Abstract
IMPORTANCE: Universal tumor screening for Lynch syndrome (LS) in colorectal cancer (CRC) is recommended and involves up to 6 sequential tests. Somatic gene testing is performed on stage IV CRCs for treatment determination. The diagnostic workup for patients with CRC could be simplified and improved using a single up-front tumor next-generation sequencing test if it has higher sensitivity and specificity than the current screening protocol. OBJECTIVE: To determine whether up-front tumor sequencing (TS) could replace the current multiple sequential test approach for universal tumor screening for LS. DESIGN, SETTING, AND PARTICIPANTS: Tumor DNA from 419 consecutive CRC cases undergoing standard universal tumor screening and germline genetic testing when indicated as part of the multicenter, population-based Ohio Colorectal Cancer Prevention Initiative from October 2015 through February 2016 (the prospective cohort) and 46 patients with CRC known to have LS due to a germline mutation in a mismatch repair gene from January 2013 through September 2015 (the validation cohort) underwent blinded TS. MAIN OUTCOMES AND MEASURES: Sensitivity of TS compared with microsatellite instability (MSI) testing and immunohistochemical (IHC) staining for the detection of LS. RESULTS: In the 465 patients, mean age at diagnosis was 59.9 years (range, 20-96 years), and 241 (51.8%) were female. Tumor sequencing identified all 46 known LS cases from the validation cohort and an additional 12 LS cases from the 419-member prospective cohort. Testing with MSI or IHC, followed by BRAF p.V600E testing missed 5 and 6 cases of LS, respectively. Tumor sequencing alone had better sensitivity (100%; 95% CI, 93.8%-100%) than IHC plus BRAF (89.7%; 95% CI, 78.8%-96.1%; P = .04) and MSI plus BRAF (91.4%; 95% CI, 81.0%-97.1%; P = .07). Tumor sequencing had equal specificity (95.3%; 95% CI, 92.6%-97.2%) to IHC plus BRAF (94.6%; 95% CI, 91.9%-96.6%; P > .99) and MSI plus BRAF (94.8%; 95% CI, 92.2%-96.8%; P = .88). Tumor sequencing identified 284 cases with KRAS, NRAS, or BRAF mutations that could affect therapy for stage IV CRC, avoiding another test. Finally, TS identified 8 patients with germline DPYD mutations that confer toxicity to fluorouracil chemotherapy, which could also be useful for treatment selection. CONCLUSIONS AND RELEVANCE: Up-front TS in CRC is simpler and has superior sensitivity to current multitest approaches to LS screening, while simultaneously providing critical information for treatment selection.