Epidermal growth factor receptor (EGFR) is a receptor tyrosine kinase that binds extracellular proteins of the EGF family and plays a critical role in oncogenesis. EGFR belongs to the ErbB family, which includes ErbB2 (human EGFR2 [HER2/neu]), ErbB3 (HER3), and ErbB4 (HER4). EGFR is a glycoprotein that contains 3 major components: an extracellular ligand binding domain, a hydrophobic transmembrane domain, and an intracellular tyrosine kinase domain. Specific ligands for each of the ErbB family receptors interact with receptors that have different affinities. EGFR remains in a state of autoinhibition in the absence of ligands like EGF and transforming growth factor alpha (TGFα). Ligand binding alters conformation of the protein and exposes the dimerization loop, which subsequently forms homodimers or heterodimers with other member of the ErbB family. Dimerization triggers the phosphorylation of tyrosine kinases in the EGFR intraceullar domain, leading to subsequent activation in downstream signaling pathways. Because of its role in the development and progression of colon cancer, EGFR has become an attractive target for therapy in the colorectal cancers (CRCs).
Cetuximab is a chimeric monoclonal antibody (immunoglobulin G1 [IgG1]) that competitively binds to the EGFR extracellular domain with higher affinity than endogenous ligands. In vitro and in vivo studies have demonstrated that binding of cetuximab to EGFR blocks phosphorylation and activation of the tyrosine kinase and subsequently inhibits downstream signal transduction, resulting in inhibition of cell growth and induction of apoptosis. Cetuximab can mediate antibody-dependent cellular cytotoxicity against human tumor cells. Panitumumab is a fully human anti-EGFR monoclonal antibody (IgG2) with high affinity to the EGFR, blocking ligand binding and inhibiting the subsequent activation of downstream signaling pathways.[3, 4] Unlike cetuximab, panitumumab does not mediate antibody-dependent cellular cytotoxicity and less often results in infusion reactions and anaphylactic reactions than cetuximab, because panitumumab is a fully humanized antibody. Multiple in vitro, in vivo, and clinical trials have demonstrated that the efficacy of anti-EGFR monoclonal antibodies depends on lack of v-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog (KRAS) mutation, although recent clinical and laboratory observations may challenge the current knowledge and clinical implications of KRAS mutation and anti-EGFR efficacy.
These anti-EGFR monoclonal antibodies have been evaluated in patients with advanced CRC as first-line, second-line, or later lines of therapy and as either a single agent or in combined chemotherapy. In this review, we discuss the results from recent phase 3 randomized clinical trials with cetuximab and panitumumab, and in particular, the controversial results from combined chemotherapy-cetuximab trials as first-line therapy in patients with metastatic CRC (mCRC). In addition, recent subgroup analyses of individual KRAS mutations and clinical outcomes in major clinical trials, including the Cetuximab Combined With Irinotecan in First-Line Therapy for Metastatic Colorectal Cancer (CRYSTAL), Phase 2 Oxaliplatin and Cetuximab in First-Line Treatment of Metastatic Colorectal Cancer (OPUS), and Panitumumab Randomized Trial in Combination With Chemotherapy for Metastatic Colorectal Cancer to Determine Efficacy (PRIME) trials, have generated conflicting results.[7, 9] In the current review, we discuss conflicting data and discrepancies with preclinical data. Anti-EGFR monoclonal antibody (MoAB)-associated skin toxicity and management strategies also are discussed. Table 1 summarizes the major clinical trials with cetuximab and panitumumab.
Earlier Trials of Anti-EGFR Agents
The first phase 1 clinical trial with cetuximab demonstrated significant antineoplastic activity in epithelial tumors and established the dosing regimen of cetuximab 400 mg/m2 as a loading dose followed by 250 mg/m2 weekly. Subsequent phase 2 trials demonstrated the efficacy of cetuximab in patients with advanced CRC.[25, 26] In a phase 3 randomized trial of 572 patients who had failed all prior therapies, cetuximab monotherapy was compared with best supportive care (BSC). The median overall survival (OS) was significantly better with cetuximab (6.1 months vs 4.6 months; hazard ratio [HR], 0.77; P = .005) in that study, in which no crossover between arms was allowed. A subsequent analysis demonstrated that the benefits of cetuximab were restricted to patients whose tumors carried KRAS wild-type alleles. Among patients who had mutated KRAS, there was no significant difference between cetuximab and BSC in terms of OS (KRAS wild type: 9.5 months vs 4.8 months; HR, 0.55; P < .001; KRAS mutation: 4.6 months vs 4.5; HR, 0.98; P = .89) or progression-free survival (PFS) (KRAS wild type: 3.7 months vs 1.9 month; HR, 0.40; P < .001; KRAS mutation: 1.8 months vs 1.8 months; HR, 0.99; P = .96).
An early phase clinical trial in panitumumab monotherapy was carried out on 148 patients with advanced CRC who were refractory to standard chemotherapy, and the investigators reported a 29% stable disease rate and a 9% partial response rate. Finally, a randomized phase 3 trial in panitumumab monotherapy was compared with BSC in patients who had chemotherapy-refractory mCRC, and a significant response rate (17% vs 0%) and PFS advantages (12.3 weeks vs 7.3 weeks; HR, 0.45; P < .0001) were reported. The lack of an OS difference was probably because of panitumumab use after crossover in the BSC group, of which 76% entered the crossover study. A later post hoc analysis confirmed that efficacy was limited to patients who did not carry a KRAS mutant allele.
After cetuximab monotherapy demonstrated efficacy in CRC, the Bevacizumab and Irinotecan Compared With Cetuximab and Bevacizumab Alone in Irinotecan-Refractory Colorectal Cancer (BOND) trial demonstrated that irinotecan plus cetuximab had significant efficacy compared with cetuximab alone in patients with irinotecan-refractory mCRC, improving the response rate (23% vs 11%) and PFS (4.1 months vs 1.5 months). The Erbitux Plus Irinotecan in Colorectal Cancer (EPIC) trial further investigated oxaliplatin-refractory patients who received second-line therapy with single-agent irinotecan with or without cetuximab. In that study, PFS was significantly better with combined therapy (4 months vs 2.6 months) as was the response rate (16.4% vs 4.2%). No significant OS benefit was observed, but nearly 50% of patients in the irinotecan alone group received cetuximab after disease progression. It is noteworthy that the trial analyzed all patients without KRAS mutation status. Recently, panitumumab with combined 5-fluorouracil (5FU), folinic acid, and irinotecan (FOLFIRI) as second-line therapy was investigated in patients with mCRC who were refractory to combined 5FU, folinic acid, and oxaliplatin (FOLFOX) compared with FOLFIRI alone. Patients who had KRAS mutations also were included in the final analysis. In patients who had KRAS wild-type tumors, combined therapy was associated with significantly longer PFS (5.9 months vs 3.9 months) and a higher response rate (35% vs 10%), and there was a trend toward better median OS (14.5 months vs 12.5 months); whereas patients who had KRAS mutated tumors failed to demonstrate better efficacy with combined therapy.
The trials described above demonstrated that the anti-EGFR MoABs, either alone or with chemotherapy, have efficacy as salvage therapy for heavily treated patients who have mCRC and as second-line therapy with noncross-resistant chemotherapy for first-line chemorefractory patients. PFS and response benefits appeared to be confined to KRAS wild-type tumors. In most of the aforementioned trials except CO.17 (cetuximab salvage therapy compared with BSC), in which crossover did not occur, the benefits in terms of PFS and response rate have yet to yield improved OS for patients with KRAS wild-type tumors. Because of significant crossover and subsequent treatment effects, the nature of these findings raises questions regarding the best endpoint with which to evaluate the efficacy of the agent of interest. For example, in the trial of panitumumab as salvage therapy, nearly half of patients in the control group received panitumumab after tumor progression, skewing OS data toward accepting the null hypothesis. Hence, at this point, we believe that the lack of an OS benefit demonstrated in cetuximab or panitumumab trials among patients with mCRC does not necessarily indicate the lack of clinical utility, and using these agents as tested can be a reasonable choice.
Anti-EGFR Trials as First-Line Therapy
Three published randomized trials investigated use of cetuximab or panitumumab in the setting of first-line treatment with different chemotherapy backbones. The CRYSTAL trial investigated cetuximab in combination with FOLFIRI for patients with mCRC compared with FOLFIRI alone. KRAS mutation status was determined from formalin-fixed, paraffin-embedded tissues retrospectively and was reported in a post hoc analysis. In the KRAS wild-type group, the addition of cetuximab to FOLFIRI significantly prolonged OS (23.5 months vs 20 months; HR, 0.796; P = .0093) and PFS (9.9 months vs 8.4 months; HR, 0.696; P = .0012) and markedly increased the response rate (57.3% vs 39.7%; odds ratio [OR], 2.069; P < .001) and the R0 resection rate (5.1% vs 2%; OR, 2.650; P = .0265). However, patients who had KRAS mutant tumors failed to demonstrate efficacy in any survival or response endpoints. In the United States, oxaliplatin-based chemotherapy has been used more commonly than irinotecan as first-line treatment. The phase 2 OPUS trial addressed the efficacy of cetuximab combined with oxaliplatin-based chemotherapy (FOLFOX4) as first-line therapy. In that trial, patients with KRAS wild-type tumors who received cetuximab plus FOLFOX4 had significantly prolonged PFS (8.3 months vs 7.2 months; HR, 0.567; P = .0064) and exhibited better response rates (57% vs 34%; OR, 2.551; P = .0027) and R0 resection rates (12% vs 3%; P = .0242) with a trend toward better OS (22.8 months vs 18.5 months; P = .39). The phase 3 PRIME trial also demonstrated efficacy for panitumumab in combination with oxaliplatin-based FOLFOX4 as first-line therapy. A PFS benefit was demonstrated in the KRAS wild-type panitumumab arm (9.6 months vs 8 months; HR, 0.80; P = .02); whereas only a trend toward better OS (23.9 months vs 19.7 months; HR, 0.83; P = .072), response rate (55% vs 48%; OR, 1.35; P = .068) and R0 resection rate (8.3% vs 7.0%) was observed in patients with KRAS wild-type tumors. The Cancer and Leukemia Group B (CALGB) 80203 trial of first-line cetuximab plus either FOLFIRI or FOLFOX, which was closed early when data on benefits of bevacizumab became available, reported a significantly higher response rate (55% vs 32% for FOLFOX; 42% vs 34% for FOLFIRI) in the cetuximab plus chemotherapy arm. However, complete data from this trial have not been published. It is noteworthy that the OPUS[18, 29] and PRIME trials demonstrated that, in patients with KRAS mutant tumors, the addition of an anti-EGFR MoAB to chemotherapy produced a detrimental effect, which was further confirmed in the second-line, phase 3 Panitumumab, Irinotecan, and Cyclosporin in Colorectal Cancer Therapy (PICCOLO) trial. These results indicate that anti-EGFR therapy should not be given to patients who have CRC with KRAS mutations in any line of treatment.
Controversies in Recent First-Line Cetuximab Trials
Recently 2 randomized phase 3 trials of cetuximab and oxaliplatin-based chemotherapy have raised more questions regarding efficacy of cetuximab in first line treatment. The Medical Research Council Continuous Chemotherapy plus Cetuximab or Intermittent Chemotherapy with Standard Continuous Palliative Combination Chemotherapy with Oxaliplatin and Fluoropyrimidine in First-Line Treatment of Metastatic Cancer MRC COIN) trials in the United Kingdom included 357 patients with KRAS wild-type tumors in a cetuximab arm, including 117 patients who received infusional 5FU with oxaliplatin and 240 patients who received capecitabine with oxaliplatin. The study investigators reported that OS (17.0 months vs 17.9 months; HR, 1.04; P = .67) and PFS (8.6 months vs 8.6 months; HR, 0.96; P = .60) did not differ between the cetuximab group and the control group, although the response rate increased from 57% to 64% in the cetuximab group (P = .047). The authors argued that the trial has not confirmed a benefit from the addition of cetuximab to oxaliplatin-based chemotherapy in first-line treatment. However, a post hoc analysis has noted improved PFS with cetuximab in patients who received infusional 5FU-based therapy (HR, 0.72; P = .037), but not in those who received capecitabine-based therapy (HR, 1.02; P = .88). A further interaction analysis demonstrated that a PFS benefit was restricted to patients who had KRAS wild-type tumors and no or only 1 metastatic site treated with infusional 5FU-based therapy (HR, 0.55; P = .011).
Another recent phase 3 trial, Nordic VII, investigated the efficacy of cetuximab in addition to bolus 5FU/folinic acid, and oxaliplatin (Nordic FLOX), which has been used exclusively in the Scandinavian countries. The Nordic FLOX regimen was administered every 2 weeks with 5FU 500 mg/m2 as a bolus infusion (<5 minutes), followed 30 minutes later by bolus leucovorin 60 mg/m2 (<10 minutes) on days 1 and 2, and oxaliplatin 85 mg/m2 over 1 hour on day 1; whereas the American FLOX regimen, as reported in the National Surgical Adjuvant Breast and Bowel Project (NSABP) C-07 trial, is commonly administered weekly as 2-hour intravenous leucovorin as a 500 mg/m2 infusion followed 1 hour later by a bolus 5FU 500 mg/m2 infusion and oxaliplatin 85 mg/m2 administered every 2 weeks as a 2-hour infusion before leucovorin and 5FU. The Nordic VII trials included 194 patients with KRAS wild-type tumors, 97 patients who received Nordic FLOX, and 97 patients who received Nordic FLOX with cetuximab. An additional 130 patients who had KRAS mutant tumors also were randomized into the control and interventional arms. In the patients with KRAS wild-type tumors, PFS (7.9 months vs 8.7 months; HR, 1.07; P = .66), OS (20.1 months vs 22 months; HR, 1.14; P = .48), and the response rate (46% vs 47%; OR, 0.96; P = .89) demonstrated an unexpected trend toward a worse outcome at all endpoints measured. Conversely, patients with KRAS mutant tumors had a trend toward better prognosis when they received cetuximab, as indicated by their PFS (9.2 months vs 7.8 months; HR, 0.71; P = .07), OS (21.1 months vs 20.4 months; HR, 0.89; P = .89), and response rate (35% vs 23%; OR, 1.44; P = .31). The trial investigators reached the conclusion that cetuximab did not add significant benefit to the Nordic FLOX regimen in first-line treatment.
Anti-EGFR therapy trials produced inconsistent results in studies that combined cetuximab with first-line oxaliplatin-based regimens, mainly depending on the chemotherapy backbones in first-line treatment, although cetuximab or panitumumab had efficacy in second-line treatment combined with chemotherapy and as single-agent salvage treatment. The significant benefit of cetuximab reported in the CRYSTAL trial contrasted with a lack of benefit from the MRC COIN and NORDIC VII trials. It was speculated that the chemotherapy back bones could explain the difference, because irinotecan may be more effective than oxaliplatin with cetuximab. However, the PRIME and OPUS trials demonstrated a benefit from EGFR-directed antibodies combined with FOLFOX4. Preclinical data also support synergistic effects of cetuximab and oxaliplatin. In vitro, cetuximab combined with oxaliplatin synergistically decreased the 50% inhibitory concentration value of oxaliplatin in subsets of colon cancer cell lines, and the observed responses depended strictly on the cell type. Response was not correlated with the level of EGFR expression but was related to the basal level of phospho-EGFR. The addition of cetuximab to oxaliplatin potentiated cytotoxic effects by stimulating oxaliplatin-DNA adduct formation, which is associated with reduced expression of the key enzyme (excision repair cross complementation group 1 [ERCC1]) in the main repair process of oxaliplatin-DNA platinum adducts. The efficacy of combined cetuximab and oxaliplatin in murine subcutaneous colon cancer xenograft models was greater than that of monotherapy alone and was independent of the responsiveness to oxaliplatin monotherapy.[30, 32] Both studies[31, 32] independently observed that cetuximab reduced the expression of ERCC1 and XPF (xeroderma pigmentosum), which are key components of the nucleotide excision repair pathway involved in the excision of platinum-DNA adducts.
The COIN trial provided another interesting observation that patients who received cetuximab with bolus and infusional 5FU (mFOLFOX6) had a significant benefit in terms of PFS (HR, 0.72; P = .037), whereas patients who received cetuximab with capecitabine-based therapy (XELOX) failed to demonstrate a benefit. The total number of patients who received XELOX (240 patients) far exceeded the total number of patient who received FOLFOX (117 patients) in that trial, which may have contributed to the discrepancies in outcomes and, ultimately, the negative outcomes. This pattern also was observed in the Nordic VII trial, in which all patients received a bolus-only 5FU and leucovorin regimen with oxaliplatin and cetuximab. We learned from those trials, including OPUS, that cetuximab demonstrated better efficacy with an infusional 5FU regimen (FOLFOX) when combined with oxaliplatin. Of note, the mode of 5FU administration was not an issue with irinotecan regimens, because infusional 5FU (FOLFIRI) clearly offered superior activity to bolus 5FU (mIFL) and capecitabine (CapeIRI) with better toxicity profiles in mCRCs; and the CRYSTAL trial demonstrated the significant benefit of cetuximab with FOLFIRI compared with FOLFIRI alone. These observations suggest that cetuximab would work more effectively with infusional 5FU than with bolus 5FU alone or capecitabine, but this has not been formally explored. The data indicating the benefits of oxaliplatin combined with panitumumab included the infusional FOLFOX regimen. In vitro experiments with colon cancer cell lines demonstrated different biochemical responses between modes of 5FU administration. The result suggested that incorporation of fluorouridine triphosphate into RNA appears to be the most important mechanism of action for 5FU bolus schedules, whereas inhibition of thymidylate synthase becomes more important as the infusion time is prolonged. It is noteworthy that multiple studies have demonstrated that anti-EGFR agents, including cetuximab, erlotinib, and lapatinib, appear to down-regulate thymidylate synthase in human colon cancer cells. The inhibition of thymidylate synthase is mediated through down-regulation of EGFR or inhibition of nuclear translocation of EGFR by anti-EGFR agents. Therefore, preclinical data implicated that there may be an unexpected, significant interaction between cetuximab and the mode of 5FU application. These in vitro differences consequently were reflected in the clinical trials using different modes of 5FU administration. However, the development of a better preclinical model is required to elucidate the molecular and biochemical interactions in anti-EGFR agents and current chemotherapy.
In the first-line chemotherapy setting, as discussed above in later line studies, benefits in PFS, response rates, or resection rates with curative intent should not be ignored because of a lack of benefit in OS, because OS can be largely affected by subsequent therapy after the trial. In particular, many patients in control arms received the investigational agent later, and most patients ultimately received comparable numbers of total active agents in subsequent lines of therapy. In the COIN trial, the OS was similar in both arms (17.0 months vs 17.9 months; HR, 1.04), although the authors did not report separate OS for the XELOX and mFOLFOX6 groups. A statistically significant number of patients in the control arm (P = .0061) received more second-line treatment, which may have affected OS. In the Nordic VII trial, a similar number of patients in both arms received active therapeutic agents in the second-line or third-line, such as irinotecan, 5FU, or bevacizumab; whereas patients in the control arm later received significantly more cetuximab (12.4% in the treatment arm vs 27.2% in the control arm), like what occurred in the panitumumab salvage therapy trial. Given this consideration, the significant PFS benefit demonstrated in the mFOLFOX6 plus cetuximab group from the COIN trial needs extra attention. A carefully re-evaluated subgroup analysis may add evidence that FOLFOX with cetuximab can have significant efficacy in first-line therapy.
The combination of cetuximab and chemotherapy increased toxicity and adverse effects, mainly skin reactions, diarrhea, and infusion-related reactions, compared with chemotherapy alone. Of note, however, trials that have demonstrated the efficacy of cetuximab have reported comparable median relative dose intensity and cumulative doses of FOLFOX and FOLFIRI between control and experimental arms.[13, 17, 19] The COIN trial reported a significant reduction in chemotherapy dose intensity in the cetuximab arm. A subgroup analysis indicated that the mFOLFOX6 plus cetuximab group received lower doses of infusional 5FU compared with the control arm (P = .016), and the XELOX plus cetuximab group received significantly reduced doses of both oxaliplatin (P = .0018) and capecitabine (P = .004). This observed reduction of chemotherapy dose intensity in the cetuximab arm may explain in part the lack of efficacy observed in that negative trial. Furthermore, the administered number of cycles of oxaliplatin and cetuximab was significantly less in the XELOX group compared with the mFOLFOX6 group (P = .016), indicating that the combined analysis of XELOX and mFOLFOX6 is more likely confounded by this discrepancy and raising the necessity of a reanalysis of the 2 separate groups.
Controversies in KRAS Mutations and Anti-EGFR Therapies
Recent retrospective, pooled subgroup analyses from prior major trials provided more complexity on anti-EGFR antibody therapy and KRAS mutational status. The KRAS G13D mutation is observed in approximately 15% to 20% of KRAS mutant CRCs, differing with other cancers, which most commonly contain KRAS codon 12 mutations. From a retrospective pooled analysis of multiple trials in the salvage setting, including patients who were treated in an off-trial setting, De Roock et al reported that patients who received cetuximab combined with BSC and/or chemotherapy with the KRAS G13D mutation had significantly better outcomes than those with other KRAS mutations. Tejpar et al, in analyzing the CRYSTRAL and OPUS trial results, demonstrated significant variations in tumor response (P = .005) and PFS (P = .046) in patients who had G13D mutant tumors compared with patients who had all other mutations. Those authors also demonstrated improved clinical outcomes with the addition of cetuximab to chemotherapy (PFS, 7.4 months vs 6.0 months; HR, 0.47; P = .039; tumor response, 40.5% vs 22%; OR, 3.38; P = .042) compared with chemotherapy alone. However, the efficacy of cetuximab in patients who had tumors with KRAS G13D mutations still was inferior to the efficacy in patients who had KRAS wild-type tumors (OS: HR, 1.61; P = .0085; objective response: OR, 0.50; P = .040). It is noteworthy that, within the chemotherapy only arms, the PFS and OS for patients who had tumors with the KRAS G13D mutation was significantly worse than the PFS and OS for patients who had tumors with other KRAS mutations or KRAS wild-type. The median PFS was 6.0 months (95% confidence interval [CI], 5.4-7.8 months) for patients who had the KRAS G13D mutation, 8.8 months (95% CI, 7.2-9.5 months) for those who had the KRAS G12V mutation, and 8.1 months (95% CI, 7.2-9.4 months) for those who had other KRAS mutations. The median OS was 14.7 months (95% CI, 12.4-19.4 months) for patients who had the KRAS G13D mutation, 17.8 months (95% CI, 15.5-21.7 months) for those who had the KRAS G12V mutation, and 17.7 months (95% CI, 15.3-20.5 months) for those who had other KRAS mutations. The worse survival and response rate for patients who had KRAS G13D mutant tumors compared with those who had KRAS G12V mutant tumors in the chemotherapy arm was statistically significant (PFS: HR, 1.92; 95% CI, 1.06-3.47; P = .031; response rate: OR, 0.39; 95% CI, 0.15-0.99; P = .47). However, the cause of the worse outcome for patients with KRAS G13D mutant tumors remains unclear and will require more careful analyses and trials.
Despite the potential association between KRAS G13D tumors and better outcomes from treatments with cetuximab, previous preclinical data do not correspond with clinical observations. KRAS mutations occur in approximately 30% to 40% of CRCs, and the KRAS G13D mutation occurs in <20% of KRAS mutant CRCs. Mutation analysis of various cancer cell lines at the Sanger Institute identified the KRAS G13D mutation in the HCT-116, DLD1, T84, LoVo, NCI-H747, and HCT-15 CRC cell lines. Several studies have consistently reported that HCT-116, DLD1, and LoVo CRC cells containing the KRAS G13D mutation respond poorly to cetuximab in vitro and in vivo.[30, 31, 40, 41] After homologous recombination, isogenic SW-48 CRC cells with KRAS G13D responded to cetuximab and demonstrated EGF-dependent activation of the KRAS-extracellular signal-regulated kinase (ERK) pathway by demonstrating marked activation of ERK in the presence of EGF, which was not observed in KRAS G12V cells. However, this observation is inconsistent with previous biochemical data indicating the failure of cetuximab to inactivate the ERK pathway in CRC cells that carried the KRAS G13D mutatoin.[40, 41] Oncogenic KRAS mutations, including G13D, strongly suppressed the EGF-stimulated activation of ERK phosphorylation and desensitized CRC cells to EGFR inhibition. Misale et al recently reported that cetuximab-resistant variant Lim1215 CRC cells acquired either KRAS G12R or G13D mutations after cetuximab-responsive parental Lim1215 cells had been continuously exposed to cetuximab. Those investigators also demonstrated that isogenic Lim1215 cells with the KRAS G13D mutation indeed displayed resistance to cetuximab, despite using the same method of homologous recombination.
In addition, the finding of cetuximab-associated clinical outcomes in patients with KRAS G13D was not reproducible in the other pooled retrospective analysis with panitumumab and chemotherapy performed by Peeters and colleagues. Those authors reanalyzed subgroups of patients that had different KRAS mutations in the PRIME trial (FOLFOX4 with and without panitumumab) in the first-line setting, FOLFIRI with and without panitumumab in the second-line setting, and BSC with and without panitumumab in the salvage setting. None of the individual KRAS mutations were associated consistently with better clinical outcomes in either the control arm or the panitumumab-chemotherapy arm. Response rates in that trial were comparable between patients with different KRAS mutations. Notably, in the panitumumab-chemotherapy arm of the PRIME trial, the KRAS G13D mutation was associated significantly with worse OS, which was opposite to the observation in a subgroup analysis from the CRYSTAL and OPUS trials. Details of these analyses remain to be published. In parallel with current laboratory and clinical data, more careful observation and re-evaluation of in vitro experiments and the development of new preclinical animal models, including the transgenic/knock-in mouse (KRAS G13D), may provide a new understanding of oncogenic KRAS in the development and treatment of CRC. In addition, as suggested by others,[7, 8] independent, prospective, randomized clinical trials are essential before therapeutic decisions to use EGFR-directed antibodies in patients with the KRAS G13D mutation can be included in clinical practice.
Misale et al and Diaz et al recently reported more interesting findings of the emergence of KRAS mutations as tumors evolve during the course of anti-EGFR treatment. With high-sensitive deep sequencing and BEAMing (beads, emulsion, amplification, and magnetics) technology, both groups were independently able to detect KRAS codon 12 or 13 mutations in the rebiopsied tumors and/or sera from patients whose tumors were initially KRAS wild-type then subsequently developed resistance against cetuximab or panitumumab. Although it has not been clear whether the mutations of conferring resistance are acquired during the course of treatment or pre-exist as de novo mechanism of resistance, the emergence of KRAS mutant clones potentially can be detected in a noninvasive manner months before radiographic progression. This raises a question regarding whether a serial test of KRAS mutation in serum or tissue is appropriate for detecting early resistance and switching to another antineoplastic regimen in patients on an anti-EGFR MoAB. It also raises a question whether a new diagnostic KRAS test should be performed at the time of initiating the anti-EGFR treatment when the patient's disease progressed on a previous line of therapy and the KRAS test was performed before that progression.
Skin Toxicities in Anti-EGFR Therapies
The anti-EGFR MoABs are associated with unique profiles of adverse effects. They include skin rash; electrolyte abnormalities, especially the magnesium-wasting syndrome; and, particularly with cetuximab, infusion reaction caused by the immunogenicity of the chimeric antibody. Among them, skin lesions,[44, 45] such as acneiform eruption, paronychial inflammation, hair abnormalities, and a marked increase in the length of the eyelashes, are prominent, sometimes dose-limiting complications that interfere significantly with patients' quality of life. Notably, none of the lesions were fatal, and the development and especially greater intensity of skin eruptions may be associated with better outcomes of cetuximab and panitumumab. The CRYSTAL and PRIME trials and other trials[9, 12] in the salvage setting also reported an association of skin toxicities with better outcomes.
Management of the dermatologic adverse effects of anti-EGFR MoABs is a particular interest of medical oncologists. The Skin Toxicity Evaluation Protocol with Panitumumab (STEPP) study, a phase 2 open-label randomized trial, evaluated the impact of a pre-emptive prophylactic skin treatment regimen of skin moisturizers, sunscreen, topical steroid, and doxycycline compared with the impact of a conventional treatment strategy. The prophylactic skin treatment regimen decreased the incidence of National Cancer Institute Common Toxicity Criteria (NCI-CTC) grade 2 or greater skin toxicities and improved quality of life assessed with the Dermatology Life Quality Index. However, there was a limitation of objectively assessing adverse skin effects, and the number of patients was small in the phase 2 trial despite the statistical significance. The prophylactic treatment may have negated the possible outcome-associated skin manifestations by masking the adverse skin effects, although the authors demonstrated no significant difference in outcomes between the pre-emptive arm and the standard therapy arm. Additional clinical trials and more experience will be required to establish prophylactic management as the standard of care.
In summary, the current National Comprehensive Cancer Network guidelines advise the use of panitumumab combined with FOLFOX or FOLFIRI but cetuximab combined only with FOLFIRI in the first-line treatment of patients with mCRC. However, the manufacturer of panitumumab recently reiterated that the labeled indication of panitumumab is as monotherapy in the salvage setting because of severe toxicity when it is combined with chemotherapy. It was reported that, in a combination of panitumumab and FOLFIRI, the incidence of NCI-CTC grade 3 diarrhea was 25%, which may lead to acute renal failure and other complications. In addition, the National Comprehensive Cancer Network guidelines strongly advise against the concurrent use of bevacizumab with either cetuximab or panitumumab in the setting of first-line treatment for advanced or metastatic disease based on results of the PACCE trial and the CAIRO2 trial, suggesting potentially adverse outcomes when combining MoABs that are directed against EGFR with bevacizumab. The advisory panel has recently removed the recommendation for the use of cetuximab with FOLFOX as initial therapy for patients with advanced or metastatic disease because of the lack of benefit and the increased incidence of grade 3 adverse events demonstrated in the MRC COIN and Nordic trials. However, a recent meta-analysis, which included the CRYSTAL, OPUS, COIN, and NORDIC VII trials, and the German AIO study (capecitabine), and the CECOG study (infusional 5FU), demonstrated that only patients who received infusional 5FU-based chemotherapy derived a benefit from cetuximab, and the receipt of either oxaliplatin or irinotecan did not affect the benefit from cetuximab. The other meta-analysis of 14 randomized controlled trials comparing the efficacy of cetuximab confirmed that the differences between trial results (heterogeneity P = .02; I2 = 62%) were explained best by the fluoropyrimidine used, and PFS benefits were confined to the trials that combined an anti-EGFR MoAB with infusional 5FU-based chemotherapy (HR, 0.77; 95% CI, 0.70-0.85; P < .00001). Therefore, the aforementioned preclinical data and meta-analysis indicate that the benefit from cetuximab is associated significantly with the use of infusional 5FU, targeting thymidylate synthase, which can be synergistically down-regulated by cetuximab or panitumumab. In contrast, oxaliplatin and irinotecan may not be attributable to the lack of benefit from cetuximab, as indicated in the meta-analyses, although a direct head-to-head comparison between cetuximab with FOLFOX and FOLFIRI remains to be reported. Bevacizumab combined with first-line chemotherapy has improved outcomes in the first-line systemic therapy of mCRC regardless of KRAS mutation status.
Ultimately, the recently completed but not yet reported CALGB 80405 trial will define the role of anti-EGFR antibody therapy with cetuximab and chemotherapy compared with bevacizumab and chemotherapy in the initial treatment of KRAS wild-type CRC. Patients were stratified according to physician-selected chemotherapy: FOLFOX versus FOLFIRI. That study will clarify the clinical efficacy and position of cetuximab as a therapeutic option compared with bevacizumab in the first-line setting, most notably when it is combined with FOLFOX, because of its common use with FOLFOX. Taken together, the results indicate that cetuximab is a reasonable option with chemotherapy as second-line and salvage therapy for patients with advanced mCRC who have KRAS wild-type tumors, whereas panitumumab remains primarily a monotherapy option in the salvage setting because of toxicity concerns. Decisions need to be made regarding delays in the use of EGFR-directed therapy for later lines of treatment. Those patients who have symptomatic disease that necessitates some degree of response, as well as those who are not deterred by the dermatologic toxicities of the EGFR-directed antibodies, may be candidates for their use in earlier lines of therapy.
Conflicting evidence exists regarding first-line anti-EGFR therapy with oxaliplatin-based chemotherapy as well as KRAS G13D mutation-associated clinical outcomes in patients with mCRC. Precise interpretation and re-evaluation of current trials is required. In particular, we must pay greater regard to the preclinical data on mechanisms of interaction between chemotherapy and the anti-EGFR antibodies, mechanisms of resistance to anti-EGFR antibodies, the role of cross-over events in OS data, and the significant dose reductions of chemotherapeutic agents in combination therapy to properly interpret the outcomes of these trials.