Predictive biomarkers for oncology are necessary to accurately identify patients who will benefit from anticancer treatment. Recently approved oncology drugs target discrete molecular aberrations or pathways in tumor cells and consequently are active on a subset of patient population, yet clinical studies have shown that not all biomarker-positive patients respond. The advancement of predictive biomarkers needs to detect novel and evolving drug resistance mechanisms, not only to guide the selection of patient subsets for specific treatments, but to identify new therapeutic targets. Going beyond the “one marker, one drug” model to incorporate genomics, transcriptomics, and receptor status assessments during biomarker-drug co-development can aid in the successful application of molecular marker-based cancer therapy. This review provides the latest update of biomarker-based cancer therapeutics approved by the US Food and Drug Administration. We provide case studies of therapeutics selectively targeting HER2, EGFR, or PD-1/PD-L1 signaling pathways. We also discuss the challenges and promising future directions in the co-development of targeted cancer therapeutics and paired predictive biomarkers.

肿瘤学的预测生物标志物对于准确识别将受益于抗癌治疗的患者而言是必不可少的。最近批准的肿瘤药物靶向肿瘤细胞中的离散分子畸变或途径，因此在一部分患者群体中具有活性，但是临床研究表明，并非所有生物标记阳性患者都能做出反应。预测性生物标志物的发展需要检测新的和不断发展的耐药机制，不仅可以指导针对特定治疗的患者亚群的选择，而且还可以确定新的治疗靶标。在生物标志物与药物的共同开发过程中，超越“单一标志物，一种药物”模型，将基因组学，转录组学和受体状态评估结合起来，可以帮助成功应用基于分子标志物的癌症疗法。这篇评论提供了由美国食品和药物管理局批准的基于生物标志物的癌症治疗剂的最新更新。我们提供了选择性靶向HER2，EGFR或PD-1 / PD-L1信号通路的疗法的案例研究。我们还将讨论靶向癌症治疗剂和配对预测性生物标记物共同开发中的挑战和有希望的未来方向。