Identification of EGFR mutations in lung sarcomatoid carcinoma

Abstract

Around the time when the short report “EGFR and KRAS status of primary sarcomatoid carcinomas of the lung: Implications for anti-EGFR treatment of a rare lung malignancy” was published in the June 2 2009 issue of International Journal of Cancer,1 we had found, during our molecular diagnostic routine on lung neoplasms, a case of primary sarcomatoid lung carcinoma that harbored an exon 19 deletion of the EGF receptor. Consequently, we read with much interest the data reported by Antoine Italiano et al. describing the molecular analysis of a small series of lung sarcomatoid carcinomas and their general findings consisting of identification of no EGFR mutations, low frequency of chromosome 7 high polysomy, and a high rate of KRAS codon 12 point mutations. To gain more insight and extend our isolated observation, we reviewed our archives of formalin-fixed paraffin-embedded tumors and selected a retrospective cohort of 22 additional consecutive primary lung sarcomatoid carcinomas, surgically treated at our institution between years 2005 and 2009, which were studied for EGFR mutations, EGFR copy number analysis, and KRAS codon 12 and 13 alterations. The characteristics of the series are detailed in Table 1.Diagnosis was independently performed by 2 expert pathologists (R.P. and P.G.) following the WHO classification criteria.2 To minimize normal cell genomic DNA contamination, tumor areas to be analyzed were selected by manual microdissection of H&E stained slides. EGFR analysis of exons 19 and 21 was performed by direct sequencing using the protocol published by Lynch et al.,3 KRAS by exon 1 direct sequencing according to Karapetis et al.4 and EGFR FISH analysis using the LSI EGFR SpectrumOrange/CEP7 SpectrumGreen probe (Vysis, Downers Grove, IL) following the manufacturer's protocol. As in the study of Italiano et al.,5 we considered FISH+ all cases showing ≥4 distinct signals in >40% of 100 screened nuclei or gene amplification defined by the presence of small gene clusters and a gene/chromosome per cell ratio ≥2 or ≥15 copies of the gene per cell in ≥10% analyzed cells. Our molecular analysis allowed the identification of one additional patient whose tumor harbored an EGFR mutation. Collectively, out of the 23 cases analyzed in all, we report of two patients with an EGFR mutation: case no. 10 exhibited an exon 19 in frame deletion of 15 nucleotides spanning codons 746 to 750 (del“ELREA”) (Fig. 1) whereas case no. 17 carried an EGFR exon 19 deletion of 18 nucleotides spanning codons 747 to 752 (del“LREATS”). Three cases demonstrated a KRAS codon 12 mutation, 8 cases had a chromosome 7 polysomy, whereas 2 cases, no. 1 and 19, showed a specific EGFR amplification (Fig. 1). For 6 cases, we could be able to easily select by manual microdissection, and analyze independently, both the adenocarcinoma and the sarcomatoid carcinoma components (cases no. 2, 4, 6, 10, 11, 17). Two out of the 6 cases examined were positive for KRAS mutation, in one of them (case no. 4), the codon 12 GGT to TGT mutation was confined to the sarcomatoid component, whereas the other (case no. 6) had the same codon 12 GGT to GAT mutation in both neoplastic histotypes. Cases no. 10 and 17 showed the same EGFR mutation in both components. No differences of EGFR copy number aberrations were detected in the different histological components. EGFR alterations in primary lung sarcomatoid carcinoma. Histopathologic features (×20 H&E) of a primary sarcomatoid carcinoma of the lung (a), case no. 1, showing EGFR amplification by dual-color FISH assay, using EGFR (red) and chromosome 7 centromere (green) probes (c). Morphologic features (×20 H&E) of sarcomatoid carcinoma of the lung (b) mixed with minor component of adenocarcinoma (arrow) and corresponding direct sequencing of EGFR exon 19 showing del 746–750 in sense and antisense orientation (d). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.] Lung sarcomatoid carcinoma of the lung is a deadly disease for which therapeutic treatment is ineffective, and therefore, molecular studies aimed to characterize their biologic features are urgently needed.6 Much interest has raised the discovery of the EGFR and KRAS roles in lung neoplasms and a consistent body of evidence exists concerning their association with tumor response to EGFR-targeted therapies.7, 8 Reasoning in terms of EGFR status, in their work, Italiano et al. reported a 100% rate of EGFR over expression in immunohistochemistry, 23% of high chromosome 7 polysomy, no EGFR mutations, and 38% of KRAS codon 12 mutations. They concluded that a high EGFR expression and high rate of KRAS mutation may contribute to the aggressive outcome of the disease, and furthermore, because of the lack of EGFR mutations and amplifications, most patients with lung sarcomatoid carcinoma are not likely to benefit from EGFR-targeted therapies. Our data, with the limitations of a small cohort, point toward different conclusions, and therefore, we feel entitled to raise some critical issues as follows: (1) KRAS rate of alterations is high in sarcomatoid tumors: in their series, Italiano et al. observe a 38% incidence of KRAS mutations of which 6 identified by canonical direct sequencing and 2 by mutant enriched PCR. If one takes in consideration, only the 6 cases identified by direct sequencing, their KRAS mutation rate would be of 28.5% and therefore not too different from other reports.9 Conclusions about the 2 cases identified by ME-PCR cannot be done because, as stated by the authors, attempts to make quantitative assessments were not made and, furthermore, the biological and clinical role of KRAS mutations found in under-represented clones, although intriguing, still needs to be clarified. In our series, we found only 3 of 23 cases with a KRAS codon 12 mutation (13%). We register that our rate is lower than expected for other lung neoplasms, but we also have to point out that, to the best of our knowledge, in the only 2 other series where the KRAS status in sarcomatoid carcinomas was studied (Pelosi et al.) and (Przygodzki et al.), the detected mutation rate was, respectively, of 22% and 9% and therefore lower than in other lung neoplasms.10, 11 Our low mutation incidence rate cannot be due to normal cell DNA contamination because manual microdissection of tumor cells is routinely performed in our laboratory and direct sequencing was performed at least 2 times on different PCR reactions. Consequently, we have to conclude that the differences in mutation incidence must be due to statistical fluctuation that is expected when small numbers are studied. (2) EGFR mutations do not occur in sarcomatoid lung tumors: some isolated evidence of the occurrences of EGFR mutations in sarcomatoid lung tumors do exist, although in oriental patients for which a much higher incidence of EGFR aberrations has been reported: Atsuhito Ushiki et al.12 recently described a case of lung sarcomatoid carcinoma treated with gefitinib that showed a poor response, in which postmortem examination demonstrated a concomitant exon 19 deletion and exon 20 T790M (resistance-associated) mutation. Takano et al., in a mixed population of NSCLC patients treated with gefitinib, reported in Table 4 of their paper one patient with pleomorphic carcinoma with an EGFR mutation, but neither mentioned which was the mutation type nor whether that particular case had responded to drug treatment. Interestingly, though, their EGFR mutation analysis was performed on Laser captured microdissected (LCM) tumors, and they reported that 26% of the 43 mutations detected in LCM samples were not identified in bulk tumor samples.13 Therefore, tumor areas appropriate selection seems to be critical for correct evaluation of EGFR mutations. In another report on an additional series of 469 patients studied for EGFR mutations, 2 patients with lung sarcomatoid carcinoma were included, of which one exhibited an EGFR exon 21 L858R mutation. In this series, a method to enrich for exon 21 point mutation detection was applied.14 In our small cohort, we found 2 patients whose tumors harbored an EGFR exon 19 deletion. Therefore, although at an unknown rate, EGFR mutations in sarcomatoid carcinomas can occur. (3) Italiano et al. found 5 cases (23%) with high polysomy but no true EGFR amplifications. In our series, we observed a higher rate of EGFR-increased copy number consisting of 8 polysomies (34.7%) and 2 cases with specific amplification. One specimen with high polysomy had a KRAS codon 12 mutation (case no. 4) showing that KRAS mutations and EGFR increased copy number, in contrast to EGFR mutations, are not necessarily mutually exclusive. Therefore, we are in agreement with Italiano et al. who demonstrated EGFR gene copy number aberrations in sarcomatoid lung carcinomas, and, furthermore, we also provide evidence that EGFR gene mutations, although at a yet unknown rate, can occur. Both genetic events have been shown to be correlated, in other lung tumor types, with response to EGFR-targeted therapies.7, 8 In our opinion, discrepancies are mainly due to the small number of cases studied in the 2 different series. Consequently, to draw final conclusions, further studies of much larger cohorts analyzed by specific methods for tumor cell selection are needed. Waiting for future confirmation of our findings, we believe that the possibility of the existence of EGFR aberrations should not be overlooked in a subset of lung sarcomatoid carcinoma patients that should not be excluded “a priori” from the potential benefit of EGFR-targeted therapies. Yours sincerely, Alvaro Leone, Paolo Graziano, Rita Gasbarra, Gregorino Paone, Giuseppe Cardillo, Andrea Mancuso, Filippo De Marinis, Serena Ricciardi, Luigi Portalone, Roberto Pisa.