PI-013 - COMPARISON OF PREDICTED OVERALL SURVIVAL TO OBSERVED CLINICAL DATA IN PATIENTS WITH BRAF V600-MUTANT METASTATIC MELANOMA TREATED WITH ENCORAFENIB ALONE OR ENCORAFENIB + BINIMETINIB.

E. Hahn¹, J. Rashid²; ¹Pfizer, Denver, CO, USA, ²Vertex Pharmaceuticals, La Jolla, CA, USA.

Background: The combination therapy of encorafenib + binimetinib is approved for patients (pts) with unresectable or metastatic melanoma with a BRAF V600E/K mutation in the US, EU, and other regions. Predicting overall survival (OS) based on modeling longitudinal tumor size data could aid in the study design of novel treatments for cancer pts. In this analysis, a sequential approach was used to model the sum of the longest lesion diameters (SLD) of target lesions and baseline factors to predict OS in pts with metastatic melanoma treated with encorafenib alone or encorafenib + binimetinib.

Methods: SLD and OS data were available from 666 pts with BRAF V600 metastatic melanoma in the COLUMBUS trial (NCT01909453). Six tumor dynamic models were evaluated to select the final tumor model. OS was assessed using Cox proportional hazards and parametric time to event models. Five parametric distributions were tested for the best fit to OS data. Stepwise covariate modeling was performed to identify influential covariates.

Results: The Stein model (Stein WD, et al. Oncologist. 2008;13:1046-54) was selected as the final tumor model. The typical estimates for the growth rate constant (Kg) and regression rate constant were 0.144 year-1 and 2.71 year-1, respectively. The parametric survival model that best described the OS data was one with a lognormal distribution. Significant associations between OS and Kg, baseline albumin, presence of liver metastasis, and log-transformed baseline lactose dehydrogenase were identified.

Conclusion: Individual predictions from the final tumor dynamic model well described the observed tumor data for pts in this large study. Kg was the most significant predictor of OS. This modeling framework could be used to predict clinical response (e.g. OS) based on early changes in tumor size for future trials.

Stein WD, Figg WD, Dahut W, Stein AD, Hoshen MB, Price D, Bates SE, Fojo T. Tumor growth rates derived from data for patients in a clinical trial correlate strongly with patient survival: a novel strategy for evaluation of clinical trial data. Oncologist. 2008 Oct;13(10):1046-54. doi: 10.1634/theoncologist.2008-0075. Epub 2008 Oct 6. PMID: 18838440; PMCID: PMC3313464.