COIOSDeep risk insights through multi-disciplinary research across the globe
cable network
Item refines Preprint · not peer-reviewed Added

AI reading breast pathology slides identified which patients gained from chemotherapy

Applied to 6,735 TAILORx trial patients without retraining, an AI model reading pathology slides identified intermediate-risk patients who gained 5 percentage points in five-year disease-free interval from chemotherapy.

A prespecified analysis of the TAILORx trial applied an AI model reading pathology slides and clinical details to 6,735 breast cancer patients. Among those given endocrine therapy alone, the model predicted time without recurrence, with a C-index of 0.736. Within the intermediate 21-gene Recurrence Score group, patients it marked as high benefit gained 5 percentage points in five-year disease-free interval from chemotherapy.

Why it is interesting: Splitting the same patients by the 21-gene score alone found no such group; the model never saw TAILORx.

Source
medRxiv, 20 September 2026
Paper
An AI Model Identifies Chemotherapy Benefit in Node-Negative HR+/HER2-Breast Cancer Patients from TAILORx, a Phase 3 Randomized Clinical Trial
Authors
Chan N, Tang C, Biswas D, Park J, Zeng KG, Cappadona J et al.
DOI
10.64898/2026.09.12.26362500
Type
Preprint
Design
Prespecified secondary analysis of the TAILORx phase 3 randomised trial, 6,735 patients with H&E slides and 21-gene Recurrence Score, using an AI model not trained on trial data
Verdict
Refines prior evidence
Driver
AI in medicine
Driver
Cancer
Attention
1 citation (Crossref) · as of 2026-10-28

Other drivers in Data, models and systems