AI Tool Shows Prostate Cancer Genetic Testing Feasibility in Community Care

Keeping pace with rapid updates to NCCN guidelines presents a significant burden for community providers processing vast amounts of new clinical data. To address this challenge, David M. Waterhouse, MD, MPH, and colleagues at Oncology Hematology Care (OHC) evaluated artificial intelligence (AI) as a tool to streamline risk stratification and identify genetic testing eligibility for patients with advanced prostate cancer.

Contextualized by findings presented at the 2026 American Society of Clinical Oncology (ASCO) Annual Meeting, he suggested that implementing an AI model to interpret guideline text yielded exceptionally high concordance with manual chart reviews. While the initiative successfully drove a culture of quality improvement at OHC and significantly raised germline and somatic testing rates from 2023 to 2025, Waterhouse noted that the manual effort required to sustain physician engagement cannot easily scale. Ultimately, he emphasized that integrating trusted, responsible technology is vital to taking work off busy clinicians’ plates while ensuring patients receive necessary genomic testing.

Waterhouse is chair of the ASCO State of Oncology Care in America (SOCCA) Committee, director of OHC’s Early Phase Clinical Trials, and principal investigator of phase 1 trials at OHC.

Transcript:

Right now, we saw the NCCN guidelines change between [2023 and 2025]. It would be hard for a community provider to be able to know every change in guidelines and every new drug; there is just too much information to process. [Therefore], building tools that will assist them with their judgment is going to be important going forward.

Our technologies are changing rapidly, but they have got to be built in a way that can be trusted and held accountable. This is just the first step in doing that kind of thing, which everyone knows is coming. We all know it is coming; we are not quite sure how it is coming.

We found resistance when we added more work to a doctor’s plate––a lot of resistance. Even though it was important to the patient, there was friction, not because they do not want to do a good job, but because [they have] too much on their plates already. So, as we move forward, the answer is going to be taking work off their plate, with their oversight, and helping them do this work. That is going to be something that comes out of this very important [work].

Overall, this is not our first quality improvement project. It is not even our first biomarker project or our first genetics project. We are building a culture of improvement at OHC. It is everybody’s job. We want anybody in the organization to be able to come to our quality team and bring their ideas and their suggestions for change. We will teach [them] how to measure it and teach [them] how to see if [they] can make that difference––and it is okay to fail. Not everything we do works. This project, to some extent, was a failure because I could not scale it, but it was successful in working on that culture.

You learn from failures just like you learn from successes. We did improve the testing rate. I fear it is going to fall back off a bit as we stop pestering the doctors. These learnings will help us do something better down the road, and it is an exciting time. Technology is here, whether we want to accept it or not, and doctors are going to have to choose whether they are going to be part of that integration or simply claim that they are victims of it. That is a choice. I hope they choose like I did, embrace it, and start asking, “How can we use it responsibly?” These are all good things that are going to come out of this.

Reference

Singstock MD, Mendenhall M, Arnal S, Waterhouse DM. Real-world genetic testing patterns in prostate cancer assessed via artificial intelligence: implications for NCCN guideline implementation. J Clin Oncol. 2026;44(suppl 16):5022. doi:10.1200/JCO.2026.44.16_suppl.5022