Primary tumor site is a useful predictor of cetuximab efficacy in the third-line or salvage treatment of KRAS wild-type (exon 2 non-mutant) metastatic colorectal cancer: a nationwide cohort study

Monoclonal antibodies targeting the epidermal growth factor receptor (EGFR), either used alone, or a combination with cytotoxic agents have been demonstrated to prolong survival in patients with metastatic colorectal cancer (CRC) harboring KRAS wild-type or expanded RAS [1–6]. However, not all patients experienced clinical benefits of the anti-EGFR antibody treatment. These studies have emphasized the importance of additional predictive biomarkers for anti-EGFR antibody treatment. Some predictive biomarkers, such as the gene expression of EGFR ligands, have been reported to correlate with patient responses after the anti-EGFR antibody treatment [7, 8]. Besides, the primary resistance mechanisms of cetuximab have been investigated rigorously. The negative predictive roles of expanded RAS [KRAS (exons 2, 3, and 4) and NRAS (exons 2, 3, and 4)] have been well established, but other biomarkers, including BRAF ^V600E mutation, amplification of KRAS, MET, and ERBB2, and cross-talk with PI3K/Akt/PETN,, still remain to be investigational [5, 6, 9–17].

Left- and right-sided CRC (LCRC and RCRC) have different clinicopathological and molecular characteristics [18–20]. Recently, clinical studies have shown that a left-sided primary tumor site was associated with the benefits of cetuximab, which is one of the anti-EGFR antibodies, in patients with KRAS wild-type (exon 2 nonmutant) CRC [21, 22]. A subgroup analysis of the AIO KRK-0306 trial revealed similar findings in patients with expanded RAS wild-type CRC [23]. The definitive reasons for this phenomenon remain unknown. Because most of the aforementioned studies have investigated Western populations, whether there is a similar association between a primary tumor site and cetuximab efficacy in the Taiwanese population has yet to be determined.

In Taiwan, patients have been reimbursed by the National Health Insurance (NHI) for cetuximab administration as the third-line or salvage therapy for KRAS wild-type (exon 2 nonmutant) metastatic CRC since August 1, 2009 [24]. In this study, we used the Taiwan Cancer Registry (TCR) and NHI databases concomitantly to evaluate the association between a primary tumor site and the clinical benefits of cetuximab in patients with KRAS wild-type (exon 2 nonmutant) metastatic CRC.

In this study, we found that patients with LCRC had more clinical benefits of the third-line or salvage cetuximab treatment regarding TTD and OS than did patients with RCRC. In addition, multivariate analysis demonstrated that a primary tumor site was an independent predictor of patient prognosis. TTD, instead of traditional progression-free survival (PFS), was used as one of the endpoints of the study because the maximal reimbursement amount of cetuximab was the standard 18-week dosage and the time of disease progression in patients were not recorded in NHI and TCR databases. Our study was in agreement with several studies on Western population. In an exploratory analysis of NCIC CTG CO.17, Brulé et al demonstrated that a left-sided tumor site (splenic flexure to rectosigmoid colon) is a strong predictive factor for long PFS in patients with refractory, metastatic, and KRAS wild-type (exon 2) colon cancer receiving cetuximab treatment [27]. Two AIO KRK studies (0104 and 0306) demonstrated that in populations with either KRAS wild-type (codon 12/13) or expanded RAS, a left primary tumor site was associated with long PFS and OS in untreated metastatic CRC patients who received cetuximab-containing regimens [21, 23]. However, based on our research, the current study is the first to demonstrate an association between the primary tumor site of CRC and the clinical benefits of cetuximab treatment in the East Asian population.

The definitive reasons for different clinical benefits of cetuximab treatment in these patients remain unclear. Several studies have revealed different molecular and clinicopathological features between left-sided and right-sided CRC [18–20, 28]. For example, RCRC is characterized by features such as microsatellite instability phenotype, RAS mutation, mitogen-activated protein kinase activation, BRAF ^V600E mutation, BRAF-like characteristics, and the CpG island methylator phenotype [18, 20]. By contrast, LCRC is characterized by chromosomal instability, amplification of EGFR and ERBB2, EGFR pathway upregulation, and WNT, MYC, and SRC pathway activation [18]. Because the aforementioned association of a primary tumor site with cetuximab treatment was also noted in an expanded RAS wild-type population, primary expanded KRAS (exon 2, 3, 4) or NRAS (exon 2, 3, 4) mutations may not explain the different clinical benefits of cetuximab treatment in patients with different primary tumor sites. Whether emergence of new KRAS mutations played a role is unknown. Several studies have suggested that the high gene expression of EGFR ligands (epiregulin and amphiregulin) predicted favorable outcomes in patients receiving cetuximab treatment, which may explain our findings [7, 8]. BRAF ^V600E has been shown to have a strong causal relationship with resistance to anti-EGFR antibodies in preclinical models, but the correlation in clinical settings is not statistically significant [9, 12, 13, 17]. There are several gene alterations involved in the EGFR signaling pathway beyond RAS and BRAF mutations, which converge biochemically on activation of RAS/MEK/ERK, but their relevance to the de novo resistance to anti-EGFR antibody treatment remains undetermined.

Although many studies have shown promising efficacies of anti-EGFR antibody therapy in KRAS or expanded RAS wild-type metastatic CRC, the heterogeneity of CRC should be considered if different benefits were found in patient subsets. Our study showed a poor cetuximab treatment efficacy in patients with KRAS wild-type RCRC (cecum to hepatic flexure), thus emphasizing an urgent unmet clinical need in such patients. Bevacizumab is a vascular endothelial growth factor inhibitor that has shown clinical benefits in patients with advanced CRC [6, 29]. In the post hoc analysis of an AIO KRK-0306 trial, patients with RCRC appeared to have a favorable outcome in the bevacizumab plus FOLFIRI arm [23]. Whether bevacizumab is a superior choice for patients with RAS wild-type RCRC must be validated in a randomized clinical trial. Recently, a combination of an anti-EGFR antibody and an MEK inhibitor has been expected to overcome the resistance emerging from KRAS mutations or cross talk with the PI3K/Akt/PETN pathway after anti-EGFR antibody treatment, and the upfront use of these regimens could be considered in these patient subsets [30].

There are limitations of the current study. First, this was a nationwide cohort study, and we identified patients with KRAS wild-type CRC according to their usage of reimbursed cetuximab. Thus, we could not perform expanded KRAS or NRAS analyses during the study period. Moreover, the schedule and dosage of cetuximab could not be uniform, which might have confounded our results, particularly for TTD. Second, bevacizumab was not reimbursed for advanced CRC until June 1, 2011 (in a first-line setting only) but was approved by the Taiwan Food and Drug Administration in 2005. Previous studies have shown that patients with different primary tumor sites might derive different benefits from bevacizumab combined with chemotherapy [23, 31]. A few patients who used self-paid bevacizumab might have confounded our results. Third, some experts have postulated that molecular features of CRC change gradually with the bowel; thus, it might be oversimplified to classify heterogeneous CRC as only a left- or right-sided group [32]. Moreover, although Brulé et al revealed primary tumor site was not a prognostic factor in refractory CRC patients in NCIC CO.17, the prognostic role of it in KRAS or RAS wild type, metastatic CRC patients remains unknown [22]. We could not exclude the probability that left-sided tumor was a favorable prognostic factor. The long interval between TTD and OS may also imply other confounding factors were ignored. However, our study clearly demonstrated that the primary tumor site (left- or right-sided) is a useful biomarker for predicting the prognosis after cetuximab treatment in patients with advanced KRAS wild-type (exon 2 nonmutant) CRC. Furthermore, our nationwide study had the advantages of evaluating OS and preventing selection bias, because no eligible patients were lost to follow-up.

This study demonstrated that a left-sided primary tumor site is a useful predictive marker for improved cetuximab efficacy for the third-line or salvage treatment among patients with KRAS wild-type (exon 2 nonmutant) metastatic CRC. Our study results emphasize the unmet medical needs in patients with a right-sided tumor site and provide factual survival data for future clinical trials.