Artificial Intelligence May Help Oncologists Better Risk-Stratify Patients With Prostate Cancer

A multi-modal artificial intelligence model may have the ability to accurately stratify patients with prostate cancer into risk subgroups, according to an expert.

During the 2022 American Society for Radiation Oncology (ASTRO) Annual Meeting, CancerNetwork^® spoke with Louis Potters, MD, FACR, FABS, FASTRO, about the potential benefits of artificial intelligence.

Potters, who is the chair of the Department of Radiation Medicine and deputy physician-in-chief of Northwell Health Cancer Institute in New Hyde Park, New York, said that further development of this prognostic tool may be helpful in more accurately identifying patients who are suitable for enrollment in future clinical trials.

Transcript:

The theme of the ASTRO meeting was really to explore the idea of artificial intelligence and neural networks in the utilization of data. During the plenary session, there was a really great study that was presented in prostate cancer looking at risk-stratification based on about 5600 patients enrolled in the NRG group cooperative studies, creating a sort of a multimodal risk-stratification to identify statistical difference and outcomes. That's now associated with a third-party software vendor who's going to be looking at creating this prognostic tool, which will help to stratify patients better than the typical risk-stratification that's been utilized in the past for prostate cancer. That will be helpful in identifying patients that will be enrolled in clinical trials.

What is also interesting is that the number of patients that would otherwise be stratified between intermediate- and high-risk disease, and the number that actually fall into a more favorable category somewhere between intermediate- and high-risk, is actually greater than what you would think we currently see with high-risk patients. We're probably over-stratifying high-risk patients today. It'll be interesting to see the validity of this data [in the] long-term and what the prognostic tool is going to be able to do for us in the clinic and for recruiting patients on clinical trials.

Reference

Tward JD, Zhang J, Esteva A, et al. Prostate cancer risk stratification in NRG Oncology phase III randomized trials using multi-modal deep learning with digital histopathology. Presented at 2022 American Society for Radiation Oncology Annual Meeting (ASTRO); October 23-26, 2022; San Antonio, TX. Abstract 2. Accessed November 7, 2022.