A point that's been discussed to death, yes. The company did decide that PFS failed because of pseudoprogression. That itself is a debatable point, maybe more relevant in GBM (wouldn't know; no immunotherapies have been approved for GBM), but when this pseudoprogression idea was in vogue a decade ago, we've adjusted the assessment criteria to better see it.
This is why you rarely see those old "we saw a big OS improvement, but no PFS benefit" these days. It's why you regularly see PFS benefit with checkpoint blockade in well-designed trials with modern endpoints.
Overall survival IS an objective endpoint, but when you lose randomization, it very quickly becomes a suspect endpoint. For example, can you answer why DCVax-L had an apparent improvement at the time of randomization, when the OS curves start to split?
Is is because DCVax-L starts working on day 1, unlike every other immunotherapy inexistence?
Or do you think there might be something different about the populations being studied?
There's a logical conclusion, but it's ignored here. The external control arm, with an inability to resolve whether the populations were balanced, leaves too many important questions unsolved to know whether the small, statistically significant improvement in OS rate is "real." If the effect size was bigger, it would be more convincing. As it stands, you cannot answer the question "were the patients enrolled in the DCVax-L study in better shape overall (in all its forms: comorbidities, age, disease biomarkers, etc) destined to do better than historical controls?"
If they had run a randomized trial, it wouldn't matter. You would say "gee, they did better than historical controls, but the experimental still beat the controls." And it would be a bona fide win, and it would be an approved regimen right now in the United States, not languishing in review with the British government.
AI
Market Data 
Markets 
