Using Predictive Biomarkers to Select Patients With Advanced Colorectal Cancer for Treatment With Epidermal Growth Factor Receptor Antibodies

Abstract

As a result of the data emerging during the course of 2008, including presentations at the 44th Annual Meeting of the American Society of Oncology, it has generally been accepted that selection of patients with metastatic colorectal cancer (mCRC) for treatment with epidermal growth factor receptor (EGFR) antibodies— cetuximab or panitumumab—is reliant on the KRAS status of the tumor. For some time, KRAS had been suggested as a predictive marker for resistance to EGFR monoclonal antibodies, but Amado et al 1 were the first to publish conclusive data demonstrating the relationship between KRAS status and panitumumab efficacy in their analysis of tumor sections from participants in a randomized phase III trial comparing panitumumab with best supportive care. They found that both response to panitumumab monotherapy and improvement in progression-free survival (PFS) were confined to patients with wildtype (WT) KRAS.KRAS mutations were detected in 43% of patients, none of whom responded to panitumumab. Analyses of KRAS status and response to cetuximab have revealed similar results. For instance, in the first-line treatment of mCRC, a retrospective analysis of the impact of KRAS status on PFS and response rate on patients treated with folinic acid, fluorouracil, and irinotecan with or without cetuximab for mCRC within the Cetuximab Combined with Irinotecan in First-Line Therapy for Metastatic Colorectal Cancer trial identifiedKRAS mutations in 35.6% of KRAS-assessable patients. 2 ForKRAS WT patients, both PFS (9.9v 8.7 months; HR, 0.68; P .017) and response rate (59.3% v 43.2%; P .0025) were significantly improved by the addition of cetuximab to folinic acid, fluorouracil, and irinotecan. In contrast, for patients withKRAS mutations, there was no significant difference in either PFS (7.6v 8.1 months; HR, 1.07;P .47) or response rate (40.2%v 36.2%; P .46) with the addition of cetuximab. For patients treated with first-line infused fluorouracil, folinic acid, and oxaliplatin with or without cetuximab in the Oxaliplatin and Cetuximab for First-Line Treatment of Metastatic Colorectal Cancer study, the improved response rate and PFS associated with cetiximab was also limited to those with KRAS WT tumors. 3 In fact, in this study, patients with mutated KRAS receiving cetuximab had poorer outcomes compared with those receiving infusional fluorouracil, leucovorin, and oxaliplatin alone. Based on these results, it is now recommended thatKRAS testing to exclude the presence of mutations should be performed when considering either panitumumab or cetuximab in the treatment of advanced colorectal cancer. Ligand-receptor activation of the EGFR at the cell surface results in homo- or heterodimerization of the receptors and triggers activation of downstream signaling pathways. 4 The incidence ofKRAS mutations in colorectal tumors is approximately 35% to 45%. KRAS mutations can result in constitutive activation of the Ras-Raf-MAPkinase pathway, one of the major EGFR downstream pathways, and therefore confer resistance to EGFR antibodies. 5 Nonmutated KRAS does not, however, guarantee benefit from treatment with EGFR monoclonal antibodies. Consequently, even with routine testing of KRAS status, a significant proportion of patients will be exposed to these drugs and their associated toxicity without deriving any benefit. In this issue ofJournalofClinicalOncology, Di Nicolantonio et al 6 evaluate the role ofBRAF mutations as prognostic or predictive factors for response to cetuximab or panitumumab and explore potential ways to circumvent inherent pathways of resistance. In this hypothesis-generating study, a retrospective analysis of 113 patients treated with either cetuximab or panitumumab was conducted. Additionally, a cellular analysis of the effect of the BRAF V600E allele on response to cetuximab or panitumumab was performed. The presence ofKRAS mutations in this population was 30%. As expected, KRAS mutations were associated with a lack of response to EGFR antibodies (P .011) and shorter PFS (P .0275) compared with KRAS WT tumors. Twenty-eight percent of KRAS WT patients responded to either cetuximab (alone or in combination with irinotecan) or panitumumab. BRAF mutations (BRAF V600E allele) were identified in 11 patients (10% of the population evaluated; 14% of KRAS WT patients). BRAF and KRAS mutations were mutually exclusive, as observed in previous studies. 7 Supporting the authors’ hypothesis that in KRAS WT tumors, BRAF mutations could have predictive value, none of the 11 BRAF-mutated tumors responded to treatment. Similarly, in the laboratory, colorectal cancer cell lines carrying the BRAF V600E allele were highly refractory to cetuximab and panitumumab. Conversely, all 22 patients who responded to KRAS WT also had BRAF WT. BRAF-mutated tumors were also associated with shorter PFS and overall survival irrespective of KRAS status (P .0107 and P .0001, respectively), suggesting a potential role of BRAF as a prognostic biomarker. Interestingly, targeting BRAF-mutated cell lines with the combination of cetuximab and sorafenib resulted in much higher response rates than those observed