Replies to post #272946 on NorthWest Biotherapeutics Inc (NWBO)

[Doc logic]

JerryCampbell,

Yes, in a large enough trial you can find stuff. However, ICT-107 was not that kind of trial and neither is DCVax-L Phase 3. Safety not being an issue for either one. ICT-107 was post hoc unblinded. DCVax-L will have been blinded. I believe both had/have merit to be considered at a minimum for a provisional approval and tightly controlled Phase 4. NWBO's position is stronger because of remaining blinded and having more mature data. Patients need options and there is more than sufficient proof that DC treatments given correctly in combo with checkpoint inhibitors improve patient outcomes. Best wishes.

[iwasadiver]

First of all Linda Powers was using a very public forum to state her positions to the FDA, and it’s very clear the FDA has begun to realize that some, if not most, of these old standards are simply outdated and culturally clung to issues with no real bearing on today’s clinical research world. The P value is, if you looked and read the paper she sited, an issue that these 800 statisticians believe needs to be changed and could even be deleterious in our adherence to current standards.

As far as subgroups and “data mining” for not identifying ahead of time such subgroups is also outdated and again clung to by those who seem to do so culturally, as if they quit looking for new ways to deal with new types of clinical research. Particularly as we move forward with increasing identification of various genetic markers and find ourselves researching the research we can fall into a downward spiral of spending years doing absolutely nothing because we keep peeling the damn onion, all the while the reality is staring us in the face. There are now various statisticians in the field of bio pharmaceuticals that are finding ways around the criticism that multiplicity causes unreliable inference.

https://www.tandfonline.com/doi/abs/10.1080/10543406.2017.1397009?src=recsys&journalCode=lbps20

https://www.ncbi.nlm.nih.gov/pubmed/29173045

https://www.bmj.com/content/351/bmj.h5651

Ms. Powers obviously believes that the goods in our trial are there and I believe the SAP used some of the statisticians who’ve spoken out and written on both the P value and the post trial Subgroup Analysis. This is not a coincidental piece of writing to the FDA

[longfellow95]

well, meth/unmeth was pre-defined, though without proper control for multiplicity.
And meth has a very good chance of demonstrating stat sig OS advantage.
If there is to be approval for a subgroup, rather than the whole GBM population, it'll surely be meth, imo.