Just a clarification, when I said absent knowledge of 9902a dataset - I meant by knowing whether a different Cox modeling approach would fail to improve the treatment effect seen, or knowing whether the final model (published in JCO) would result in significance which a different modeling approach would not. Hope it's clear.
Market Data 
Markets 
AI
