Mismatch Repair Deficiency in Colorectal Cancers: Is Somatic Genomic Testing the Grab-Bag for All Answers?

Abstract

Colorectal cancer (CRC) is a large public health problem and is the third most common cancer in men and women. It accounts for approximately 50,000 deaths each year and is largely preventable with recommended population screening. Multiple studies have estimated that approximately 3% of CRCs are attributable to Lynch syndrome (LS). This autosomal dominant genetic disorder is associated with greatly increased risks of developing colorectal, endometrial, and other cancers. Despite the increased risk of cancers seen in patients with LS, long-term follow-up studies show that adherence to current surveillance recommendations is highly successful. Unfortunately, unless there is strong clinical suspicion, many cases of LS are missed. There is therefore a clear need to explore universal screening for LS in the population. Universal screening is feasible because the diagnosis of LS is primarily based on the presence of a germline mutation in a mismatch repair (MMR) gene. The tumors of patients with LS characteristically demonstrate MMR deficiency, defined as the presence of microsatellite instability (MSI) or loss of expression of the MMR protein (detected via immunohistochemistry [IHC]), which are the cellular hallmarks of this disorder. In a small, proof-of-principle study recently published, researchers reported that they could predict the benefit of an anti–PD-1 inhibitor called pembrolizumab by scanning patients’ tumor samples for defects in MMR. Regardless of their type of cancer, patients whose tumors were MMR deficient (MMR-D) were more likely to respond to the immune-boosting anti–PD-1 drug than those with tumors proficient in MMR. This has generated renewed interest in assessing patients for evidence of MMR deficiency. In the article that accompanies this editorial, Stadler et al hypothesized that a next-generation sequencing (NGS) panel for colorectal tumors that identifies RAS/BRAF and other actionable somatic mutations could also reliably identify tumors with DNA MMR protein deficiency on the basis of increased mutational load. The authors identified all CRCs that underwent genomic mutation profiling with a custom NGS assay (MSK-IMPACT). Tumor mutational load, with exclusion of copy number changes, was determined for each case and compared with MMR status as determined by routine IHC. Two hundred twenty-four samples underwent analysis with both the MSK-IMPACT 341-gene assay and IHC. Thirteen percent (n 5 28) exhibited MMR-D by IHC. Using the 341-gene assay, 100% of the 193 tumors with , 20 mutations were MMR proficient (MMR-P). Of 31 tumors with $ 20 mutations, 28 (90%) were MMR-D. The three remaining tumors were easily identified as being distinct from the MMR-D tumors, with . 150 mutations each. All of these latter tumors harbored the P286R hotspot POLE mutation, consistent with the ultramutator phenotype. Among MMR-D tumors, the median number of mutations was 50 (range, 20 to 90) compared with six (range, 0 to 17) in MMR-P/POLE wild-type tumors (P , .001). With a mutational load cutoff of $ 20 and , 150 for MMR-D detection, sensitivity and specificity were both 1.0 (95% CI, 0.93 to 1.0). The authors successfully demonstrate that a cutoff for mutational load can be identified via multigene NGS tumor profiling, providing an alternative means of screening for MMR-D in the same assay that is used for tumor genotyping. The authors’ findings that mutational load correlates with MMR-D status is not novel in and of itself. What is perhaps of greater clinical importance is determining how such information should be used to screen for MMR status in patients with CRC, if at all. Are we assessing the MMR status to select for potential immunotherapy agents, or to identify possible underlying LS? Should we sequence primary CRC, metastases, normal germline, or all of the above? Unfortunately, this question cannot be easily addressed by this study alone, because its experimental design used a mix of primary and metastatic CRC sites. Within the MMR-P/ POLE wild-type group, the multigene assay was performed on the primary tumor in 59% of cases and on a metastatic lesion in 41% of cases; the mean and median mutational load did not differ between primary tumors and metastases. Among the 28 MMR-D cases, the multigene assay was performed on the primary lesion in 86% of cases and on a metastatic lesion in 14% of cases. The authors did remark that in a subset analysis, they found a significant difference between the MMR-P and MMR-D cases, irrespective of whether primary tissue or metastasis was analyzed.