Hi! I’m flipper44, and I have cognitive-dissonance.
Or.
It’s the final endpoint. (Add trumpets here if you so desire)
What if I told you both Ex and FEMike are correct? What if there is another piece that must be added to the jigsaw puzzle? My guess is everyone here is going through the same reevaluation of their prior scientific investment paradigm.
Upon closer examination, it appears the that the latest data give us compelling evidence that the placebo crossover arm did much better than expected. Ex states this means the trial is irretrievably ruined, because you can’t differentiate patients improved by DCVax-l as different and distinct from SOC patients that just had the good fortune to live longer notwithstanding crossover.
Most trial advocates then argue, that is why the experts added the first two new external control endpoints. Ex then points out, you are still left with a placebo and placebo crossover arm that combined did better or nearly as well as the treatment arm, and 33% of the placebos never even received DCVax-l.
FEMike and Ex both have valid observations. In addition to the ECA arm, how can we distinguish therapeutic response to DCVax-l from good luck?
What if it is primarily the last endpoint that could further illuminate what the data is telling us.
Some trial endpoint background might be helpful. This trial started enrolling in 2007. By 2015, it became apparent to scientists that the trial’s primary endpoint would likely be confounded by psPD. They were also concerned, because the crossover might confound an unblinded overall survival comparison with placebo because 67% of placebo eventually crossed over to DCVax-l.
Skipping ahead, from 2018 through 2020, they worked on the first (yes first) SAP which would include modifying the primary endpoint to overall survival measured by treatment arm compared against an external control arm, but they also kept the overall survival treatment arm versus the placebo arm as a distant secondary endpoint.
Most of you knew all that. OK, but what some of you may have missed, or forgot, is that they also removed some other endpoints, but added, as a final endpoint, tumor response. (Note: they also added two endpoints dealing with confirmed PFS)
Why would they add this in 2019/2020? I mean, we know they previously had different primary, secondary and tertiary endpoints. About seven in all.
Why switch to tumor response? Many probably blew it off at the time as a fluke. Perhaps they hoped something might come of it. Here’s the funny thing, they had been observing (in a blinded manner) tumor responses for 13 years! This was not an oversight, imo. Something was happening to the tumors.
More background. Scientists know, that with therapies like CIs and yes, even dendritic cells, if you wait a number of years, some tumors will go from stable disease, to partial response to complete response. It doesn’t happen frequently, but it does happen. It can literally take five years from completion of therapy to get there in some extreme cases.
Ok, with that in mind, go back to the trial. When original treatment arm patients had tumor recurrence, they were often taken off therapy. Prior to that recurrence, at best you could only judge tumor response by watching the patients who still had residual tumor after surgery. Not ideal, but I would suggest most of these were mesenchymal, because mesenchymal can’t be neatly debulked.
Anyway, the mother-load of tumor observation would likely have to come from the placebo arm crossovers, because most recurrences are inoperable.
Moreover, recurrent tumors are often mesenchymal as well, due to a mutation transition process I won’t go into right now.
At this sage, these tumors would never thereafter partially or completely respond from SOC therapy alone. We know this (ok maybe one in a million) from countless case observations around the world.
So what if they were seeing some tumor responses in some of the long term survivors? What if the tumor responses were clearly nothing seen in the past with SOC. Maybe some of these responses would need years to observe. Maybe some would be quick, but perhaps some would require years of observation to go from stable, to partial to, in a few cases, even complete responses?
Upon unblinding, what if they learned they were only patients treated with DCVax-l — Aka original treatment arm and/or cross-over arm?
This would have profound implications. It would mean that you could also count/attribute the crossover longer term survival to DCVax-l therapy. It would confirm the external control arm findings.
Now, beyond the endpoint, what if the large majority of these responding tumors were otherwise the most aggressive kind? Aka: mesenchymal.
This would further increase confidence.
