>> The notion that immunotherapy produces late-separating survival curves while chemo produces early-separating survival curves is more spin than science. <<
That's probably true , though I wouldn't be surprised if it becomes more science-like as evidence accumulates since there is a plausible MOA to explain it.
Recognizing a real , meaningful separation in survival curves is the important thing , whether it occurs early or late , when you're talking about drugs for terminal conditions. Optimization , combos , etc. , can increase the spread and move the separation up to earlier time points , but only if the drugs are available to be optimized. What bothers me is the idea that ingrained review methodologies may stifle progress. One thing I'd suggest as a part of every BLA where survival is an endpoint -- a final group picture of Tx. arm survivors vs. control arm survivors , just to keep it real.
On a related note , I sense that a "rebranding" may be occurring with regard to certain chemo drugs which fits in with your comment above. Every week a new paper comes out describing the effects of current chemotherapy drugs on various immune parameters like T-reg depletion , Mac1 vs. Mac2 predominance , myeloid vs. plasmacytoid , cytokine profiles , etc. In a few years they may be calling it ' Chemmunotherapy '.