Trust and Transparency Key for Leveraging AI to Expand Access to Precision Oncology

As medicine incorporates artificial intelligence (AI), one of the most exciting applications is in oncology, where leaders hope it holds the potential to bring precision medicine to more patients by matching them with clinical trials and targeted therapies. Experts on the panel “Access for Everyone: Using AI in Precision Oncology” during the Patient-Centered Oncology Care 2025 conference highlighted the ways their institutions are using AI tools and their thoughts on future directions.

To kick off the panel, moderator Davey Daniel, MD, chief medical officer of OneOncology, a self-proclaimed “AI novice,” asked the speakers to provide an example of how they’ve heard of AI being leveraged for precision oncology.

• Vivek Subbiah, MD, chief of early-phase drug development, Sarah Cannon Research Institute, said that AI is assisting his team in sifting through clinical trials to find the right match for patients by their biomarkers.
• Stephen Speicher, MD, MS, senior medical director and head of health care quality and patient safety, Flatiron Health, mentioned the opportunity to bring precision medicine to the point of care by optimizing test ordering.
• John L. Villano, MD, PhD, director of clinical neuro-oncology, University of Kentucky, discussed the insights gleaned from molecular testing reports.
• Alyssa Schatz, DrPH, MSW, vice president of policy and advocacy, National Comprehensive Cancer Network, has been monitoring AI’s impact on oncology care and patient access, quality, and safety.

Translating Technology Into Policy

As the capabilities of AI expand, the regulatory environment around these tools is “highly complex, even more so than many of the other policy issues that we deal with today, and it’s also evolving,” Schatz said. States are implementing legislation around how AI can be used in health care decisions, such as payer coverage determinations or prior authorizations. On the federal level, she noted that whereas the Biden administration was more cautious, the Trump administration is more eager to have health care providers experiment with AI. In fact, the federal AI Action Plan released in July calls out the health care sector as “especially slow to adopt” AI.¹

Still, there is an appetite for AI adoption, Speicher said, pointing to an American Medical Association survey that found clinician use of AI doubled from 2023 to 2024.² But the panelists agreed that trust and transparency are key: Health care providers including oncologists need to understand the sources of data and how the tool works in order to feel comfortable using it. Schatz suggested that legislation on this issue is a while away, so the onus is currently on the developers to build trust in their products.

AI Transforming Oncology—for Patients, Providers, and Pharma

The opportunities for leveraging AI in precision oncology are significant and growing, the panelists agreed. According to Subbiah, molecular testing reports contain actionable transcriptomics information that is often hidden beyond the first few pages, so “it gives us options for personalizing care, we can match the patient to a drug trial or a CDK [cyclin-dependent kinase] inhibitor, so that wasn’t apparent 5 or 10 years ago.”

AI tools like ChatGPT are also becoming a resource for patients seeking to better understand their disease and treatment. Villano said that these patients tend to ask more questions, which can take more time during a clinic visit, “it’s beneficial for patients to have that discussion through each of the molecular tests that is done and how impactful it is. It does take more time, but patients do learn from it, especially if they’re really interested.”

Schatz also mentioned the potential for AI to improve the clinical trial enrollment process, which is currently burdensome. The ClinicalTrials.gov site is unwieldy even for clinicians and navigators, so she hopes AI tools could automate the access pathway and make it more available to patients.

The research and development process can also benefit from AI, Subbiah noted, such as by using AI scoring systems to “rescue” drug candidates that have failed in a clinical trial and apply them to a different indication or subpopulation.

Next Steps for AI in Precision Oncology

For cancer centers looking to incorporate AI into their workflow, Speicher suggested asking the technology developers questions around their training models, funding stability, clinical oversight, internal compliance, and safety for clinical use cases. “It’s a mutual relationship, you’re spending a lot of time working with them to build these products, and a good vendor is one that will work with you over time,” he said. It’s also essential that these tools have been repeatedly vetted by clinical experts and not just built by engineers working in a silo.

The panelists agreed that it’s an exciting time for the oncology field, considering the expanding treatment options and technological capabilities. Villano encouraged the audience to keep an open mind as AI is incorporated more in patient care instead of being scared of it.

References

1. Winning the Race: America’s AI Action Plan. The White House; July 2025. Accessed September 26, 2025. https://www.whitehouse.gov/wp-content/uploads/2025/07/Americas-AI-Action-Plan.pdf

2. AMA augmented medical intelligence research. American Medical Association. February 2025. Accessed September 26, 2025. https://www.ama-assn.org/system/files/physician-ai-sentiment-report.pdf