Replies to post #98589 on NorthWest Biotherapeutics Inc (NWBO)

[Doktornolittle]

Fantastic stuff Flipper! Regardless of the relevance to the DCVax-L trial, this information and the spirit of the presentations is tremendously encouraging for the future of immunotherapies and other cancer treatments.

Possibly automated cancer subgroup detection via automated image analysis. Possibly going on right now in post data analysis in the DCVax-L trial... including distinguishing real from pseudo-progression... I assume that is what you are suggesting.

Another related benefit for DCVax-L could be the following: I have been concerned about efficacy in the Mesenchymal subgroup because I have not heard anybody say that the subgroup can be identified in time to select DCVax-L prior to surgery. That would seem to be a logistical problem to me. If so, NWBO might need to get approval for the entire GBM population, which is more difficult. But this presentation suggests that there may be a non-invasive way to identify such subgroups before surgery using "Radiomics"; if not now, then possibly in the near future. This might influence the FDA to allow approval for that subgroup even though the logistics might initially be more burdensome than the FDA would like.

[antihama]

Here's the transcription (99% accurate)

@1:06 What else we can do w these images….Radionomics is a relatively new tech in imagining. Quantitative molecular imaging provides a potential platform for linking specific imaging traits w a specific gene expression patterns and these imaging features may actually serve as molecular surrogates that contribute to diagnosis, prognosis, and likely gene expression associated treatment response. So here’s a pt on TKI… these pts didn’t have the classic response pattern. Certainly, not the debt of response that was accessed but actually the tumors became ghosted and this is a way of looking at the individual pixels in here to define a response phenotype based upon on certain futures in this basically the graying of the tumor…
Potentially w the exact same scans that we have we MAY be able to distinguish between true progression and pseudo based upon on the patterns that we actually see in the individual pixels on the same imaging set.
So in conclusion what we need is better data collection of the actual images in the appropriate meta data and this should include images that were obtained post progression, that’s absolutely critical. We need to standardize iRECIST, as a 1st step w very specific rules of response and progression and then we also have to enhance the data collection so that our case report forms are more robust and complete. For instance, just beginning w the basic concept of report - all the sites of disease and all the different methods that we see progression. And we should explore alternatives - Radionomics is one that I briefly touched upon solely because it’s available, it’s actually available on the actual images. @1:08:35