Increased Detection Sensitivity for KRAS Mutations Enhances the Prediction of Anti-EGFR Monoclonal Antibody Resistance in Metastatic Colorectal Cancer

Cetuximab and panitumumab are monoclonal antibodies (MoAbs) that inhibit the activation of the epidermal growth factor receptor (EGFR) and its downstream pathways, namely the RAS-RAF-MAPK and the PI3K-PTEN-AKT axes (1). As the response rate to anti-EGFR MoAbs varies from 10% to 20% in patients with metastatic colorectal cancer (mCRC; ref. 2), several studies have been performed to identify markers that can predict the efficacy of these agents.

It is widely accepted that a lack of response to anti-EGFR MoAbs occurs in the presence of oncogenic KRAS mutations (3–11). This finding led the European Medicines Agency and, subsequently, the Food and Drug Administration to limit the use of cetuximab and panitumumab only to patients with wild-type KRAS tumors (12, 13).

In addition, the presence of oncogenic deregulation of EGFR and other members of its downstream signaling pathways, such as BRAF, PIK3CA, and PTEN, might influence the response to cetuximab and panitumumab and could, therefore, help to identify nonresponder (NR) patients (4, 14–19). However, the evaluation of these additional molecular markers does not seem to be sufficient to fully predict the response to EGFR-targeted agents (20, 21). This lack could be explained by inappropriate or nonstandardized methodologies and by the limited sensitivity of current sequencing methods for detecting DNA point mutations (22).

Despite a general consensus favoring the introduction of KRAS testing in clinical practice as a powerful means to select patients before drug administration, validated and standardized techniques for KRAS analysis are still lacking (23–25). Currently, the most widely applied method for assessing KRAS gene status is direct sequencing, which has a relatively low sensitivity because mutant alleles must be present in at least 20% of cells to be reproducibly detected (25).

More sensitive methods are available for KRAS analysis. Some are laboratory-made techniques, such as mutant enriched-PCR (ME-PCR), and others are CE-marked commercial tests for diagnostic use, such as matrix-assisted laser desorption/ionization—time-of-flight (MALDI-TOF) technology or ARMS. Several studies have compared different methodologies of KRAS analysis but without showing a correlation with the clinical response to anti-EGFR agents in mCRC patients (26, 27).

In this study, highly sensitive methods for the detection of KRAS mutations (ME-PCR and MALDI-TOF MS) have been evaluated to identify patients that are unlikely to benefit from anti-EGFR MoAbs. In addition to KRAS mutations in codons 12 and 13, we analyzed the more infrequent mutations occurring at codon 61 by direct sequencing. Finally, to better understand the impact of KRAS tests on predicting the efficacy of EGFR targeted drugs, we also analyzed the mutational status of BRAF and PIK3CA and protein expression of PTEN in the same cohort.

We retrospectively analyzed 111 patients with histologically confirmed mCRC collected from 1996 to 2009. All tumors were colorectal adenocarcinomas, diagnosed at the Institute of Pathology of Locarno (Switzerland; n = 52), the Civic Hospital of Legnano (Italy; n = 22), the University Foundation of Chieti (Italy; n = 30), and at the University School of Medicine of Novara (Italy; n = 7). Patients gave informed consent and were treated with cetuximab- or panitumumab-based regimens at the referred institutions. All patients had EGFR expression in at least 1% of neoplastic cells, as detected by immunohistochemical studies.

In chemotherapy-refractory patients, cetuximab was administered as a single agent or in combination with chemotherapy-based regimens (irinotecan or oxaliplatin in 95 and 7 patients, respectively) given at the same dose and schedule used at progression. Therefore, the patients included in this study were selected based primarily on evidence that the treatment outcome could only be attributed to the administration of either cetuximab or panitumumab.

With the exception of 3 patients who received cetuximab as frontline monotherapy, the others had failed at least 1 prior chemotherapy regimen.

Cetuximab, as monotherapy or in combination with another chemotherapeutic regimen, was administered at a loading dose of 400 mg/m² over 2 hours, followed weekly by 250 mg/m² over 1 hour. Panitumumab (6 mg/kg) was given i.v. every 2 weeks until progression.

Treatment was continued until progressive disease (PD) or toxicity occurred, according to the standard criteria (28). The clinical response was assessed every 6 to 8 weeks by radiologic examination. The Response Evaluation Criteria in Solid Tumors (RECIST; 28) were adopted for clinic evaluation, and the objective tumor response was classified as partial response (PR), stable disease (SD), or PD. Patients with SD or PD were defined as NR (28). Two independent oncologists and radiologists verified the clinical response for all patients in a blinded manner.

Formalin-fixed paraffin-embedded tumor blocks were reviewed for quality and tumor content. A single representative tumor block for each case, containing at least 70% neoplastic cells, was selected for immunohistochemical, cytogenetic, and molecular analyses. Tumor macrodissection was performed in tumor blocks containing less than 70% of neoplastic cells (to reduce the presence of nonneoplastic tissues).

Genomic DNA was extracted using the QIAamp Mini kit (Qiagen) according to the manufacturer's instructions.

Mutational analyses by direct sequencing were performed at the Institute of Pathology in Locarno (Switzerland).

Mutations in KRAS (exons 2–3, containing hotspot codons 12, 13, and 61), BRAF (exon 15, containing codon 600), and PIK3CA (exons 9 and 20, containing codons 542, 545, and 1047) were detected by direct sequencing (sensitivity of about 20%) of genomic DNA as previously described (17,19,29). The list of primers used for the mutational analyses is reported in Table 1. Direct sequencing has a sensitivity of approximately 20%.

PTEN protein expression status was analyzed using immunohistochemistry (IHC) on 3 μm tissue sections as reported previously (4).

Mutations at codons 12 and 13 of the KRAS gene were investigated using MALDI-TOF MS (specific only for G12V, G12D, G12A, G12C, G12R, G12S, G12F, G13D, G13V; sensitivity of about 10%), ME-PCR and eME-PCR (both with a sensitivity of 0.1%).

The analysis of KRAS using MALDI-TOF was performed by staff from the Institute for Cancer Research and Treatment (IRCC), Candiolo, Italy, at Sequenom GmbH, Hamburg, Germany. This advanced technology for high-throughput mutational analysis of tumor samples is based on a combination of PCR and MALDI-TOF MS, using the MassARRAY System (Sequenom GmbH). KRAS exon 2 was amplified by PCR using OncoCarta PCR primers (Sequenom). The thermocycling conditions were as follows: 94°C for 2 minutes; 45 cycles of 94°C for 30 seconds, 56°C for 30 seconds, and 72°C for 60 seconds; and a final extension at 72°C for 5 minutes. The primary PCR reaction mix was treated with shrimp alkaline phosphatase to deactivate unincorporated dNTPs, and a single-base primer extension step was performed. This method is based on the annealing of a primer adjacent to the mutation site and extension with 1 mass-modified dNTP. The reaction was cycled at 94°C for 30 seconds; followed by 40 cycles of 94°C for 5 seconds, 52°C for 5 seconds, and 80°C for 5 seconds; and a final extension at 72°C for 3 minutes. CLEAN resin (Sequenom) was added to the mixture to remove extraneous salts that could interfere with the MALDI-TOF analysis. Finally, the extended primers were dispensed onto a 384-well SpectroCHIP II array (Sequenom), using a MassARRAY Nanodispenser (Samsung), and analyzed by a MALDI-TOF mass spectrometer. The data were processed using a MassARRAY Typer 4.0 Analyzer (Sequenom). MALDI-TOF MS has a sensitivity of approximately 10%.

KRAS analyses by ME-PCR were performed at the Institute of Pathology in Locarno (Switzerland). ME-PCR consists of 2 amplification steps (semi-nested PCR), in which artificial restriction sites are introduced into the wild-type amplicon using mismatched primers (30). The restriction site, BstNI for codon 12 or BglI for codon 13, introduced during the first PCR step, is positioned immediately next to the KRAS codon being analyzed to distinguish between the wild-type and mutant sequences. Wild-type amplicons were then digested by BstNI or BglI, whereas mutant products were enriched for a second round of amplifications. The ME-PCR and digestion conditions have been reported previously (30). The list of primers used for the mutational analyses is shown in Table 1. All samples were subjected to automated sequencing by an ABI PRISM 3130 (Applied Biosystems) using reverse primers. All mutated cases were confirmed twice with independent PCR reactions. ME-PCR has a sensitivity of up to 0.1% (30).

KRAS analyses by eME-PCR were performed at the Clinical Research Center in Chieti (Italy). An eME-PCR technique was recently developed to further increase the detection sensitivity of KRAS mutations at codons 12 and 13. The method has previously been described in detail and used for the detection of KRAS mutations in lung adenocarcinomas (31). This technique has been engineered to obtain the highest sensitivity and specificity through the optimal selection of reagents and several modifications to the original protocol. eME-PCR has a sensitivity of approximately 0.1%.

The objective tumor response was used as the endpoint of our study. A 2-tailed Fisher's exact test was used to calculate the P values for the association between gene alterations and the clinical response to anti-EGFR MoAbs. The level of significance was set at P = 0.05. The progression-free survival (PFS) and overall survival (OS) analysis were performed according to the Kaplan–Meier method, and survival curves were compared using the log-rank test. OS was defined as the time from the start of treatment with cetuximab or panitumumab until the last follow-up, whereas PFS was calculated from the start of treatment with cetuximab or panitumumab until the first documented tumor progression or death. The data were analyzed using the SAS System V9.1 (SAS Institute Inc.).

This study analyzed 111 patients, including 27 patients already considered in our previous work (4). The patients' characteristics, treatment regimens, and response by treatment with anti-EGFR MoAbs are summarized in Table 2. After cetuximab- or panitumumab-based therapy, 21 patients (19%) achieved PR.

Direct sequencing identified KRAS exon 2 mutations in primary tumors in 43 cases (39%). These mutations occurred only in codon 12 in 31 cases (74%) and only in codon 13 in 11 cases (26%); a double-point mutation involving both codons was detected in 1 case (2%). For the analysis of the clinical response, the case with the double mutation was grouped with the codon 12 mutations. Exon 3 of KRAS could not be analyzed in 3 cases because of a lack of material. In exon 3, a KRAS point mutation was detected in 4/108 evaluable cases (4%). Three cases showed a mutation involving the third nucleotide of codon 61, leading to a Q61H change in 2 cases and a Q61L change in 1 case; the other patient showed a mutation in codon 60 (G60D). BRAF mutations were found in 9/111 patients (8%), all represented by V600E substitutions. All KRAS mutated cases were BRAF wild type and vice-versa, confirming data from the literature (32).

KRAS analyses by ME-PCR and eME-PCR (hereafter collectively referred to as ME-PCRs) were performed on the entire cohort of mCRC patients. However, it was only possible to analyze KRAS gene status by MALDI-TOF MS in 53 patients because of a lack of histological tissue samples from the remaining patients (Table 3). All 3 methodologies showed an increase in sensitivity compared with direct sequencing due to either the application of mass spectrometry (MALDI-TOF MS) or the disruption of wild-type alleles through 2 sequential enzymatic reactions and subsequent re-amplification of mutated alleles (ME-PCRs). ME-PCRs, but not MALDI-TOF MS, are more sensitive than other commercial kits based on ARMS, such as the DxS KRAS mutational test, for which the sensitivity is approximately 1%. Moreover, the 3 methods all showed a reproducibility of 100%, as all the mutations found by direct sequencing of codon 12 or 13 of the KRAS gene were confirmed by MALDI-TOF MS (in the 53 evaluable patients) and by ME-PCRs (in the entire cohort), and the mutations found by all 3 methodologies were confirmed in 2 independent experiments.

Focusing our attention on the subgroup of patients who had been found to be wild type in exon 2 of KRAS by direct sequencing, we detected additional KRAS mutations in 4/36 (11%) analyzable cases (2 G12A and 2 G12V mutations) using the MALDI-TOF MS method. In the entire cohort of 68 patients who showed no mutations in codon 12 or 13 of the KRAS gene by direct sequencing, the ME-PCR found the same mutations detected by MALDI-TOF MS and 5 additional mutated cases (3 with G12D, 1 with G13C and 1 with G13S changes). Finally, eME-PCR detected the same mutated cases as ME-PCR, aside from 1 case with a G12D mutation, and 4 additional KRAS codon 12 mutations (2 G12S, 1 G12D, and 1 G12V changes) (Table 3, Fig. 1). Overall, in the group for which it was possible to analyze K-RAS status by MALDI-TOF, ME-PCR, and eME-PCR, the MALDI-TOF MS technique failed to identify 5/9 mutations detected only by ME-PCRs. All KRAS mutations detected by the 3 highly sensitive methodologies were found in BRAF wild-type patients, further confirming the mutual exclusivity of mutations in these 2 genes in CRC. In 1 patient (#35, Table 3), a KRAS exon 3 mutation was detected in addition to a KRAS exon 2 G12D mutation; the latter was found only by ME-PCR.

Overall, through the application of highly sensitive KRAS analysis methods, we detected additional KRAS alterations in up to 13/68 patients (19%) after direct sequencing had shown them to be wild type in codons 12 and 13 of KRAS.

Metastatic lesions (lung and brain in 1 case, and distant extraintestinal metastatic lymph nodes in another) were available in 2 of the 13 patients whose KRAS mutations were only detectable with the highly sensitive technologies (patients #1 and #3, respectively, Table 4). The analysis of the metastatic lesions by direct sequencing of KRAS was able to reveal the same mutation that was only found in the primary tumors by ME-PCR (Fig. 2).

Because of a lack of histological tissue samples, cases 2 and 21 were not evaluated for PIK3CA and PTEN protein expression, respectively. We detected PIK3CA mutations in 11/109 patients (10%) by direct sequencing and loss of PTEN expression in 26 (of 90 evaluable) patients (29%) by IHC. PIK3CA mutations were found in exon 9 in 8 cases (73%) with the classical E545K mutation in 6 cases, E545G in 1 case, and E542K in 1 case. Mutations in exon 20 were found in 3 cases (27%), and all were H1047R changes.

The relationship between the KRAS, BRAF, and PIK3CA mutations and PTEN expression with clinical outcome was evaluated in terms of the objective tumor response (Table 5).

With the exception of 2 patients with KRAS G13D mutations and 1 patient with the rare KRAS G60D mutation in exon 3, the patients with KRAS, BRAF, or PIK3CA mutations, as detected by direct sequencing, did not respond to cetuximab- or panitumumab-based therapy. All but 2 patients showing a loss of PTEN expression by IHC did not respond to anti-EGFR MoAbs. In the entire cohort, only KRAS exon 2 status was strongly linked to the response (P = 0.002, 2-tailed Fisher's exact test, Table 5). In the group of patients wild type for KRAS exon 2, the KRAS mutations found by ME-PCR were significantly linked to nonresponse (P = 0.01, 2-tailed Fisher's exact test), whereas BRAF mutations were borderline linked to anti-EGFR MoAb resistance (P = 0.053, 2-tailed Fisher's exact test). PIK3CA and PTEN alterations were not associated with response (P = 0.3 and 0.1 respectively, 2-tailed Fisher's exact test; Table 5). In the group of KRAS wild-type patients, as determined by ME-PCR, BRAF mutations were associated with nonresponse (P = 0.02, 2-tailed Fisher's exact test), whereas mutations in the other molecular markers were not. Overall, in the subgroup of patients for which all the molecular markers were evaluated (90 cases), alterations in at least 1 molecular marker were found in 61/90 (68%) cases, 57/74 (77%) of those in NR patients.

All cases with KRAS mutations detected by MALDI-TOF MS and ME-PCR occurred in NR patients, therefore increasing the rate of identified NR patients from 45% (based on the detection of KRAS mutations by direct sequencing alone) to 60% (by adding the analysis of KRAS exon 2 by ME-PCR; Fig. 3). Of the 13 patients showing a KRAS mutation by ME-PCR, 5 showed a concomitant alteration in PIK3CA, KRAS codon 61, or PTEN protein expression; 1 was not evaluable for PTEN expression, whereas the 7 remaining cases did not carry any alterations in the pathways downstream of EGFR. Notably, KRAS codon 61 mutations were detected in 3 NR patients and occurred concomitantly with H1047R PIK3CA mutations, PTEN loss, and mutations in KRAS codon 12 that were detected by ME-PCR. Overall, the analysis of the molecular markers by direct sequencing or immunohistochemistry and of KRAS mutations by ME-PCR enabled us to identify 87% of the NR patients (Fig. 3).

Kaplan–Meier curves indicated that there were no statistically significant differences in PFS and OS between patients with a KRAS wild-type tumor and patients carrying a KRAS mutated tumor detected either by direct sequencing or ME-PCR (data not shown).

KRAS is the major negative regulator of EGFR-targeted therapies in mCRC patients (3–11). However, no standardized procedures for KRAS mutational testing have been proposed and established (23, 24). Indeed, most of the works dealing with the role of KRAS in predicting the response to anti-EGFR MoAbs have been performed by means of direct sequencing or allele-specific PCR (25, 33), which have a sensitivity of approximately 10% to 20% for detecting single amino acid substitutions, whereas few studies have used ARMS (1% sensitivity) for KRAS analysis in a clinical setting (8, 34). KRAS mutational analysis techniques, such as ME-PCR or PCR-RFLP, characterized by increased sensitivity compared with direct sequencing and ARMS, have only been used for detecting micrometastasis in CRC patients (35), either in stool or in serum as a prototype screening (36–38), but have never been used for the assessment of predictive markers in primary tumors. Our study focused on the possibility of improving the detection of KRAS mutations in mCRC patients treated with cetuximab or panitumumab using 3 highly sensitive techniques, MALDI-TOF MS, characterized by a sensitivity of 10% (39), and 2 different ME-PCRs techniques, characterized by a sensitivity threshold of 0.1% (40). Increased efficacy of detecting mutated minor clones would lead to more accurate identification of NR patients.

Several steps were undertaken to ensure that optimal procedures were used for mutational testing through direct sequencing. These steps included strict adherence to current recommendations for tissue handling (24) and the involvement of experienced pathologists in selecting representative tissue samples and performing tumor macrodissection. Furthermore, mutational analyses were performed using widely accepted protocols (4, 17), and the laboratory is registered by external quality control audits.

In our cohort, direct sequencing revealed KRAS mutations in 43/111 cases (39%), a result in line with the published data on this topic (4, 17, 20). Using more highly sensitive techniques, we found the same alterations detected by direct sequencing, and additional KRAS mutations in up to 13 cases were identified, depending on the methodology used, bringing the number of KRAS mutations identified to 56/111 cases (50%). The fact that it was possible to perform MALDI-TOF MS analysis only in 50% of cases renders it somewhat difficult to predict the relative contribution of the ME-PCR techniques and MALDI-TOF MS in the entire cohort. However, if we focus on the 53 patients for whom all the analyses were available, MALDI-TOF MS missed 5/9 mutations that were found only by ME-PCRs.

Focusing on works investigating the KRAS mutational status of mCRC patients treated with EGFR-targeted therapies, the mutational rate reported in the literature is 32% to 42% when direct sequencing was used and 40% to 45% when ARMS was applied. Therefore, the use of ARMS slightly increases the number of KRAS mutations detected. In our cohort, when considering the cases in which it was possible to use direct sequencing and the 3 highly sensitive methods, the KRAS mutational status was 32% with direct sequencing (17/53), 40% with MALDI-TOF MS (21/53), and 50% with ME-PCRs (26/53). Therefore, we can postulate that the use of MALDI-TOF MS may lead to the identification of a number of KRAS mutations similar to that obtained by ARMS whereas the application of ME-PCRs significantly increases the detected KRAS mutational rate.

In addition, when considering BRAF, a marker recently linked to MoAb resistance (17, 41), the MAPK pathway was altered in 65/111 patients (58%), thus showing that this axis plays a pivotal role in CRC development. The 3 highly sensitive techniques confirmed all of the mutations in KRAS exon 2 found by direct sequencing, indicating that the increase in the sensitivity of our methodologies did not compromise the accuracy of detecting specific KRAS mutations.

The data generated by the use of more sensitive techniques are reproducible because all of the mutations found were confirmed at least twice with 2 independent experiments. Moreover, ME-PCR techniques were also used to analyze KRAS mutations in plasma samples from CRC patients, and the mutations found corresponded in each case to the same mutation detected in the primary tumor. No mutations were found in patients with KRAS wild-type tumors (42). There are some limitations of the high-sensitive techniques we used. ME-PCRs are time-consuming (3 working days) and require sequencing confirmation and considerable manual input to avoid contamination. MALDI-TOF MS is a fast, high-throughput methodology, characterized by lower sensitivity than ME-PCRs and elevated costs. Moreover, it requires 180 ng DNA for the KRAS analysis, whereas for ME-PCR, 100 ng is sufficient.

With regard to the response to the anti-EGFR MoAbs, it is important to stress that although the patients included in our cohort are characterized by different chemotherapy background, they represent a homogeneous group in terms of evaluating the cetuximab or panitumumab response, as they experienced resistance to previous chemotherapy regimens.

By correlating the molecular and clinical data, we found that almost all of the KRAS exon 2 mutations detected by direct sequencing and all of the KRAS exon 2 mutations detected by the highly sensitive techniques occurred in NR patients. Therefore, the use of ME-PCRs increased the rate of identification of NR patients from 45% (detected by direct sequencing only) to 62%. As all BRAF mutations were found in NR patients, confirming previous studies (3, 17, 41), by adding the BRAF test to the KRAS analyses, we identified 63/90 (70%) patients refractory to MoAbs. Our data confirm the results of studies comparing different KRAS mutation analysis methods, showing a greater proportion of KRAS mutated cases when more highly sensitive methods are used instead of direct sequencing (26, 27). The novelty of this study is that we showed the clinical impact of using a highly sensitive KRAS analysis method for patient selection.

Kaplan–Meier analysis showed that patients with KRAS mutations, detected by either method, showed similar survival curves in terms of both PFS and OS, with P-values not reaching statistical significance. Therefore, our data do not support the prognostic role of KRAS in the follow-up of mCRC patients treated with anti-EGFR therapies. However, there are conflicting data in the literature on this topic, so our results are not unexpected (9, 43).

Interestingly, in our cohort, 2 patients carrying both a KRAS mutation in codon 13 (the classical G13D change), as identified by direct sequencing, benefited from the treatment. This finding is in line with the literature; a recent review (21) reported that a very small number of patients carrying KRAS-mutated tumors responded to either cetuximab or panitumumab. In those patients, codon 13 mutations were predominantly present. A recent study investigating this issue confirmed our results, thus indicating that, in rare instances, KRAS codon 13 mutations may not impair the response to cetuximab and panitumumab (44).

The identification of a greater number of KRAS-mutated cases by ME-PCR in primary tumor specimens may be explained by the heterogeneity of the tumors (45–49) and suggests that clones bearing KRAS mutations might be undetectable when direct sequencing is used (31). Cells from these clones may display an increased capability to disseminate into peripheral organs where they could predominate (distant metastasis). Our results seem to support this hypothesis; the 2 cases in which KRAS mutations were identified only by means of ME-PCR in the primary tumor and for which a metastatic lesion was available showed the same KRAS mutation in the metastatic specimens simply by using direct sequencing. The clinical relevance of highly sensitive methodologies is strengthened by the evidence that mutations in KRAS gene found only by ME-PCR were never concomitant with BRAF mutations, thus confirming the mutual exclusivity between alterations in these 2 genes, even when high-sensitive methodologies of KRAS analysis were used.

Recent data have suggested that alterations in the PI3K pathway (e.g., PIK3CA gene mutations and the loss of PTEN protein expression) may represent additional negative genetic regulators of EGFR-targeted therapies (4, 17, 19). Therefore, to analyze the clinical significance of KRAS mutations that could be detected only by ME-PCRs, we also investigated PIK3CA mutations and PTEN protein expression in the same patients. All but 2 patients with PTEN loss, characterized by alterations in the PI3K axis, were NR, thus confirming previous reports (18, 19, 41). In particular, of the 13 cases with a KRAS mutation that could be detected only by ME-PCR, 7 did not carry any mutations in the PI3K pathway; therefore, in these cases, the nonresponsiveness is attributable to the presence of the KRAS mutation detected by ME-PCR. These results confirm that the use of ME-PCR methods may represent a useful tool for improving the identification of patients unlikely to benefit from MoAb administration. Overall, by examining the PI3K pathway in addition to the KRAS and BRAF analyses, we were able to identify up to 87% of NR patients.

Because recent reports suggest that rare mutations in KRAS exon 3 could be associated with a lack of response to treatment with cetuximab plus irinotecan (41, 50), we extended our analysis of this gene. In the present cohort of patients, we identified 3 mutations in codon 61, all occurring in NR patients (who also showed a PIK3CA point mutation or a loss of PTEN expression), and 1 mutation in codon 60, in a patient who responded to therapies. Although they were obtained in a relatively limited number of cases, our results seem to indicate that the examination of KRAS exon 3 does not improve the identification of NR patients. These results also suggest that such mutations play a minor role in EGFR-targeted drug efficacy.

Overall, our results suggest that patients might also benefit from the development of sensitive methodologies for the analysis of other predictive molecular markers involved in pathways downstream of EGFR, for example, a ME-PCR technique is also available for BRAF (51). However, this assay is not feasible for all mutational analyses because it involves the generation of a restriction site specific for the wild-type allele, which depends on the surrounding sequence. Other more sensitive methods (e.g., allele-specific PCR) may still be developed.

In conclusion, our data point to the usefulness of increasing the sensitivity of methods to detect mutations in KRAS for enhancing predictions of resistance to cetuximab or panitumumab in mCRC. There is, therefore, an urgent need for the establishment of widely accepted guidelines for KRAS testing, focused on defining the sensitivity threshold that is required the accurate identification of NR patients. We must also emphasize that, as our work is a preliminary and retrospective study, our data will need to be confirmed in larger prospective studies.

No potential conflicts of interest were disclosed.

Oncosuisse OCS-01921-08-2006 (M. Frattini), OCS-02301-08-2008 (M. Frattini), Fondazione Ticinese per la Ricerca sul Cancro (Tessin Foundation for Cancer Research; M. Frattini), the Italian Association for Cancer Research (AIRC; A. Bardelli, A. Marchetti), the Italian Ministry of Health (A. Bardelli, A. Marchetti), Regione Piemonte (A. Bardelli and F.D. Nicolantonio), the Italian Ministry of University and Research, Fondazione Monte dei Paschi di Siena, Siena, Italy (A. Bardelli), and the Association for International Cancer Research (AICR-UK; A. Bardelli), EU FP7 Marie Curie CAN-GENE (A. Bardelli), AIRC 2010 Special Program Molecular Clinical Oncology 5xMille, Project n° 9970 (A. Bardelli).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.