分析了数百种癌细胞系的基因表达谱以及细胞系对药物治疗的反应，以确定其表达与耐药性相关的基因。在 809 个癌细胞系的 GDSC 数据集中，36 个基因的表达与许多抗癌药物的耐药性（IC50 增加）相关。这在 860 个细胞系的 CTRP 数据集中得到了验证。从癌细胞系 UAB36 中 36 个基因的相关系数得出的多基因评分预测了细胞系对他莫昔芬的耐药性。尽管 36 个基因是从细胞系行为中选择的，但 UAB36 成功预测了三个不同的接受他莫昔芬治疗的患者队列中乳腺癌患者的生存情况。 UAB36 优于两个现有的预测基因特征，并且是乳腺癌患者结果的预测因子，独立于影响结果的已知临床协变量。这种方法应该为许多癌症类型对特定药物的耐药性提供有前途的多基因生物标志物。© 2024。作者。

Gene expression profiles of hundreds of cancer cell-lines and the cell-lines' response to drug treatment were analyzed to identify genes whose expression correlated with drug resistance. In the GDSC dataset of 809 cancer cell lines, expression of 36 genes were associated with drug resistance (increased IC50) to many anti-cancer drugs. This was validated in the CTRP dataset of 860 cell lines. A polygenic score derived from the correlation coefficients of the 36 genes in cancer cell lines, UAB36, predicted resistance of cell lines to Tamoxifen. Although the 36 genes were selected from cell line behaviors, UAB36 successfully predicted survival of breast cancer patients in three different cohorts of patients treated with Tamoxifen. UAB36 outperforms two existing predictive gene signatures and is a predictor of outcome of breast cancer patients independent of the known clinical co-variates that affect outcome. This approach should provide promising polygenic biomarkers for resistance in many cancer types against specific drugs.© 2024. The Author(s).