“AI can help clinicians translate unprecedented biological and clinical complexity into actionable insights,” says Robert L. Ferris, MD, PhD.

As an oncologist and health system leader, I am increasingly excited by the opportunity for artificial intelligence (AI) to transform precision oncology. Modern cancer care already depends on an extraordinary volume of information, including pathology, imaging, genomics, treatment history, clinical data, and emerging biomarkers. The challenge is no longer simply generating information; it is integrating that information rapidly and meaningfully to make better decisions for individual patients. AI has the potential to become an essential partner in that process, helping clinicians translate unprecedented biological and clinical complexity into actionable insights.

Next-Generation Sequencing

The expansion of next-generation sequencing has made this opportunity particularly compelling. We can now identify mutations, amplifications, fusions, copy-number alterations, and tumor-agnostic biomarkers, but a genomic alteration cannot be interpreted in isolation. Its significance depends on tumor type, co-mutations, prior therapies, and the broader clinical context. The future of precision oncology therefore lies beyond simple mutation-to-drug matching. AI can help integrate these diverse dimensions to estimate treatment response, identify resistance mechanisms, and support increasingly individualized therapeutic strategies.

AI-Enabled Molecular Tumor Board

One of the most immediate applications is the AI-enabled molecular tumor board. Today, clinicians must synthesize sequencing reports, medical records, scientific literature, clinical trials, and prior treatment history—an increasingly time-consuming task. Retrieval-augmented generation and large language models could rapidly organize this evidence, identify relevant therapeutic options, and surface clinical trials. Importantly, these systems should augment rather than replace oncologists. Hallucinations and inaccurate interpretations remain important limitations, making clinical oversight and verification essential. We have tested several products on the market and they’re advancing rapidly.

Clinical Trial Matching

Clinical trial matching represents another compelling opportunity. AI could interpret complex eligibility criteria and compare them with patient characteristics, potentially identifying trials that might otherwise be overlooked. Similarly, multimodal models that integrate genomic, transcriptomic, cellular, and other biological information may move us toward predicting how an individual tumor will respond to specific therapies rather than relying on biomarkers alone. This could ultimately reduce exposure to ineffective treatments and enable more adaptive treatment strategies as tumors evolve.

Single-Cell Sequencing and Patient-Derived Organoids

I am particularly enthusiastic about the convergence of AI with single-cell sequencing and patient-derived organoids, to guide patient therapy decision marking. Tumors are biologically heterogeneous, containing populations of cells with different vulnerabilities and mechanisms of resistance. Combining single-cell data, organoid drug testing, and genomic profiling could provide an increasingly comprehensive picture of an individual patient’s cancer and identify rational treatment combinations. AI may also help address the enormous complexity of combination therapy by prioritizing drug pairs or regimens with the greatest potential for synergy while incorporating toxicity, sequencing, interactions, and patient-specific factors.

What Health Systems Need to Make AI Work in Oncology

For health systems, however, realizing the promise of AI will require more than sophisticated algorithms. Models must be prospectively validated, benchmarked consistently, and demonstrated to improve meaningful clinical outcomes. Data quality, interoperability, privacy, bias, reimbursement, computational infrastructure, and equitable access will all influence implementation. We must ensure that AI benefits patients across academic and community settings rather than widening existing disparities.

Looking ahead, I envision oncology systems in which AI seamlessly integrates genomics, transcriptomics, pathology, imaging, single-cell data, organoid responses, and longitudinal electronic health records. Such systems could help clinicians understand each patient’s cancer at an unprecedented level of resolution and continuously adapt treatment as disease biology changes. AI-assisted tumor boards, trial matching, response prediction, and therapeutic design could become routine components of cancer care.

The possibilities are genuinely exciting. AI will not replace the oncologist; rather, its greatest value may be in extending our ability to understand complex disease and make informed, individualized decisions. If we pair technological innovation with rigorous validation, responsible implementation, and continued human judgment, AI can help us build an oncology system that is more personalized, adaptive, efficient, and ultimately more capable of delivering the right treatment to the right patient at the right time.

Source: AI as a Partner in Oncology | Targeted Oncology – Immunotherapy, Biomarkers, and Cancer Pathways