Frequent PIK3CA Mutations in Colorectal and Endometrial Tumors With 2 or More Somatic Mutations in Mismatch Repair Genes

Stacey A Cohen,Emily H Turner,Mallory B Beightol,Angela Jacobson,Ted A Gooley,Stephen J Salipante,Sigurdis Haraldsdottir,Christina Smith,Sheena Scroggins,Jonathan F Tait,William M Grady,Edward H Lin,David E Cohn,Paul J Goodfellow,Mark W Arnold,Albert de la Chapelle,Rachel Pearlman,Heather Hampel,Colin C Pritchard
DOI: https://doi.org/10.1053/j.gastro.2016.06.004
IF: 29.4
Gastroenterology
Abstract:Background & aims: Some colorectal and endometrial tumors with microsatellite instability not attributable to MLH1 hypermethylation or germline mutations contain 2 or more somatic mutations in genes encoding mismatch repair (MMR) proteins. We sought to define the molecular phenotype of this newly recognized tumor subtype. Methods: From 2 prospective studies of the efficacy of screening for Lynch syndrome, we identified patients with colorectal and endometrial tumors who had 2 or more somatic (but not germline) mutations in genes encoding MMR proteins (double somatic). We determined the frequencies of tumor mutations in PIK3CA, BRAF, KRAS, NRAS, and PTEN by targeted next-generation sequencing and used logistic-regression models to compare them with those from patients with Lynch syndrome, MLH1-hypermethylated, or microsatellite-stable tumors. We validated our findings using independent data sets from The Cancer Genome Atlas. Results: Among colorectal cancer cases, we found that 14 of 21 (67%) patients with double somatic tumors also had PIK3CA mutations, compared with 4 of 18 (22%) tumors from patients with Lynch syndrome, 2 of 10 (20%) tumors with MLH1 hypermethylation, and 12 of 78 (15%) tumors with microsatellite stability (P < .0001 for patients with double somatic tumors vs other subgroups). Mutations in PIK3CA were detected in all 13 patients with double somatic endometrial cancers (P = .04 compared with other subgroups). We did not detect BRAF mutations in patients with double somatic colorectal tumors or Lynch syndrome. We found highly similar results in a validation cohort from The Cancer Genome Atlas (113 patients with colorectal tumors, 178 endometrial tumors); 100% of double somatic cases had a somatic mutation in PIK3CA (P < .0001 compared with other subgroups). Conclusions: Most patients with colorectal or endometrial tumors with 2 or more somatic (but not germline) mutations in MMR proteins also have mutations in PIK3CA; mutations in PIK3CA are detected at substantially higher frequencies in these double somatic tumors than in other microsatellite-instability subgroups. PIK3CA mutation status might be used to identify a specific group of colorectal tumors, and to select treatment or determine prognosis.
What problem does this paper attempt to address?