https://doi.org/10.7554/eLife.100490.4
We translate membrane-mechanics biology into clinical tools. Our deep-learning pipelines analyze IHC and H&E pathology to detect TRPV4 localization patterns and to stratify DCIS risk at the lesion and patient levels. By linking image-derived features to time-to-progression and therapy markers, we aim to provide decision support that complements pathologist expertise. In tandem, we quantify how mechanical priming and receptor enrichment create druggable vulnerabilities, closing the loop from mechanism to biomarker to intervention.