Toward Noninvasive Genomic Screening of Lung Cancer Patients

Abstract

Understanding the molecular profile of non–small-cell lung cancer (NSCLC) is becoming ever more relevant to patient care and treatment decision making. The development of molecularly targeted therapies against the epidermal growth factor receptor (EGFR) tyrosine kinase has led to the discovery of mutations and other genomic alterations that affect the efficacy of these agents. A set of activating mutations in EGFR, primarily overlapping deletion mutations in exon 19 and the single L858R point mutation in exon 21, are associated with increased response and survival after tyrosine kinase inhibitor (TKI) therapy, whereas the T790M point mutation or insertion mutations in exon 20 of EGFR, amplification of MET, and mutations in K-ras are associated with failure to respond. As a result, the field is moving toward acceptance of molecular testing as an integral component of clinical care. However, several practical obstacles to this paradigm shift remain. These include the acquisition and availability of appropriate tissue samples and the time required for analysis. The lung cancer diagnostic procedure of choice is frequently a fine needle aspiration, either via bronchoscopy or using the computed tomography–guided percutaneous approach, which often yields insufficient material for performing one or multiple molecular analyses. When considering EGFR mutation screening, the problem of limited tissue is confounded by reliance on the direct sequencing method, which is not highly sensitive and often results in falsely negative or noninformative results when performed on cytology specimens. Direct sequencing can often take up to 3 to 4 weeks to yield results, which is not clinically optimal. Consequently, there is great interest in establishing alternate methods of EGFR mutation screening other than direct sequencing, as well as developing blood-based methods of analyzing these somatic mutations. In this issue of the Journal of Clinical Oncology, Bai and colleagues present the results from a prospective clinical trial designed to assess the value of noninvasive EGFR mutation analysis via denaturing high-performance liquid chromatography (DHPLC) analysis of peripheral blood specimens. DHPLC has previously been used to analyze tumor tissue biopsies for EGFR mutation with promising results. Janne et al demonstrated that by predigesting samples with a DNA endonuclease and then performing DHPLC, they could detect EGFR mutations with 100% sensitivity and 87% specificity compared to standard direct sequencing. Similarly, both Cohen and Chin found that DHPLC was more sensitive, more time efficient, and less costly than direct sequencing of tumor tissue. In contrast to standard direct sequencing, DHPLC involves denaturing polymerase chain reaction products with heat followed by cooling to permit renaturation and the formation of duplexes. The heteroduplexes formed in the presence of a heterozygous EGFR mutation differ in HPLC column retention time compared with the homoduplexes formed with wild-type EGFR, and are therefore detectable as an aberrant peak on the data readout. Although homozygous EGFR mutations are not detectable with this method, they are extraordinarily rare. Noting that a variation in elution time, or an aberrant peak, exists on DHPLC analysis does not define the specific mutation present, and direct sequencing of that peak is still necessary to confirm the exact mutation. In this case, Bai and colleagues first established standard readout patterns observed on DHPLC with the most common activating EGFR mutations, and then classified their subsequent results on the basis of these patterns. The study team enrolled 230 consecutive patients with advanced NSCLC from the Beijing Cancer Hospital over a 3-year period. All patients had primary tumor samples available for DHPLC analysis and comparison to the plasma DHPLC results. The primary aim of the study was to determine if plasma DHPLC analysis was a valid surrogate for tumor tissue EGFR mutation analysis. They detected EGFR mutations in 34% of the plasma samples and 33% of the tumor tissue samples, although this did not represent the same cohort of patients using each method. There was 80% concordance between DHPLC analysis of tumor tissue and of plasma samples, and six of the 30 discordant samples were analyzed by direct sequencing to confirm that the mismatched results were accurate. The investigators then assessed the patient population for clinical outcome and confirmed that EGFR mutations identified in the plasma were more frequent among patients with adenocarcinoma and low-smoking histories, and predicted for a significantly higher response and progression-free survival among patients treated with second-line gefitinib, consistent with results from prior studies analyzing tumor tissue. To our knowledge, this study represents the first prospective clinical trial examining the ability to detect EGFR mutations from a blood-based assay and signifies an important advance for the field. As noted earlier, tumor biopsy tissue from NSCLC patients is often too limited to perform molecular analyses. Furthermore, though mechanisms of resistance to EGFR-targeted therapies are well described and identification of these could inform choice of the most appropriate salvage therapies or clinical trials, serial molecular analyses from repeat tumor biopsies are simply not feasible outside the setting of a tertiary research hospital. Consequently, a reliable blood-based EGFR JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 27 NUMBER 16 JUNE 1 2009