafenib; pbo, placebo; RMST, restricted mean survival time; tram, trametinib. RMST: Subgroup Analysis • RMST was improved with dabrafenib + trametinib across all AJCC 7 stage III substages, with the greatest difference in RMST between arms reported in patients with stage IIIB and IIIC disease (Table 1; Figure 3A-C) Table 1. RMST at 60 Months Overall Stage IIIAa Stage IIIBa Stage IIICa Dab + Tram (n = 438) Pbo (n = 432) Dab + Tram (n = 83) Pbo (n = 71) Dab + Tram (n = 169) Pbo (n = 187) Dab + Tram (n = 181) Pbo (n = 166) Events, n 187 259 25 26 69 117 90 112 RMST (95% CI), mo 41.5 (39.4-43.6) 28.7 (26.3-31.2) 50.4 (46.7-54.2) 42.2 (36.4-47.9) 41.2 (37.7-44.7) 29.0 (25.4-32.6) 38.0 (34.6-41.3) 22.8 (18.9-26.8) RMST difference (95% CI), mo 12.8 (9.5-16.0) 8.2 (1.4-15.1) 12.2 (7.2-17.2) 15.1 (10.0-20.3) dab, dabrafenib; pbo, placebo; RMST, restricted mean survival time; tram, trametinib. a Per American Joint Committee on Cancer’s AJCC Cancer Staging Manual, 7th edition. Figure 3. RMST at 60 Months in the (A) Stage IIIA, (B) Stage IIIB, and (C) Stage IIIC Subgroups (AJCC 7) 83 81 80 79 76 74 70 66 62 62 61 60 58 55 53 52 51 51 50 42 36 28 25 11 2 0 71 59 56 53 50 45 44 42 41 41 39 38 35 35 34 34 33 31 31 25 22 14 13 5 2 0 0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60 63 66 69 72 75 Months Since Randomization 8.2-month RMST difference Dab + tram (n = 83) Pbo (n = 71) 169 161 151 144 134 124 110 103 99 95 91 89 86 86 84 81 80 78 75 61 52 37 26 16 9 5 2 0 187 161 135 120 102 94 84 79 74 73 68 67 65 63 60 60 58 57 56 51 42 33 25 16 6 1 0 0 0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60 63 66 69 72 75 78 81 Months Since Randomization 12.2-month RMST difference Dab + tram (n = 169) Pbo (n = 187) 181 165 156 149 141 128 115 107 99 95 88 85 83 83 78 76 71 70 68 60 44 34 28 13 5 2 0 166 109 84 72 64 58 54 52 50 49 48 47 44 43 43 42 42 42 42 37 32 25 17 7 4 0 0 0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60 63 66 69 72 75 78 Months Since Randomization 15.1-month RMST difference Dab + tram (n = 181) Pbo (n = 166) Dab + tram (n = 83) Pbo (n = 71) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 A. Stage IIIA S ur vi va l P ro ba bi lit y No. at risk Dab + tram (n = 169) Pbo (n = 187) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 S ur vi va l P ro ba bi lit y No. at risk B. Stage IIIB Dab + tram (n = 181) Pbo (n = 166) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 C. Stage IIIC S ur vi va l P ro ba bi lit y No. at risk AJCC, American Joint Committee on Cancer; AJCC 7, AJCC Cancer Staging Manual, 7th edition; dab, dabrafenib; pbo, placebo; RMST, restricted mean survival time; tram, trametinib. Cure-Rate Analysis • The estimated cure rate in the overall stage III population was 51% (95% CI, 46%-56%) in the dabrafenib + trametinib arm compared with 35% (95% CI, 30%-40%) in the placebo arm (Figure 4) Figure 4. Cure-Rate Model Analysis Months Since Randomization S ur vi va l P ro ba bi lit y 0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 0.0 0.2 0.4 0.6 0.8 1.0 Dab + tram KM Estimated Cure Rate (95% CI) 51 (46-56)Dab + tram (n = 438) 35 (30-40)Pbo (n = 432) Pbo KM Dab + tram cure-rate model Pbo cure-rate model 16% absolute difference dab, dabrafenib; KM, Kaplan-Meier; pbo, placebo; tram, trametinib. Cure-Rate Analysis: Subgroup Analysis • The estimated cure rate was improved with dabrafenib + trametinib across all AJCC 7 stage III substages, with the greatest difference between arms reported in patients with stage IIIB and IIIC disease (Figure 5A-C) Figure 5. Cure-Rate Model Analysis for the (A) Stage IIIA, (B) Stage IIIB, and (C) Stage IIIC Subgroups (AJCC 7) S ur vi va l P ro ba bi lit y 0.0 0.2 0.4 0.6 0.8 1.0 Months Since Randomization 0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 84 84 Estimated Cure Rate (95% CI) Estimated Cure Rate (95% CI) A. Stage IIIA 63 (49-77)Dab + tram (n = 83) 52 (33-71)Pbo (n = 71) 11% absolute difference Dab + tram KM Pbo KM Dab + tram cure-rate model Pbo cure-rate model S ur vi va l P ro ba bi lit y 0.0 0.2 0.4 0.6 0.8 1.0 Months Since Randomization 0 6 12 18 24 30 36 42 48 54 60 66 72 78 Dab + tram KM Pbo KM Dab + tram cure-rate model Pbo cure-rate model B. Stage IIIB 54 (46-62)Dab + tram (n = 169) 33 (26-40)Pbo (n = 187) 21% absolute difference S ur vi va l P ro ba bi lit y 0.0 0.2 0.4 0.6 0.8 1.0 Months Since Randomization 0 6 12 18 24 30 36 42 48 54 60 66 72 78 Estimated Cure Rate (95% CI) C. Stage IIIC 43 (35-51)Dab + tram (n = 181) 28 (21-36)Pbo (n = 166) 15% absolute difference Dab + tram KM Pbo KM Dab + tram cure-rate model Pbo cure-rate model AJCC, American Joint Committee on Cancer; AJCC 7, AJCC Cancer Staging Manual, 7th edition; dab, dabrafenib; KM, Kaplan-Meier; pbo, placebo; tram, trametinib. Conclusions • RMST and cure-rate model analyses complement and enhance conventional statistical approaches, including Kaplan-Meier and Cox regression analyses, and may facilitate clinical interpretation of treatment effects for oncologists and patients • Results from RMST and cure-rate modeling analyses suggest that treatment with dabrafenib + trametinib leads to durable RFS benefit compared with placebo – RMST analysis suggests that over a 60-month period, patients treated with dabrafenib + trametinib gain 12.8 months of RFS on average compared with placebo – There was an absolute increase of 16% in the proportion of patients who were cured in the dabrafenib + trametinib arm vs the placebo arm • These analyses provide insights into long-term clinical benefits of adjuvant therapy with dabrafenib + trametinib; overall survival analysis is currently ongoing References 1. Hauschild A, et al. ASCO 2020 [abstract 10001]. 2. Weber JS, et al. ESMO 2019 [abstract 1310O]. 3. Eggermont AM, et al. ASCO 2020 [abstract 10000]. 4. A'Hern RP. J Clin Oncol. 2016;34(28):3474-3476. 5. Kim DH, et al. JAMA Cardiol. 2017;2(11):1179-1180. 6. Pak K, et al. JAMA Oncol. 2017;3(12):1692-1696. 7. Uno H, et al. J Clin Oncol. 2014;32(22):2380-2385. 8. Othus M, et al. Clin Cancer Res. 2012;18(14):3731-3736. 9. Huang L, et al. Cancer. 2008;112(10):2289-2300. Acknowledgements We thank the patients and their families for participating in this study. We also thank all investigators and site staff for their contributions. We thank Lali Sandalic (Novartis Pharmaceuticals Corporation) for providing statistical analysis support. We also thank Maurizio Voi, MD (Novartis Pharmaceuticals Corporation), for guidance and critical review. We thank Zareen Khan, PhD, from ArticulateScience LLC, for medical editorial assistance with this presentation, which was funded by Novartis Pharmaceuticals Corporation (East Hanover, NJ), in accordance with Good Publication Practice (GPP3) (https://www.ismpp.org/gpp3) guidelines and International Committee of Medical Journal Editors recommendations. This study was sponsored by GlaxoSmithKline; dabrafenib and trametinib are assets of Novartis as of March 2, 2015. COMBI-AD is registered at ClinicalTrials.gov (NCT01682083) and conducted in accordance with Study Protocol BRF115532. First Author Disclosures JM Kirkwood reports grants and personal fees from Amgen, Bristol Myers Squibb, Checkmate Pharmaceuticals, and Novartis, and grants from Castle Biosciences, Immunocore LLC, and Iovance, outside of the presented work. Presented online as part of the Winter Clinical Dermatology Virtual Conference 2021; January 16-24, 2021. Previously presented online as part of the ESMO Virtual Congress 2020; September 19-21, 2020. 3 ways to instantly download an electronic copy of this poster to your mobile device or email a copy to your computer or tablet Text: Q143b7 To: 8NOVA (86682) US Only +18324604729 North, Central and South Americas; Caribbean; China +447860024038 UK, Europe & Russia +46737494608 Sweden, Europe Visit the web at: https://novartis.medicalcongressposters.com/Default.aspx?doc=143b7 Scan this QR code Copies of this e-poster obtained through QR, AR and/or text key codes are for personal use only and may not be reproduced without written permission of the authors. Standard data or message rates may apply. http://novartis.medicalcongressposters.com/Default.aspx?doc=ae593