CanSurv: Survival Modeling and Cure Analysis

CanSurv Software

CanSurv is a statistical software tool for analyzing population-based cancer survival data. As part of the JPsurv Suite, it provides flexible methods for modeling survival using grouped and individual-level data.

Key Features

  • Parametric survival models
  • Cox proportional hazards models
  • Mixture cure models
  • Inclusion of prognostic and demographic covariates
  • Graphical diagnostics and model assessment

Supported Data

Survival data generated from SEER*Stat survival sessions. Supports grouped (life table) and individual-level data, and both relative and cause-specific survival.

Methods Overview

Parameters and standard errors are estimated using the Newton-Raphson method. Supported models include parametric, Cox, and mixture cure models, including alternative latency distributions.

Outputs and Diagnostics

Actuarial and model-based survival curves, k-year survival estimates, and deviance residuals for model evaluation.

Important Considerations

Mixture cure models estimate both the cure fraction and survival among uncured patients. Estimates may be sensitive to the choice of latency distribution and length of follow-up. Adequate follow-up or large sample sizes are recommended.

Getting Help

Part of the JPsurv Suite

CanSurv provides survival modeling and diagnostics. For trend analysis, see JPsurv.