Cancer immunotherapy with immune-checkpoint blockade (ICB) is considered a promising strategy for cancer treatment. Identifying predictive biomarkers and developing efficient computational models to predict the ICB response are important issues for successful immunotherapy. Here, we present a concise and intuitive survey of the computational issues for ICB response prediction, providing a summary of the available predictive biomarkers and building of one-stop machine-learning models that integrate biomarkers calculable from high-throughput sequencing (HTS) data. Several points for discussion are highlighted to inspire further research for improving ICB treatment. Continuing efforts are required to improve ICB response prediction and to identify novel predictive biomarkers by taking advantage of the rapid development of computational models and HTS techniques for effective and personalized cancer immunotherapy.

带有免疫检查点封锁（ICB）的癌症免疫疗法被认为是一种有前途的癌症治疗策略。鉴定预测性生物标志物并开发有效的计算模型以预测ICB反应是成功进行免疫治疗的重要问题。在这里，我们对ICB反应预测的计算问题进行了简明而直观的调查，总结了可用的预测生物标记，并建立了一站式机器学习模型，该模型集成了可从高通量测序（HTS）数据计算得出的生物标记。着重讨论的几点要点，以激发进一步的研究来改善ICB的治疗方法。