• © 2005 by American Society of Clinical Oncology

In Reply:

Weber et al state that there are limitations in the use of single-strand conformation polymorphism analysis (SSCP) for the mutational screening of EGFR. The authors should consider that in our article we used a dual technical approach (direct sequencing and SSCP assay) for the analysis of all of the 860 non–small-cell lung carcinomas investigated.¹ We unambiguously demonstrated that not only could SSCP analysis detect all of the EGFR mutations observed by direct sequencing, but also an additional 20% of point mutations that were later confirmed by sequencing of the shifted bands.

The SSCP analysis is a very sensitive technique for the detection of gene mutations.^2,3 In addition, this method is highly reliable, as we have shown in a multicenter quality control study for detection of p53 mutations.⁴ However, the setting of critical parameters is required; namely, the amplicon length, the temperature during electrophoresis, and the composition of the gel.^2,5 The length of the amplicon is particularly important; for maximum efficiency it should fall within the range of 150 to 250 bp. Fortunately, all of the exons codifying the EGFR tyrosine kinase domain, contrary to the statement of Weber et al, are very short, under 186 bp (GenBank NT033968), in a ideal range for SSCP analysis. The length of the amplicons used in our study was 227 bp for exon 18, 192 bp for exon 19, and 212 bp for exon 21. The other two parameters (running temperature and gel composition) were optimized for the screening of EGFR mutations and were reported in detail.¹

We disagree with Weber et al about the absence of real hot spots for mutations in the EGFR gene. At present, 297 mutations in the tyrosine kinase domain of EGFR have been reported in the literature. Of these, 120 (40.4%) are located at codon 858 (L858R), 16 (5.4%) at codon 719 (G719C), and 136 (45.8%) are small in-frame deletion between codons 746 and 759. Taken together, these three alterations account for 92% of the mutations so far detected in the EGFR gene. Moreover, it is widely recognized that the SSCP analysis can be successfully used, not only to evaluate the presence of hot spot mutations, but also for the screening of unknown mutations.⁵ In particular, the accuracy of SSCP is extremely high for the detection of small DNA deletions/insertions, which represent the most frequent genetic abnormalities found in the tyrosine kinase domain of EGFR and ErbB2 genes (personal observation).

We did not use heteroduplex analysis. However, we and others^6,7 have shown that SSCP is more sensitive than direct sequencing, allowing the detection of as little as 5% to 10% mutant alleles in a DNA sample, while sequencing requires at least 25% to 30% of mutated DNA, as reported by Weber et al in their letter. In our opinion, the high sensitivity of SSCP is not a limit, but an advantage, especially when undissected tumor tissues, small biopsies, or aggressive subclones of neoplastic cells within the tumor mass have to be analyzed for the presence of mutations.

As pointed out by Weber et al, we did not examine exon 20. We decided to evaluate exons 18, 19, and 21 of EGFR since in previous reports, only these exons were found to be affected by somatic mutations. Infrequent mutations in exon 20 have later been reported in lung cancer patients (approximately 4% of all the mutations so far detected). Interestingly, these mutations were often coupled with mutations in other exons. Very recently, a particular mutation in exon 20 (T790M) has been associated with acquired resistance to gefitinib.⁸ We have therefore developed a couple of primers for SSCP screening of exon 20 (available by request). The high sensitivity of SSCP, as discussed above, could allow the detection of T790M mutations in a small fraction of growing neoplastic cells resistant to gefitinib. This may allow for an early prediction of chemoresistance during pharmacologic treatment.

We agree that investigators studying EGFR somatic mutations for prediction of response to tyrosine kinase inhibitors should come together in a consortium. However, all over the world, and especially in Eastern countries, there is an urgent need to know the status of the EGFR gene in all patients with lung adenocarcinoma. The SSCP analysis we developed can be easily diffused as a screening method since it is accurate, relatively simple, quick, and does not require expensive equipment.