Insight

Olivier Elemento: The (True) Current State and Roadmap for AI in Cancer Treatment Decisions

Olivier Elemento/LinkedIn

Apr 17, 2025, 03:13

Olivier Elemento: The (True) Current State and Roadmap for AI in Cancer Treatment Decisions

Olivier Elemento, Director of Englander Institute for Precision Medicine, shared a post on LinkedIn:

“The (True) Current State and Roadmap for AI in Cancer Treatment Decisions.

A few months ago, Isaac Kohane asked me to write an opinionated review about “The Use of Artificial Intelligence for Cancer Therapeutic Decision-Making” for NEJM AI. I was fortunate to have colleagues Sean Khozin and Cora Sternberg join me in this effort, and today I’m pleased to share that it’s been published.

Let’s be honest: despite the AI hype cycle, we’re still far from having AI truly transform cancer treatment decisions. While diagnostic applications in radiology and pathology are showing promising maturity, the broader implementation in treatment selection just isn’t there yet. Current AI models, especially LLMs, struggle with consistent adherence to clinical guidelines and most critically, lack validation through properly designed clinical trials.

We remain convinced, however, that AI has the potential to dramatically improve cancer care outcomes if we take the right approach. We need robust multicenter datasets that reflect real-world complexity, practical validation frameworks, and most importantly, a fundamental shift toward rigorously testing these systems like we would any other medical intervention.

The gap between AI’s promise and current reality in oncology is substantial, but with the right roadmap, we can bridge it. Whether incremental or transformative, the path forward must be built on evidence, not hype.”

Authors: Olivier Elemento, Sean Khozin, Cora N. Sternberg

Olivier Elemento: The (True) Current State and Roadmap for AI in Cancer Treatment Decisions