I think just to give some background, I think neurology is in a very exciting time. What we are trying to do here is build AI tools that can hopefully be assistive in neurology practices. For that to happen, we are working on basically creating tools that are robust, that are generalizable, that can actually work in the real world. One of the challenges in terms of finding the right patients for clinical trials is that the clinical diagnostic criteria does not necessarily meet the clinical trial enrollment criteria...
I think just to give some background, I think neurology is in a very exciting time. What we are trying to do here is build AI tools that can hopefully be assistive in neurology practices. For that to happen, we are working on basically creating tools that are robust, that are generalizable, that can actually work in the real world. One of the challenges in terms of finding the right patients for clinical trials is that the clinical diagnostic criteria does not necessarily meet the clinical trial enrollment criteria. So what happens is that when clinical trials are done in different sites, the physicians are trying to recruit these patients to get to those clinical trials, and unfortunately, the diagnosis that is happening in clinical trial sites does not allow the right selection of these patients for these trials, because of which the screen failure rate of these clinical trials is pretty high. So what we are trying to do is to build AI that uses routinely collected clinical workup data, neurology workup data, to see if we can identify those patients that might meet the clinical trial enrollment criteria. So this work that we have done is basically trying to combine different modalities of data that you can obtain in a routine neurology workup to see if we can build advanced AI methods to process all this information to see if we can identify those patients that might be eligible for these clinical trials. So this work is mainly focused on identifying those patients that are both PET positive on amyloid as well as tau to see if we can sort of take that routinely collected data and then make those predictions on how and who might turn out to be screen positive.
This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.
