Combination therapy is a promising therapeutic strategy to enhance the efficacy of immune checkpoint blockade (ICB); however, predicting drugs for effective combination is challenging. Here we developed a general data-driven method called CM-Drug for screening compounds that can boost ICB treatment efficacy based on core and minor gene sets identified between responsive and nonresponsive samples in ICB therapy. The CM-Drug method was validated using melanoma and lung cancer mouse models, with combined therapeutic efficacy demonstrated in eight of nine predicted compounds. Among these compounds, taltirelin had the strongest synergistic effect. Mechanistic analysis and experimental verification demonstrated that taltirelin can stimulate CD8⁺ T cells and is mediated by the induction of thyroid-stimulating hormone. This study provides an effective and general method for predicting and evaluating drugs for combination therapy and identifies candidate compounds for future ICB combination therapy.

联合治疗是一种有前景的治疗策略，可增强免疫检查点阻断（ICB）的功效；然而，预测药物的有效组合具有挑战性。在这里，我们开发了一种称为 CM-Drug 的通用数据驱动方法，用于根据 ICB 治疗中有反应和无反应样本之间识别的核心和次要基因集筛选可以提高 ICB 治疗功效的化合物。CM-Drug 方法使用黑色素瘤和肺癌小鼠模型进行了验证，九种预测化合物中的八种证明了联合治疗功效。在这些化合物中，他替瑞林具有最强的协同作用。机理分析和实验验证表明，他替瑞林可以刺激CD8 ⁺ T细胞，并且是通过促甲状腺激素的诱导介导的。这项研究为预测​​和评估联合治疗药物提供了一种有效且通用的方法，并确定了未来 ICB 联合治疗的候选化合物。