New Updates on the Evolving Landscape of AI in Oncology

In this latest edition of his Fellows Forum blog, Dr. Waqas Haque pens his thoughts on how to approach AI in oncology with equity and fairness, a multicancer early detection test and the idea of “precision equity,” a strategy for making genetic therapies more accessible and affordable, and more!

In the evolving landscape of oncology, technological advancements are continuing to push the boundaries of what is possible in cancer treatment. We at Oncology Data Advisor are excited to play a small role in the dissemination of new findings and breakthroughs at the intersection of cancer care and artificial intelligence. We were recently recognized with the APEX Award for our series Exploring AI in Oncology: The Convergence of Technology and Cancer Care (see press release here).

Despite the potential of AI, technological advancements are only as valuable as the accessibility they afford to patients who need them. As we welcome AI tools that streamline what providers don’t have to do, as outlined in Dr. Rahul Banerjee’s recent update, it is imperative to address broader limitations involving access to care. A Perspective article in Nature Reviews: Clinical Oncology cautions about approaching AI in oncology with equity and fairness. AI algorithms often inherit biases present in training data, which can result in disparities in cancer diagnosis and treatment, particularly for underserved populations. To mitigate these issues, the development and deployment of AI in oncology should be grounded in robust, evidence-based research and should include diverse, representative datasets. Policies designed to improve access to cancer care must prioritize global data consortia, rigorous validation methods, and context-appropriate AI models.

Another JAMA Oncology Viewpoint from last month discusses advancing “precision equity” through liquid biopsies for cancer screening, especially in underserved communities. Traditional screening methods face access challenges among low-income groups including poor uptake and logistical barriers; for example, only 6% of individuals eligible guideline-based lung cancer screening actually undergo low-dose CT lung scans. Liquid biopsies, which detect multiple cancers via a routine blood test, offer a potential scalable solution. The article discusses the Mercy Clinic Pilot MCD Program, which offers the GRAIL Galleri test in conjunction with a care navigator for individuals with relevant social and health risk factors (smoking or alcohol use, cirrhosis, and cancer in a first-degree relative, among others). This approach has begun to show positive results, particularly in communities with historical environmental and occupational exposures, fostering equitable access to cancer care. At the same time, scalable tests intended to benefit patients at large must be rigorously evaluated to ensure that they actually improve outcomes. An NEJM op-ed emphasizes that without robust evidence from randomized controlled trials (RCTs) demonstrating reduced cancer-specific mortality, widespread use may lead to overdiagnosis and unnecessary interventions.

Looking beyond AI, a recently released Nature article, “A Roadmap for Affordable Genetic Medicines,” outlines a comprehensive strategy to make genetic therapies more accessible and affordable. Despite the promise of genetic treatments, high costs and manufacturing limitations restrict their global reach. The authors propose several key solutions: academic institutions should include access provisions in their licensing agreements, and manufacturing innovations should reduce costs. Regulatory frameworks need to adopt a risk-based approach to facilitate faster approvals, and maintaining perpetual investigational new drug (IND) status for ultra-rare disorders could provide said access to therapies. These measures could lower prices, streamline production, and ensure equitable access to cutting-edge treatments.