[sachingulaya]

For measuring survival my old lab championed the Kaplan-meier estimator for its ability to censor non-cancer causes of death. The 5yr survival can be heavily skewed if patients are dying from secondary causes such as heart attacks, getting hit by a bus, etc. Oftentimes patients with cancer have other serious comorbidities.

[danieltillett]

Exactly. Most people getting cancer are pretty old. Time to death might make sense in paediatric cancers, but not in 80 year olds.