AI as an intervention: improving clinical outcomes relies on a causal approach to AI development and validation.

The primary practice of healthcare artificial intelligence (AI) starts with model development, often using state-of-the-art AI, retrospectively evaluated using metrics lifted from the AI literature like AUROC and DICE score. However, good performance on these metrics may not translate to improved clinical outcomes. Instead, we argue for a better development pipeline constructed by working backward from the end goal of positively impacting clinically relevant outcomes using AI, leading to considerations of causality in model development and validation, and subsequently a better development pipeline.

Pre-malignant conditions diagnosed following a positive cancer signal from a multi-cancer early detection test.

Blood-based tests for multi-cancer early detection (MCED) are being developed to facilitate the detection of various cancer types. The Detecting cancers Earlier Through Elective mutation-based blood Collection and Testing study (DETECT-A) study evaluated an MCED test in 9,911 women, age 65-75, without personal history of cancer. In a analysis, we report on the detection of precancerous neoplasms consequent to MCED testing and follow-up.