The revolution of precision medicine has produced unprecedented seismic shifts in the treatment paradigm of advanced cancers. Among the major killers, colorectal cancer (CRC) is far behind the others. In fact, the great successes obtained in breast, NSCLC, melanoma, and genitourinary tract tumors have been observed only in fewer than 5 % metastatic colorectal cancer (mCRC): those with the mismatch repair deficiency (dMMR), a well-known predictive factor for to the outstanding efficacy of checkpoint inhibitors (CPI).

The treatment of the remaining vast majority mCRC patients is still based upon only two molecular determinants: the RAS and BRAF mutational status. New promising biomarkers include HER2, tumor mutational burden (TMB) for its possible implications on CPI efficacy, and the extremely rare NTRK fusions. The Consensus Molecular Subtypes classification (CMS) is a good example of the efforts to combine different molecular features of this disease, although its relevance in clinical practice is still under investigation.

In this Review, we focus on all these prognostic and predictive biomarkers, analyzing data from the most important clinical trials of the last years. We also try to rank them according to their prognostic and predictive power.

精准医学的革命在晚期癌症的治疗范式中产生了前所未有的巨变。在主要杀手中，结直肠癌 (CRC) 远远落后于其他人。事实上，仅在不到 5% 的转移性结直肠癌 (mCRC) 中观察到在乳腺癌、非小细胞肺癌、黑色素瘤和泌尿生殖道肿瘤中取得的巨大成功：那些具有众所周知的预测因素错配修复缺陷 (dMMR) 的患者检查点抑制剂 (CPI) 的卓越功效。

其余绝大多数 mCRC 患者的治疗仍然仅基于两个分子决定因素：RAS 和 BRAF 突变状态。新的有前途的生物标志物包括 HER2、肿瘤突变负荷 (TMB) 可能对 CPI 疗效产生影响，以及极其罕见的 NTRK 融合。共识分子亚型分类 (CMS) 是努力结合这种疾病的不同分子特征的一个很好的例子，尽管它在临床实践中的相关性仍在研究中。

在这篇综述中，我们专注于所有这些预后和预测性生物标志物，分析来自过去几年最重要的临床试验的数据。我们还尝试根据它们的预后和预测能力对它们进行排名。