Long-Term Extrapolation of Overall Survival (OS) and Progression-Free Survival (PFS) for the Lunar Trial in Metastatic Non-Small Cell Lung Cancer (NSCLC) Following Progression on Platinum-Based Therapy
Author(s)
Wang B1, Wu E2, Nino de Rivera Guzman J3
1Elysia Group Ltd., Taipei, Taiwan, 2Real Chemistry, New York, NY, USA, 3Novocure, Glendale, CO, USA
Presentation Documents
OBJECTIVES:
To estimate OS (Overall survival) and PFS (Progression-free survival) over a lifetime horizon (20 years) for patients with metastatic NSCLC following progression on or after platinum-based therapy in the LUNAR trial, comparing Tumor Treating Fields (TTFields) with an immune checkpoint inhibitor (ICI) or docetaxel versus ICI or docetaxel alone.METHODS:
The trial data featured median follow-up periods of 10.6 months for the TTFields + ICI or docetaxel group and 9.5 months for the ICI or docetaxel alone group, necessitating extrapolation. Due to uncertainty in the long-term patient distribution, separate OS and PFS curves were generated for the subgroups (TTFields + docetaxel, docetaxel alone, TTFields + ICI, and ICI alone). Kaplan-Meier data were utilized for OS up to 36 months and for PFS up to six weeks, after which fitted parametric curves were applied for further analysis. Parametric models including exponential, Weibull, log-normal, log-logistic, generalized gamma, and gamma were selected based on statistical best fit using Akaike Information Criterion values, visual inspection, and clinical expert opinions. Lastly, mortality was assessed beyond the probabilities indicated by the OS curve.RESULTS:
OS trial data was capped at 36 months due to the significant drop in patient numbers, and PFS at 6 weeks, aligning with initial radiographic assessment and previous methods in NSCLC studies. For OS data, log-normal distributions were used for TTFields + docetaxel, gamma distributions for docetaxel, generalized gamma for TTFields + ICI, and log-normal for ICI. PFS extrapolations employed gamma distributions for TTFields + docetaxel and ICI, and exponential distributions for TTFields + ICI.CONCLUSIONS:
The extrapolation methodology allowed for robust long-term estimation of OS and PFS for patients in the LUNAR trial, providing essential insights for health economic evaluations. The chosen parametric models balanced statistical accuracy, clinical plausibility, and expert recommendations, ensuring reliable projections for the patient subgroups studied.Conference/Value in Health Info
2024-11, ISPOR Europe 2024, Barcelona, Spain
Value in Health, Volume 27, Issue 12, S2 (December 2024)
Code
EE542
Topic
Economic Evaluation
Topic Subcategory
Cost-comparison, Effectiveness, Utility, Benefit Analysis, Trial-Based Economic Evaluation
Disease
Oncology