Unveiling the Dual Nature of AI in Oncology

Patrick I. Borgen, MD, chair of the Department of Surgery at Maimonides Medical Center and head of the Maimonides Breast Center, recently shared insights into the evolving role of artificial intelligence (AI) in oncology. In a report by OncLive, Dr. Borgen discusses the groundbreaking potential and inherent challenges of AI in cancer treatment.

At the 41st Annual Miami Breast Cancer Conference, Anant Madabhushi, PhD, from the Winship Cancer Institute of Emory University, highlighted significant advancements in AI for cancer image analysis. The research presented suggests that AI could rival or even outperform traditional genomic profiling in predicting critical cancer characteristics. This positions AI as a potential disruptor in the tumor analysis industry, which heavily relies on genomic, proteomic, and expression assays. AI’s capacity to analyze microscopic images at unprecedented speeds and reduced costs could revolutionize oncology practices.

Balancing Promise with Caution

Despite these promising developments, Dr. Borgen emphasizes the importance of recognizing AI’s current limitations. At the upcoming 42nd Annual Miami Breast Cancer Conference, Joshua Feinberg, MD, from Maimonides Medical Center, will present findings on the performance of AI platforms such as ChatGPT and Google Gemini in breast cancer board examinations. These platforms achieved around 70% accuracy, raising concerns about their reliability in clinical decision-making. A 30% error rate underscores the risks of depending on AI-driven tools for medical guidance, particularly in tasks like interpreting X-rays and pathology slides.

The Path Forward

While AI holds immense potential for enhancing cancer diagnostics and treatment planning, understanding its strengths and weaknesses remains crucial. As the medical community continues to integrate AI into oncology, these insights stress the need for a balanced approach, ensuring technological advancements align with clinical efficacy.