Modeling Overall Survival Under Immuno-Oncology Therapies with Long-Term Survival Models

• Activity Number: 351 - Statistical Issues Specific the Therapeutic Areas- 3
• Type: Contributed
• Date/Time: Tuesday, July 31, 2018 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Junshui Ma* and Keaven Anderson
• Companies: Merck & Co., Inc. and Merck & Co., Inc.
• Keywords: Immuno-oncology; Long-term Survival; Cure Rate; Non-proportional Hazard

Abstract

Accumulating evidence suggests a portion of cancer patients achieved Long-Term Survival (LTS) under immunotherapies. For example, published results showed that ~19% metastatic melanoma patients survived over 9 years under an anti-CTLA4 treatment. Similar durable benefits of immunotherapies have been observed across other tumor types, including Non-small Cell Lung Cancer (NSCLC). This scenario violates the assumption of many popular parametric survival models (e.g. exponential and Weibull models), in which survival probability goes to zero with time. Thus, an alternative, but well-established, class of parametric survival models, called Long-Term Survival (LTS) models, should be used to better model the survival data of immunotherapies. In this presentation, the clear advantages of using LTS models over popular parametric survival models will be shown in at least two aspects: (1) LTS models fit the survival data of immunotherapies better; (2) LTS models predict future survival results reasonably good. Finally, we will discuss how the conventional thinking of proportional hazard (PH) and Non-PH should be adjusted under the framework of LTS models.