On World Cancer Research Day, Medscape released data indicating that cancer dominates physician queries to AI‑powered platforms in every market studied. The analysis, which examined millions of AI searches from nine countries, shows clinicians are turning to advanced tools to help manage the increasingly information‑rich field of oncology.
Why cancer leads AI searches
Cancer remains one of the most information‑intensive specialties. With a wide variety of possible causes, symptoms, cell types, and disease progression patterns, doctors face a daunting amount of data when diagnosing and treating each patient. Dr. Maurie Markman, a professor of medical oncology at City of Hope Comprehensive Cancer Center, explained that “physicians recognize that every patient’s cancer may be biologically unique,” and that treatment decisions now require consideration of biomarkers, tumor genetics, emerging research, and evolving guidelines.
Growth of treatment options and data overload
The past two decades have seen a dramatic expansion in approved cancer therapies. According to a JAMA article, the number of cancer medicines receiving regular approval rose from 26 between 2006‑2010 to 115 between 2021‑2025 – a 340% increase. Precision‑medicine advances, such as tissue‑agnostic therapies that target specific genetic markers regardless of tumor origin, have further broadened options for patients.
Next‑generation sequencing (NGS) now allows oncologists to test dozens of genes simultaneously, uncovering mutations that can guide targeted treatment. While this technology offers unprecedented personalization, interpreting the complex molecular data can be challenging, prompting many clinicians to seek AI assistance.
AI models showing promise
Researchers at Harvard Medical School have developed the Clinical Histopathology Imaging Evaluation Foundation (CHIEF) model, which reads digital tumor slides to detect cancer cells and molecular profiles with nearly 94% accuracy across 15 databases covering 11 cancer types. Another AI system combining stimulated Raman histology, optical imaging, and deep learning can diagnose intra‑operative brain tumors in about 150 seconds with 94.6% accuracy, outperforming traditional pathology timelines.
These tools help physicians quickly synthesize large volumes of literature, compare treatment approaches, and navigate evolving guidelines, supporting precision‑medicine efforts.
Risks and limitations
Despite the promise, AI adoption carries risks. A 2024 study of colonoscopy centers found that adenoma detection rates rose when AI assistance was used but fell below baseline after the tool was removed, suggesting possible overreliance that could erode clinicians’ independent diagnostic skills.
Human bias in algorithm development also raises safety concerns. A 2024 review highlighted that many AI models were trained on publicly available image datasets that under‑represent darker skin tones, potentially leading to inaccurate predictions for patients of color.
Looking ahead
Future innovations such as single‑cell sequencing and liquid biopsies aim to make precision oncology more accessible and effective. As the volume of cancer‑related data continues to grow, AI is likely to remain a valuable adjunct for oncologists—provided its limitations are acknowledged and clinicians maintain their own expertise.
Original reporting: KTVZ (Central Oregon) — read the source article.
