LBA14 - Molecular subtype and chemotherapy-related toxicity in stage 3 colon cancers: NCCTG N0147

Date 27 September 2014
Event ESMO 2014
Session Gastrointestinal tumours, colorectal
Topics Anticancer Agents
Complications/Toxicities of Treatment
Colon and Rectal Cancer
Biological Therapy
Presenter Frank Sinicrope
Citation Annals of Oncology (2014) 25 (5): 1-41. 10.1093/annonc/mdu438
Authors F. Sinicrope1, Q. Shi2, T. Smyrk3, R. Goldberg4, M. Mahoney2, D. Sargent2, S. Alberts1
  • 1Medical Oncology, Mayo Clinic, 55905 - Rochester/US
  • 2Health Sciences Research-cancer Center Statistics, Mayo Clinic, 55905 - Rochester/US
  • 3Laboratory Medicine And Pathology, Mayo Clinic, 55905 - Rochester/US
  • 4Division Of Medical Oncology, Ohio State University School of Medicine, James Cancer Hospital and Solove Research Institute, Columbus/US



The association of pathway-based colon cancer subtypes with treatment-related adverse events (AEs) was examined in 2,381 stage 3 patients from a randomized trial of FOLFOX +/- cetuximab.


Tumors were categorized into 5 subtypes using DNA mismatch repair (MMR) status and mutually exclusive mutations in BRAFV600E or KRAS. Three subtypes had proficient (p) MMR: mutant BRAF, mutant KRAS, or without either mutation. Two subtypes had deficient (d) MMR: sporadics with mutant BRAF or hypermethylation of MLH1, or familial without BRAF mutation or MLH1 hypermethylation. Association between grade 3+ AEs and subtypes was determined by Chi-squared test (univariate) and logistic regression (multivariate, interactions).


Subtype distribution was similar between treatment arms (p = .62). Among sporadic dMMR patients, 77% completed > 6 treatment cycles vs 87 -91% of patients in other subtypes (age adjusted p = .029). Overall grade 3+ AEs among patients receiving < 12 cycles was highest for sporadic dMMR (81%) and lowest for familial dMMR (40%) subtypes (p = .016). The sporadic dMMR subtype had the highest rate of grade 3+ diarrhea (p < .0001). Statistically significant interactions were found between subtypes and primary tumor location (p = .008). Among distal but not proximal cancers, sporadic dMMR pts had the highest AE rate (78%); mutant BRAFV600E pMMR pts had the lowest (33%) [overall p =.005]. Grade 3+ AEs in patients > 70 yrs were similar among subtypes (p = .60). Odd ratios (ORs) for comparison of sporadic dMMR to other subtypes are shown (Table). Only the association of distal tumor location with higher AE rates for sporadic dMMR patients was statistically significant by multivariable analysis (p = .011).

Univariate associations between overall grade 3+ AEs and molecular subtypes.

Completed < 12 cycles Distal tumor location
Subtype & frequency % OR (95% CI) P OR (95% CI) P
Non mutant KRAS, BRAF pMMR (55%) .41 (.21-.79) 0.008 .41 (.08-199) 0.268
Mutant KRAS pMMR (27%) .41 (.21-.84) 0.014 .46 (.09-2.25) 0.336
Mutant BRAF pMMR (7.9%) .56 (.23-1.38) 0.21 .14 (.03-0.77) 0.024
Familial dMMR (2.7%) .16 (.05-.53) 0.003 .23 (.03-1.77) 0.158
Sporadic dMMR (7.5%) ref ref


Differences in treatment-related toxicity were found by tumor subtype with sporadic dMMR patients receiving fewer treatment cycles and experiencing greater toxicity, especially among distal cancers after adjustment for covariates.


All authors have declared no conflicts of interest.