The abstract is read at the publisher; the summary is JClub's.
TEAM, an interactive trustworthy AI pathology copilot, improves biomarker-driven prognostic stratification and therapeutic response prediction, outperforming prior methods across 88 benchmarks.
- Why it matters: Precision oncology needs models that jointly profile morphological biomarker signatures and predict outcomes, with uncertainty and interactive refinement for clinical translation.
- What they did: TEAM is pretrained on 55,648 pan-cancer WSIs and 1,750,648 ROIs (360 million patches), learns risk-aware embeddings using clinical metadata and a relative risk prior, and propagates patch-level aleatoric and model epistemic uncertainty to slide-level predictions.
- The result: Using profiled biomarkers as intermediate features and vision-language agentic reasoning with a clinician-in-the-loop web interface, TEAM achieves stronger results on 88 benchmarks across 51 multi-institutional cohorts.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.