LAUSR

Mutation Frequencies in Patients With Early-Onset Colorectal Cancer To the Editor Pearlman et al1 show what can be achieved through the cooperation of cancer registries with pathologists and other clinicians. The study participants were 450 newly identified young men and women with colon cancer who were treated at one of 51 Ohio hospitals. Both blood and paraffinembedded tumor samples were submitted to the core molecular and pathology laboratories. Using a panel of 25 cancer susceptibility genes as a genetic screening test, 72 patients were found to carry a pathogenic mutation; of these, 59 mutations were in genes associated with colon cancer and 13 mutations were in genes associated with breast or pancreatic cancer but not (classically) with colon cancer. We expect it might be difficult to follow the recommendation of coupling microsatellite instability (MSI) analysis with germline mutation analysis outside research studies. Mismatch repair (MMR) testing requires the pathologist to dissect tumor tissue samples and assay these with 7 molecular markers. In the past, sequencing was expensive and MMR was used as a tool for triaging patients who might undergo germline testing. The cost of sequencing is no longer a critical issue, and therefore it is rational to offer sequencing to all patients with colon cancer and to individuals with a family history of colon cancer, regardless of MSI status. Mismatch repair testing of the 450 patients yielded 9 patients with MMR but without a mutation (2%). The authors’ recommendation for universal MMR testing is based on personalizing treatment according to MSI status. An early study indicates that cancers with MMR deficiencies often respond to anti-PD1 (programmed cell death protein 1) immunotherapy.2 This is encouraging, but the cited study was a small phase 2 study with heavily treated metastatic colon cancer patients. In the case of breast cancer, there are blood-based tests for genetic susceptibility (eg, BRCA1 and BRCA2) and tumor-based tests for treatment response (eg, ERBB2/HER2, estrogen receptor) and the patient populations eligible for the 2 tests are different. By proposing that the standard genetics protocol include contemporaneous assessment of MSI and germline mutations, we create the potential for excluding many patients for the sake of obtaining a small amount of additional information of uncertain value. The practical question is should we test all individuals with colon cancer for a panel of susceptibility genes? It is straightforward to endorse testing for the 6 colon cancer genes (MLH1, MSH2, MUTYH, MSH6, APC, and SMAD4), but it is not clear that we should test patients with colon cancer for unrelated genes such as BRCA1 and BRCA2 just because it is easy to do so. Cross-test counseling is complicated, and the interpretation of variants is an ongoing challenge. Of the 450 patients, 67 had a variant of unknown significance in APC, CHEK2, or ATM. We favor a simple policy that makes available inexpensive testing for 6 colon cancer susceptibility genes to all patients with colon cancer and to their close relatives.