MRE11-Deficiency Associated with Improved Long-Term Disease Free Survival and Overall Survival in a Subset of Stage III Colon Cancer Patients in Randomized CALGB 89803 Trial

Purpose: Colon cancers deficient in mismatch repair (MMR) may exhibit diminished expression of the DNA repair gene, MRE11, as a consequence of contraction of a T-11 mononucleotide tract. This study investigated MRE11 status and its association with prognosis, survival and drug response in patients with stage III colon cancer. Patients and Methods: Cancer and Leukemia Group B 89803 (Alliance) randomly assigned 1,264 patients with stage III colon cancer to postoperative weekly adjuvant bolus 5-fluorouracil/leucovorin (FU/LV) or irinotecan+FU/LV (IFL), with 8 year follow-up. Tumors from these patients were analyzed to determine stability of a T-11 tract in the MRE11 gene. The primary endpoint was overall survival (OS), and a secondary endpoint was disease-free survival (DFS). Non-proportional hazards were addressed using time-dependent covariates in Cox analyses. Results: Of 625 tumor cases examined, 70 (11.2%) exhibited contraction at the T-11 tract in one or both MRE11 alleles and were thus predicted to be deficient in MRE11 (dMRE11). In pooled treatment analyses, dMRE11 patients showed initially reduced DFS and OS but improved long-term DFS and OS compared with patients with an intact MRE11 T-11 tract. In the subgroup of dMRE11 patients treated with IFL, an unexplained early increase in mortality but better long-term DFS than IFL-treated pMRE11 patients was observed. Conclusions: Analysis of this relatively small number of patients and events showed that the dMRE11 marker predicts better prognosis independent of treatment in the long-term. In subgroup analyses, dMRE11 patients treated with irinotecan exhibited unexplained short-term mortality. MRE11 status is readily assayed and may therefore prove to be a useful prognostic marker, provided that the results reported here for a relatively small number of patients can be generalized in independent analyses of larger numbers of samples.