Bringing Trustworthy Artificial Intelligence to the Clinical Forefront at JCO: A Guide for Studies Testing Artificial Intelligence Models

This editorial, published in the Journal of Clinical Oncology, establishes a framework for evaluating and reporting AI and machine learning models in oncology research, outlining the standards that submissions to JCO must meet to demonstrate clinical trustworthiness. Authored by the journal's editorial leadership, the guide emphasizes a critical distinction between AI tools that predict risk and those that genuinely support clinical decision-making, calling for a higher bar that prioritizes actionable guidance over prognostic scores alone. It outlines editorial priorities including rigorous validation, transparency in model development, and demonstration of real-world clinical utility.

The publication reflects a broader and increasingly urgent conversation in oncology about how AI tools should be developed, tested, and reported before they reach the clinic. As AI-enabled devices and algorithms proliferate across cancer care, this guide sets a meaningful standard for the field, pushing researchers and developers to move beyond benchmark performance and demonstrate that their tools can meaningfully improve patient outcomes at the point of care.