Fig. 1: PFSratio workflow.
a, b Proposed workflow (a) and worked example (b) to properly design, conduct and analyze a clinical trial with PFSratio as its primary endpoint. A single-arm, phase II platform trial testing novel RAS inhibitors in different KRAS-mutant advanced tumors progressing to 1st line therapy is hypothesized, and results are generated through a simulation with randomly generated data. The first step is trial design and sample size calculation. Then, the optimal setting is a to start enrolling patients at 1st line treatment initiation, following the same tumor assessment schedule during PFS1 and PFS2. Once the study has adequate follow up, exploration of PFS1 and PFS2 is recommended, followed by PFSratio analysis selecting the proper method to avoid biased results. Abbreviations: BTC biliary tract cancer, CI confidence interval, CRC colorectal cancer, GBVE Gumbel’s type B bivariate extreme-value, kappa shape of the hazard function, KM Kaplan Meier, PFS progression free survival, PFSratio progression free survival ratio, PD progressive disease, PDAC pancreatic ductal adenocarcinoma, rho pearson correlation index between PFS1 and PFS2, R median ratio between PFS1 and 2, detailed with corresponding δ1 and δ0.
