Proof of Concept / Clinical Biostatistics
Kaplan-Meier & Cox Regression for Uveal Melanoma
A synthetic clinical trial simulation modeled after IMCgp100-202 (tebentafusp, Kimmtrak), the landmark Phase III trial in metastatic uveal melanoma. This proof of concept generates patient-level data from Weibull survival distributions with covariate effects, then computes Kaplan-Meier estimators with Greenwood variance, log-rank tests, Cox proportional hazards modeling, and subgroup forest plots — all running entirely in the browser with zero backend dependencies.
Disclaimer: This is a proof-of-concept demonstration using synthetic data generated from Weibull distributions. All patient records, survival times, hazard ratios, and statistical outputs are simulated and do not represent actual clinical trial results from IMCgp100-202 or any other study. Individual patient trajectories are modeled, not derived from real patient data. This tool is NOT intended for clinical decision-making and should NOT be used to inform treatment decisions for uveal melanoma or any other condition.
Generating synthetic trial data & computing survival curves...
Kaplan-Meier product-limit estimator with Greenwood's formula for variance estimation and pointwise 95% confidence intervals. Log-rank test for between-group comparison using the Mantel-Haenszel chi-squared statistic with 1 degree of freedom.
Cox proportional hazards model estimated via Newton-Raphson optimization of the partial likelihood with Breslow tie handling. Standard errors derived from the observed information matrix. Model discrimination assessed via Harrell's concordance index.
Weibull survival distributions (shape=1.1) calibrated to published aggregate results. Covariate effects are multiplicative on the scale parameter: ECOG 1 (0.75x), elevated LDH (0.7x), liver mets (0.9x), age >70 (0.85x). 2:1 randomization with ~20% random censoring and 36-month maximum follow-up.