IMUC was taking big risks when they picked OS as PE for a randomized ph2 trial. Had they picked PFS, we'd be looking at a positive trial today.
The ITT population = 124 patients. In a 2:1 randomization, 1/3 went into the placebo-arm, 2/3 got active treatment. To reach the median survival point in each arm 1/2 of the randomized patients have to have died: If 42 patients were assigned to the placebo arm, the median was reached when 21 deaths were recorded, if 82 patients were assigned to receive active treatment, the median survival point was reached at 41 deaths. All in all 62 events on record. The biometric plan was designed such that the study would show a statistically significant mOS difference with this n=124 for a greater difference than the observed 2 to 3 months. For the observed difference to be statistically significant one would have to have made this a much larger (many more patients) and more time consuming and costly trial. That was the risk, IMUC were taking when they launched this protocol. Based on the early clinical experience they could hope for a much larger difference in mOS. The bet was, if (1) the trial comes out positive for the primary endpoint of OS, we have a product to take to the market right away, if (2) we fail on the PE but have proof of this principle being active and effective, we hopefully have enough additional insight to design our phase 3 program for success.
Even with longer follow-up, the mOS endpoint will not become positive. BUT: We will have secondary survival endpoints which in cancer immunotherapies are particularly interesting and important, e.g. 6 months, 1yr, 2yr, 5yr survival rates comparing the two groups. It is an intrinsic problem in running studies with immunotherapies in cancer, that the benefits may only show in patients who survive long enough to enjoy them.
So, on to a much more educated phase 3 trial protocol now with a very active proven principle called ICT-107! But for this, the company needs more $$$..