longfellow, you are making some good points regarding the need for equivalence testing regarding a possible trial's methodology change. Most likely after the trial is over we will get some information as to whether any changes (i.e. manufacturing) were made and whether they were minor enough and within the guidelines of the trial to avoid equivalence testing but were enough to improve results.
With regards to the hold, I agree that the time between treatment of the last group and imposition of the hold was quite short. However, as a group the last successful 106 (or 108, whatever) may have had many patients who demonstrated strong inflammatory responses. is it
possible that between 11/14 and 8/15 an unusually large number of patients demonstrated inflammatory reactions that within a few months showed up as psPD? By that time they would have become very familiar with this DCVax-L related psPD and may have established that within a few months. They may have therefore guessed correctly by 8/15 via previous observations that this last cohort will have an unusually large number of long term survivors. Whether the reasons for this looming success was improved surgery or another reason, it was possibly decided for ethical reasons to treat the last 31 patients only with DCVax-L. This is just a possibility that comes to mind.
Finally, with regard to the question as to whether the last 108 demonstrated unusually good survival capacities between years 2 and 3 or from day 1 through month 36, I am thoroughly confused.
Lykiri recently posted that one of Dr. Bosch's presentation and Dr. Liau's 2018 data has the answer. Accordingly the first 223 patients had at 2 years a 40% survival rate, the last 54 patients in the trial (the last 50% of the 108) had a 65%-70% survival rate and finally the entire trial (331), a 46.4% survival rate. Unfortunately I was not able to find Dr. Bosch's presentation. In fact, Dr. LIau's recent 2018 SNO update only changes the 2 year survival figure from 46.2% at 3/17 to 46.4% at 10/18.
This mere 0.2% increase in the 2 year survival is totally a function of the last 54 patients who by 10/18 had passed the 2 year timeline. One can calculate accordingly that the last 54 had at 2 years about a 48% survival rate and certainly not a 65%-70% survival.
However, it occurred to me that perhaps the 2 year survival figure of 46.2% recorded at 3/17 was a derived figure that was based on performance up to 3/17 and predicted the final 2 year survival percentage of the entire trial. In that case, a projection of a 65%-70% survival at 2 years for the last 54 patients could have been factored in and they ended up being only off by 0.2% (46.2% vs. 46.4%). However, at 3/17 their 3 year survival rate was 24.2% and it ended up being 28.2% by 10/18. If they had a basically correct derived figure at 3/17 for the 2 year survival rate, why would they have so underestimated the final 3 year survival rate. Finally they never mentioned in the publication that any of their 3/17 survival rates for 3 years on trial were derived so I seriously doubt that they were.
In summary I do know that 42 of the last 108 (38.9%) reached 36 months but don't understand anything about the attrition rate of this group up to that time.
AI
