These data raise the question of whether

the survival of these 2 subtypes of stage III N1 patients treated with FOLFOX might be similar to a stage II population. In a review of data for stage II colon cancers from adjuvant chemotherapy trials that evaluated FOLFOX, 25, 40, 41 and 42 reported DFS rates are similar to those observed in our stage III N1 tumors without BRAFV600E or KRAS mutations or in the dMMR subtypes. This finding suggests that N1 pMMR tumors without BRAFV600E or KRAS mutations may have an intrinsically better prognosis, irrespective of therapy, or alternatively, may receive greater benefit from FOLFOX vs the other subtypes. The situation in dMMR tumors is more complex given data suggesting lack of 5-FU benefit 43 and the unknown benefit, if any, of oxaliplatin Apoptosis inhibitor combined with 5-FU/leucovorin in stage III dMMR patients. 19 Although the prognostic impact of molecular subtypes in N1 cancers was similar to the overall cohort, we unexpectedly observed poor DFS for N2 dMMR

sporadic tumors, which was not significantly different from the poor prognosis of N2 pMMR tumors with mutant BRAFV600E or mutant http://www.selleckchem.com/products/dabrafenib-gsk2118436.html KRAS. However, this finding was not observed among N2 dMMR tumors of the familial subtype that maintained their favorable HRs, and an explanation awaits further research. The mutant KRAS pMMR subtype had the highest percentage of African Americans compared with the other subtypes, consistent with data indicating higher rates of KRAS mutations in CRCs from

African Americans. 44 and 45 Conflicting data have been reported for the frequency of dMMR/MSI in CRCs from African Americans compared with whites, 45 yet our study does not demonstrate a difference in the rate of African Americans by MMR status. Our data for mutant KRAS, albeit preliminary due to small patient numbers of non-white race, suggest that colon cancers from African Americans may be associated with this poor prognostic subtype. Our C1GALT1 findings support limited data demonstrating the ability of subtype classifications to predict clinical outcomes. Recently, a CRC subtype classification20 was applied to tumor tissues from the Iowa Women’s Health Study, which found differences in age at diagnosis, tumor site, and histologic grade across 3 CRC subtypes defined by combinations of MSI, CIMP, BRAF, and KRAS status. However, no statistically significant differences in survival were found across the tumor subtypes in this smaller cohort that was limited to women. 20 In contrast to our study, the authors defined a mutant BRAFV600E serrated subtype without regard to MSI status and did not distinguish the MSI-high familial subtype as a distinct group. 20 Data shown here and elsewhere 21 and 37 suggest that the serrated neoplasia pathway can give rise to colon cancer subtypes with divergent prognoses. Our subtype classification was more informative than analysis of individual biomarkers.