Transcriptome-informed multi-modal AI for predicting neoadjuvant therapy response from breast cancer biopsies
arXiv cs.AI · 3 Oct 2026, 03:52 AEST
A new AI model predicts neoadjuvant therapy response in breast cancer using transcriptome data. The model shows promise in improving discrimination of treatment responders across various cohorts.
Key facts
- Scarcity of labeled data limits development of deep learning biomarkers in oncology.
- We develop a two-stage AI model predicting pathological complete response (pCR) to neoadjuvant therapy in breast cancer.
- Developed using 1,080 patients (five cohorts) and evaluated in 1,412 patients (nine cohorts), the model achieves a pooled AUROC of 0.79 (95% CI, 0.73-0.85), discriminating responders within molecular subtypes.
- Ablations show transcriptome-wide inference improves discrimination over clinical variables alone or one-stage pathology models.
Read aloud
A new A I model predicts neoadjuvant therapy response in breast cancer using transcriptome data. The model shows promise in improving discrimination of treatment responders across various cohorts. Scarcity of labeled data limits development of deep learning biomarkers in oncology. We develop a two-stage A I model predicting pathological complete response pCR to neoadjuvant therapy in breast cancer. Developed using one thousand eighty patients five cohorts and evaluated in one thousand four hundred twelve patients nine cohorts, the model achieves a pooled AUROC of zero point seven nine ninety five percent CI, zero point seven three-zero point eight five, discriminating responders within molecular subtypes. Ablations show transcriptome-wide inference improves discrimination over clinical variables alone or one-stage pathology models.
Sources in this cluster
- Transcriptome-informed multi-modal AI for predicting neoadjuvant therapy response from breast cancer biopsies
arXiv cs.AI · 3 Oct 2026, 03:52 AEST