Multi-objective domain adaptation for pan-cancer transfer of pharmacogenomic drug response to patient tumours

MODAE+ transfers pharmacogenomic drug-response models from 1,293 CCLE cell lines to 9,696 TCGA tumours, preserving cancer-lineage structure while improving survival-linked sensitivity predictions.

• Why it matters: Drug-sensitivity measurements from cell lines do not reliably carry over to patient tumours due to biological heterogeneity and cross-domain distribution shift.
• What they did: MODAE+ is a pan-cancer multi-objective domain adaptation framework that jointly optimizes domain alignment, deconfounding, drug-response prediction, survival prediction, and cancer-lineage classification using CTRP sensitivity for 544 compounds in 738 matched cell lines.
• The result: For TCGA patients receiving shared drugs, predicted sensitivity is linked to overall survival in 46.3% of cancer–drug pairs by log-rank testing and 41.5% by multivariable Cox analysis, and signature concordance shows Spearman correlations of 0.791 (paclitaxel) and 0.724 (doxorubicin).

AI summary of the abstract · may contain errors · not medical advice · read the paper.