在过去的几十年中，癌症治疗的发展主要集中在精确定位单个癌症相关分子上。尽管有了巨大的进步，但这种定向治疗仍然显示出不完整的精度和最终由于靶点异质性或变异而导致耐受性的发展。然而，最近发展的基于细胞的治疗，比如嵌合抗原受体（CAR）T细胞，为重新构架癌症靶向策略提供了革命性的机会。装备有合成回路的免疫细胞本质上是活的计算机，可以编程识别肿瘤基于多个信号，包括肿瘤细胞本质和微环境。此外，细胞还可以被编程启动广泛但高度局部化的治疗反应，可以限制逃逸的潜力，同时仍然保持高精度。虽然这些新兴智能细胞工程能力尚未在临床上得到充分实现，但我们在这里认为，当它们与基因组数据的机器学习分析相结合时，它们将变得更加强大，这可以指导设计最具有区分性和可操作性的治疗识别程序。癌症分析和合成生物学的融合可能会导致更加微妙的肿瘤识别范式，更类似于面部识别，具有更有效地解决治疗癌症的复杂挑战的能力。© 2022. Springer Nature Limited.

In the past several decades, the development of cancer therapeutics has largely focused on precision targeting of single cancer-associated molecules. Despite great advances, such targeted therapies still show incomplete precision and the eventual development of resistance due to target heterogeneity or mutation. However, the recent development of cell-based therapies such as chimeric antigen receptor (CAR) T cells presents a revolutionary opportunity to reframe strategies for targeting cancers. Immune cells equipped with synthetic circuits are essentially living computers that can be programmed to recognize tumours based on multiple signals, including both tumour cell-intrinsic and microenvironmental. Moreover, cells can be programmed to launch broad but highly localized therapeutic responses that can limit the potential for escape while still maintaining high precision. Although these emerging smart cell engineering capabilities have yet to be fully implemented in the clinic, we argue here that they will become much more powerful when combined with machine learning analysis of genomic data, which can guide the design of therapeutic recognition programs that are the most discriminatory and actionable. The merging of cancer analytics and synthetic biology could lead to nuanced paradigms of tumour recognition, more akin to facial recognition, that have the ability to more effectively address the complex challenges of treating cancer.© 2022. Springer Nature Limited.