Heterogeneity of KRAS, NRAS, BRAF and PIK3CA mutations in metastatic colorectal cancer and potential effects on therapy in the CAPRI GOIM trial.

BACKGROUND Evidence suggests that metastatic colorectal carcinoma (mCRC) has a high level of intratumor heterogeneity. We carried out a quantitative assessment of tumor heterogeneity for KRAS, NRAS, BRAF and PIK3CA mutations, in order to assess potential clinical implications. PATIENTS AND METHODS Tumor samples (n = 182) from the CAPRI-GOIM trial of first-line cetuximab + FOLFIRI in KRAS exon-2 wild-type mCRC patients were assessed by next-generation sequencing that allows quantitative assessment of mutant genes. Mutant allelic frequency was normalized for the neoplastic cell content and, assuming that somatic mutations usually affect one allele, the Heterogeneity Score (HS) was calculated by multiplying by 2 the frequency of mutant alleles in neoplastic cells. Therefore, HS virtually corresponds to the fraction of neoplastic cells carrying a specific mutation. RESULTS The KRAS HS ranged between 12 and 260 with mean value of 87.1 and median value of 84.4, suggesting that in most CRC, the majority of neoplastic cells carry mutant KRAS. Similar findings were observed for NRAS (HS range 35.5-146.7; mean 102.8; median 117.1). In contrast, in BRAF (HS range 17.1-120; mean 54.8; median 54.3) and PIK3CA (HS range 14.3-120; mean 59.5; median 47.3) mutant cases, only a fraction of neoplastic cells seem to carry the mutant allele. The response rate was 70% in KRAS mutant patients with an HS <33 (low KRAS; n = 10) and 45.7% in KRAS HS >33 patients (high KRAS; n = 35); median progression-free survival were 7.97 and 8.37 months, respectively. Low-KRAS tumors had a higher frequency of additional mutations in PIK3CA when compared with high-KRAS (6/10 versus 8/35). CONCLUSIONS KRAS and NRAS mutations are usually present in the majority of neoplastic cells, whereas BRAF and PIK3CA mutations often affect a limited fraction of transformed cells. Resistance to cetuximab in low-KRAS patients might be driven by the complex mutational profile rather than KRAS mutation load.