Next steps for ArteraAI and prostate cancer study data

In this video, Charles Parker, MD, and Gerhardt Attard, MD, PhD, discuss future directions following the study “External validation of a digital pathology-based multimodal artificial intelligence (MMAI)-derived model in high-risk localized (M0)/metastatic (M1) prostate cancer (PCa) starting androgen deprivation therapy (ADT) in the docetaxel (Doc) or abiraterone (AAP) phase III STAMPEDE trials,” which was presented at the 2023 European Society for Medical Oncology Annual Congress in Madrid, Spain. Parker is a clinical research fellow and PhD student at UCL Medical College in London, the United Kingdom. Attard is the John Black Charitable Foundation Endowed Chair in Urological Cancer Research at University College London.

Transcription:

What are the next steps for this research?

Parker: There are a couple of really exciting directions we want to take this work in. Firstly, the model was an excellent prognostic tool, [but it] doesn't tell us the biology of what's driving the differences in these patients. So one way that we can work on answering this is we can use the other strength of the STAMPEDE trial and the STRATOSPHere biomarker work, which is that as part of the workflow, we did serial sections of the box that we received from each patient. So we have as close as possible the same tissue that had an H&E stained diagnostic slide, sections taken for immunohistochemical studies, sections taken for low-pulse whole-genome sequencing, and sections taken for transcriptomic profiling. So for each of these patients, we have really deep multi-omic data, which we can now start to integrate. Now we know which cohorts of patients that have a different clinical behavior, we can then start to integrate across all of these studies and really dive down into what's driving the difference in disease in these patients. At the other end, we have to bear in mind that all this work has been done on trial patients who are pre selected and a little bit fitter than the general population, so what we really need to do as well is see how this test behaves in the real world. So we're quite excited at looking at the possibility of how we can generate prospective data from this with pilot studies at UCL and the University of Manchester, 2 of the lead centers for the STAMPEDE study, to see how these tests can be used in the clinical setting, and what kind of data we get from that. So there are a couple of exciting directions for us to take with this.

Attard: And to build on that a bit more, the test is prognostic. The next question is, which treatment should we give, and we have this backbone of treatments, but then there's a number of decisions we need to make; for example, should this patient also have chemotherapy? And the test can tell us he is at higher risk of dying from prostate cancer, so he's a patient you want to look for additional treatments for, but will chemotherapy work or will it not? And the experiment hasn't asked that question. So that's the next step. And we may achieve that just using the same inputs we've had for this model. As Charles has explained, we've also been collecting a wide variety of different modalities of data that we may use for those predictive questions. And also to improve prognostication; potentially, there's improvements that could be made. You always want to improve. But I think the predictive question is probably the next really burning one.

This transcription was edited for clarity.