Purpose: Epidermal growth factor receptor (EGFR) mutations, especially in-frame deletions in exon 19 (DEL) and a point mutation in exon 21 (L858R), predict gefitinib sensitivity in patients with non–small cell lung cancer (NSCLC). In this study, we verified the accuracy of EGFR mutation analysis in small samples by high-resolution melting analysis (HRMA), which is a rapid method using PCR amplification with a dye to analyze the melting curves in NSCLC.

Experimental Design: We designed a prospective study to compare the sensitivity and specificity of HRMA and DNA sequencing with laser capture microdissection. Eligible patients with lung lesions were screened by bronchoscopy or percutaneous needle biopsy to histologically confirm the diagnosis, followed by surgical resection of the NSCLC. Small diagnostic specimens were analyzed for EGFR mutations by HRMA, and the surgically resected specimens were examined for mutations by HRMA and DNA sequencing.

Results: The analyses for EGFR mutations were conducted in 52 eligible cases of the 92 enrolled patients. EGFR mutations were detected in 18 (34.6%) patients. The results of HRMA from surgically resected specimens as well as DNA sequencing revealed 100% sensitivity and specificity. On the other hand, the sensitivity and specificity of HRMA from the small diagnostic specimens were 83.3% and 100%, respectively.

Conclusions: In this study, we showed that HRMA is a highly accurate method for detecting DEL and L858R mutations in patients with NSCLC, although it is necessary to consider the identification of patients with a false-negative result when the analysis is conducted using small samples.

Somatic mutations in the kinase domain of the epidermal growth factor receptor (EGFR) have been reported in patients with non–small cell lung cancer (NSCLC; refs. 13). Although many types of EGFR mutations have been identified, they seem to be concentrated in exons 18 to 21 of EGFR; ∼85% to 90% of EGFR-mutant patients have mutations in two hotspots: a short in-frame deletion in exon 19 (DEL) and a point mutation at codon 858 in exon 21 (L858R; ref. 4). Several studies have revealed that EGFR mutations are strongly associated with the tumor response and clinical outcome in patients with NSCLC receiving treatment with EGFR tyrosine kinase inhibitors, such as gefitinib (Iressa, AstraZeneca; refs. 57). The mutational status of EGFR, especially the presence/absence of DEL and L858R, is a strong predictor of the sensitivity to EGFR tyrosine kinase inhibitor, and the detection of EGFR mutations is useful for decision-making by both patients and physicians (4, 8). Recently, a laboratory test for EGFR mutations has become clinically available for guiding treatment decisions.

Until now, screening for these mutations has most commonly been conducted using DNA sequencing methods. In our previous study, we used methanol-fixed, paraffin-embedded surgical specimens and performed direct sequencing and pyrosequencing with laser capture microdissection (LCM) to ensure high-quality genetic analysis of archived tissues (5, 9). However, these approaches are not useful in clinical practice for two reasons. First, although the sequencing methods require a high ratio of tumor-to-normal tissue DNA for optimal results, the diagnostic specimens obtained from cases of advanced NSCLC may contain only a small amount of tumor cells and are highly contaminated with normal cells. Secondly, EGFR mutation analysis based on DNA sequencing requires special instruments and is also time-consuming and expensive. Therefore, some simple and highly sensitive nonsequencing methods to detect EGFR mutations have been reported (1022). However, the accuracy of these methods for clinical use have not been assessed in prospective studies.

High-resolution melting analysis (HRMA) using the LCGreen I (Idaho Technology) dye was introduced as an easy, quick, and inexpensive method for the screening of mutations (23), and we established and validated the HRMA method to detect DEL and L858R mutations in cases of NSCLC (9, 10). Our cell line study revealed that DEL and L858R mutations could be detected using HRMA in the presence of 10% and 0.1% of mutant cells, respectively (10). We also showed that the two major mutations could be identified by HRMA retrospectively using DNA extracted from archived Papanicolaou-stained cytologic slides with 88% sensitivity and 100% specificity (9). Furthermore, it was shown that among patients treated with gefitinib, the response rate (78% versus 8%), time-to-progression (median, 9.2 versus 1.6 months), and overall survival (median, 21.7 versus 8.7 months) were significantly better in patients with EGFR mutations than with wild-type EGFR (P < 0.001), as detected by HRMA (9). These results suggest that this easy, quick, and inexpensive method which was done using diagnostic small samples of advanced NSCLC tumors is one of the most useful and precise methods to detect EGFR mutations in clinical practice.

In this study, we designed a prospective study to detect two major EGFR mutations by HRMA using small diagnostic cytologic or biopsy specimens and surgically resected specimens, and the results were compared with the results of DNA sequencing methods combined with LCM, which we consider as the “gold standard” for such detection, applied to methanol-fixed, paraffin-embedded surgically resected specimens. We evaluated the diagnostic sensitivity, specificity, predictive values, and accuracy of the detection of EGFR mutations using HRMA and revealed that this method is feasible for clinical use to detect EGFR mutations in small samples obtained from patients with NSCLC.

Patients and materials. Patients with lung lesions, which were suspected clinically to be operable NSCLC, were enrolled in this prospective study. The patients were scheduled for bronchoscopy or percutaneous needle biopsy to establish the histologic diagnosis, and informed consent was obtained from each of the patients prior to these diagnostic procedures. Thereafter, the patients diagnosed with NSCLC underwent lung surgery at our hospital. In this study, mutational analysis of EGFR was done by HRMA or DNA sequencing methods combined with LCM in all the patients in which both the preoperatively obtained diagnostic specimens and the resected specimens were histologically confirmed by a certified pathologist to contain malignant cells.

Based on a protocol approved by the Institutional Review Board of the National Cancer Center, we did mutational analyses of EGFR to detect DEL and L858R in the eligible patients. The Papanicolaou-stained cytologic slides (n = 35), formalin-fixed, paraffin-embedded transbronchial or percutaneous needle biopsy specimens (n = 34), and methanol-fixed, paraffin-embedded surgically resected specimens subjected to LCM using a PixCell II LCM system (Arcturus Engineering, Inc.; n = 52) were collected prospectively. DNA was extracted using the QIAamp DNA Micro Kit (Qiagen), as described in our previous report (10).

HRMA. PCR was done to amplify exons 19 or 21 of EGFR using LCGreen I (Idaho Technology) on a LightCycler (Roche Diagnostics) and primers designed as previously described (10). If the first PCR products were not available for the mutational analyses of the melting curves, we did a second PCR using the same primers. These PCR products were denatured at 95°C for 10 min and cooled to 40°C to promote the formation of heteroduplexes. The LightCycler capillary was transferred to an HR-1 (Idaho Technology), an HRMA instrument, and heated at a transition rate of 0.3°C/s. Data were acquired and analyzed using the accompanying software (Idaho Technology). After normalization and temperature-adjustment steps, melting curve shapes from 78.5°C to 85.5°C were compared between the tumor samples and control samples. Human Genomic DNA (Roche Diagnostics) was used as the negative control sample with wild-type EGFR. Samples revealing skewed or left-shifted curves as compared with the control samples were judged to have mutations without positive controls (9, 10). All analyses were done in a blinded fashion by two researchers (T. Fukui and T. Takano). After independent evaluation by the two researchers, the final judgment was arrived at by consensus after joint viewing of the melting curves from both.

DNA sequencing methods with LCM. In our previous study, we did a direct sequencing or pyrosequencing of EGFR in patients with recurrent NSCLC after primary surgery (5). Based on the results of our previous study, we consider direct sequencing with LCM for the detection of DEL and pyrosequencing with LCM for the detection of L858R as the gold standard in relation to EGFR mutational analysis. DNA was extracted from methanol-fixed, paraffin-embedded surgical specimens by LCM, according to a previously described method (24). Direct sequencing of the PCR products for DEL was done using ABI PRISM3700 and 3100 DNA sequencers (Applied Biosystems). Pyrosequencing to analyze L858R was done using Pyrosequencing PSQ 96MA (Pyrosequencing; refs. 5, 25). The EGFR mutational analysis using DNA sequencing methods was done in a blinded fashion by a researcher (H. Sakamoto) according to a previously described method (5), and then compared with the corresponding results obtained using HRMA.

Statistical analysis. The primary end point of this study was the sensitivity and specificity of the results obtained using HRMA as compared with those of the results obtained using DNA sequencing with LCM. The sample size was calculated using a statistical power level of 0.80 and two-sided α level of 0.1 on the basis of an estimated sensitivity of at least 0.80 and an expected value of 0.95 for HRMA, a minimum of 20 patients with EGFR-mutated tumors were required. Because the percentage of NSCLC patients with EGFR mutations was expected to be 40% in this study population composed of only Japanese, approximately 50 patients with NSCLC were needed. Therefore, considering a specificity of at least 0.80 and the expected value of 0.95 for HRMA, 30 patients with wild-type tumors showed a statistical power level of 0.90 using a two-sided α level of 0.1.

The associations between mutational status and patient characteristics were assessed by a χ2 test using the SPSS statistical package (SPSS version 11.0 for Windows; SPCC, Inc.).

Patient characteristics. From December 2005 to December 2006, 92 patients with clinically suspected operable NSCLC were enrolled in this study. The following diagnostic procedures were done preoperatively in 90 patients: bronchoscopy (n = 57), percutaneous needle biopsy (n = 27), or bronchoscopy followed by percutaneous needle biopsy (n = 6). The patient characteristics are shown in Table 1. All the patients were Japanese. Among the patients, a definitive diagnosis was established in 85 patients by bronchoscopy in 43 of 59 patients (72.9%) and by percutaneous needle biopsy in 25 of 31 patients (80.6%); in 18 of the 85 (21.2%) patients, the histologic diagnosis could not be established preoperatively by bronchoscopy and/or percutaneous needle biopsy, the patients underwent lung surgery for suspicious malignant lung lesion, and examination of the resected specimens revealed the diagnosis of primary NSCLC in 17 and malignant lymphoma in 1 of the 18 patients. Among the 76 patients diagnosed to have primary NSCLC, 73 consented to undergo lung surgery. Finally, the analysis for EGFR mutations was done on 52 patients with a definitive histologic diagnosis of primary NSCLC, established both by examination of the preoperative diagnostic specimens and of the corresponding resected specimens (Fig. 1).

Table 1.

Patient characteristics

(A) Characteristics of all the patients enrolled in this study (n = 92)
All (n = 92)BF (n = 64)PNB (n = 34)*
Age, year, median (range) 64 (34-84) 64 (38-84) 62 (41-79)
Gender (male/female) 58/34 41/23 23/11
Smoking history (N/F/C) 29/30/33 23/19/22 7/14/13
Tumor size, mm, average (range) 27.2 (10.2-73.4) 28.3 (13.8-56.6) 24.5 (10.2-73.4)
Accuracy of the diagnostic procedure (%) 66/85 (77.6) 43/59 (72.9) 25/31 (80.6)
Accuracy of the cytologic slides (%) 54/85 (63.5) 31/59 (52.5) 23/30 (76.7)
Accuracy of the biopsy specimens (%) 42/62 (67.7) 35/54 (64.8) 7/9 (77.8)

(B) Characteristics of the patients who underwent analysis of the EGFR mutations in this study (n = 52)

All (n = 52)

BF (n = 38)

PNB (n = 17)

Age, year, median (range) 64.5 (34-84) 64.5 (34-84) 64 (47-78)
Gender (male/female) 36/16 25/13 14/3
Smoking history (N/F/C) 16/17/19 15/11/12 1/7/9
Tumor size, mm, average (range) 27.0 (11.0-56.6) 28.3 (20.6-56.6) 24.1 (11.0-48.8)
Postoperative diagnosis (Ad/Sq/LCNEC) 45/5/2 34/4/0 12/3/2
Pathologic stage (IA/B, IIA/B, IIIA/B) 19/13, 3/5, 9/2 15/8, 3/2, 8/2 7/5, 0/2, 3/0
(A) Characteristics of all the patients enrolled in this study (n = 92)
All (n = 92)BF (n = 64)PNB (n = 34)*
Age, year, median (range) 64 (34-84) 64 (38-84) 62 (41-79)
Gender (male/female) 58/34 41/23 23/11
Smoking history (N/F/C) 29/30/33 23/19/22 7/14/13
Tumor size, mm, average (range) 27.2 (10.2-73.4) 28.3 (13.8-56.6) 24.5 (10.2-73.4)
Accuracy of the diagnostic procedure (%) 66/85 (77.6) 43/59 (72.9) 25/31 (80.6)
Accuracy of the cytologic slides (%) 54/85 (63.5) 31/59 (52.5) 23/30 (76.7)
Accuracy of the biopsy specimens (%) 42/62 (67.7) 35/54 (64.8) 7/9 (77.8)

(B) Characteristics of the patients who underwent analysis of the EGFR mutations in this study (n = 52)

All (n = 52)

BF (n = 38)

PNB (n = 17)

Age, year, median (range) 64.5 (34-84) 64.5 (34-84) 64 (47-78)
Gender (male/female) 36/16 25/13 14/3
Smoking history (N/F/C) 16/17/19 15/11/12 1/7/9
Tumor size, mm, average (range) 27.0 (11.0-56.6) 28.3 (20.6-56.6) 24.1 (11.0-48.8)
Postoperative diagnosis (Ad/Sq/LCNEC) 45/5/2 34/4/0 12/3/2
Pathologic stage (IA/B, IIA/B, IIIA/B) 19/13, 3/5, 9/2 15/8, 3/2, 8/2 7/5, 0/2, 3/0

NOTE: Never smokers were defined as patients who had never smoked, former smokers were defined as patients who had stopped smoking at least 1 y before the diagnosis, and current smokers were defined as patients who were still smoking at the time of the diagnosis.

Abbreviations: BF, bronchoscopy; PNB, percutaneous needle biopsy; N, never smoker; F, former smoker; C, current smoker; Ad, adenocarcinoma; Sq, squamous cell carcinoma; LCNEC, large cell neuroendocrine carcinoma.

*

Including six patients in whom bronchoscopy was done followed by percutaneous needle biopsy.

Including three in whom bronchoscopy was done followed by percutaneous needle biopsy.

Fig. 1.

Flowchart of the analyses conducted in 92 enrolled patients with lung tumors in this study.

Fig. 1.

Flowchart of the analyses conducted in 92 enrolled patients with lung tumors in this study.

Close modal

Mutational analyses. We analyzed 35 cytologic samples and 34 biopsy specimens obtained from 52 patients by HRMA, and did both HRMA and DNA sequencing with LCM in the 52 resected specimens corresponding to the 52 patients. Among the 52 surgically resected specimens analyzed by DNA sequencing with LCM, there were 18 (34.6%) samples with EGFR mutations, 5 with DEL mutations, and 13 with L858R mutations. As shown in Table 2, the EGFR mutations were detected more frequently in women, never-smokers, and patients with a histologic diagnosis of adenocarcinoma. All results from HRMA done in a blinded fashion by two researchers (T. Fukui and T. Takano) were consistent.

Table 2.

EGFR mutation status among the patient subgroups

nEGFR mutations*
P
DELL858RTotal%
Total 52 13 18 34.6 —
Gender
Women 16 11 68.8 0.001
Men 36 19.4
Smoking history
Never 16 11 68.8 0.001
Former 17 35.3
Current 19 5.3
Histology
Ad 44 13 18 100 0.025
Sq
LCNEC
nEGFR mutations*
P
DELL858RTotal%
Total 52 13 18 34.6 —
Gender
Women 16 11 68.8 0.001
Men 36 19.4
Smoking history
Never 16 11 68.8 0.001
Former 17 35.3
Current 19 5.3
Histology
Ad 44 13 18 100 0.025
Sq
LCNEC

Abbreviations: DEL, deletional mutations in exon 19; L858R, a point mutation at codon 858 in exon 21; Ad, adenocarcinoma; Sq, squamous cell carcinoma; LCNEC, large cell neuroendocrine carcinoma.

*

The EGFR mutations were analyzed by DNA sequencing with LCM.

Comparison between never smokers and others.

Comparison between adenocarcinoma and others.

HRMA could be conducted using small diagnostic samples from all 52 patients, although the analysis needed to be conducted using the second PCR product in 15 cases. In the analysis of exon 19, 5 samples revealed different curves from the control and 47 samples revealed almost the same curves as the control; therefore, we judged that the five former patients had DEL mutations (Fig. 2A). In the analysis of exon 21, 10 samples revealed a left-shift from the control and 42 samples revealed almost the same curves as the control; therefore, we judged that the 10 former patients had L858R mutations (Fig. 2B). All the 52 surgically resected specimens analyzed by DNA sequencing with LCM could also be analyzed by HRMA, although the analysis needed to be conducted using the second PCR product in two cases. DEL mutations were detected in 5 patients (Fig. 2C) and L858R mutations in 13 patients (Fig. 2D) among the 52 patients. Of the 52 specimens, both cytologic slides and biopsy specimens were analyzed in 17 cases. Discrepant results were obtained by HRMA in one of the cases, with L858R mutation being detected in the cytologic slides but not in the biopsy specimens. We included this patient in the population with L858R mutations.

Fig. 2.

Adjusted melting curves obtained by HRMA of the samples in this study to detect EGFR mutations (↑), in-frame deletions in exon 19 (A, small samples; C, resected specimens) and a point mutation in exon 21 (B, small samples; D, resected specimens). Each sample that revealed a skewed or left-shifted curve from those of the control sample was judged to have a mutation.

Fig. 2.

Adjusted melting curves obtained by HRMA of the samples in this study to detect EGFR mutations (↑), in-frame deletions in exon 19 (A, small samples; C, resected specimens) and a point mutation in exon 21 (B, small samples; D, resected specimens). Each sample that revealed a skewed or left-shifted curve from those of the control sample was judged to have a mutation.

Close modal

The results of HRMA were consistent with the results of DNA sequencing with LCM in all the surgically resected specimens analyzed by the two methods. On the other hand, HRMA using small diagnostic specimens revealed the wild-type curve in three cases, although analysis of the corresponding surgically resected specimens analyzed by pyrosequencing with LCM revealed the L858R mutation (Table 3). Thus, the results for these samples obtained by HRMA were considered as false-negative results. Neither method of analysis yielded any false-positive cases. The results of the EGFR mutational analysis by HRMA compared with DNA sequencing with LCM using surgically resected specimens were shown in Table 4. The sensitivity, specificity, and accuracy of HRMA using small diagnostic specimens were 83.3%, 100%, and 94.2%, respectively. Using surgically resected specimens, those of HRMA were all 100%.

Table 3.

Results of the EGFR mutation analyses in patients with EGFR mutation–positive tumors

No. of patientsSmall samples
Surgically resected specimens
HRMAHRMASequence with LCM
13 DEL DEL DEL1*
26 DEL DEL DEL1*
32 DEL DEL DEL2
40 DEL DEL DEL2
47 DEL DEL DEL1*
L858R L858R L858R
Wild-type L858R L858R
12 L858R L858R L858R
18 L858R L858R L858R
21 L858R L858R L858R
23 L858R L858R L858R
25 Wild-type L858R L858R
27 L858R L858R L858R
28 L858R L858R L858R
31 Wild-type L858R L858R
41 L858R L858R L858R
53 L858R L858R L858R
54 L858R L858R L858R
No. of patientsSmall samples
Surgically resected specimens
HRMAHRMASequence with LCM
13 DEL DEL DEL1*
26 DEL DEL DEL1*
32 DEL DEL DEL2
40 DEL DEL DEL2
47 DEL DEL DEL1*
L858R L858R L858R
Wild-type L858R L858R
12 L858R L858R L858R
18 L858R L858R L858R
21 L858R L858R L858R
23 L858R L858R L858R
25 Wild-type L858R L858R
27 L858R L858R L858R
28 L858R L858R L858R
31 Wild-type L858R L858R
41 L858R L858R L858R
53 L858R L858R L858R
54 L858R L858R L858R

Abbreviations: DEL, deletional mutations in exon 19; L858R, a point mutation at codon 858 in exon 21.

*

DEL1: del E746-A750 (del 2235-2249).

DEL2: del E746-A750 (del 2236-2250).

The analyses by HRMA were done using second PCR products.

Table 4.

Comparison of the sensitivity, specificity, predictive values, and accuracy between HRMA and DNA sequencing with LCM (n = 52)

HRMA using small samplesHRMA using surgically resected specimens
True-positive 15 18
True-negative 34 34
False-positive
False-negative
Sensitivity 83.3 (68.9-97.8) 100
Specificity 100 100
NPV 91.9 (84.5-99.3) 100
PPV 100 100
Accuracy 94.2 (88.9-99.5) 100
HRMA using small samplesHRMA using surgically resected specimens
True-positive 15 18
True-negative 34 34
False-positive
False-negative
Sensitivity 83.3 (68.9-97.8) 100
Specificity 100 100
NPV 91.9 (84.5-99.3) 100
PPV 100 100
Accuracy 94.2 (88.9-99.5) 100

NOTE: The results of these analyses were compared with those of DNA sequencing with LCM (used as the gold standard in this study). Data are presented as % or % (90% confidence interval). True-positive is defined as the correct detection of DEL in exon 19 or L858R in exon 21.

Abbreviations: NPV, negative predictive value; PPV, positive predictive value.

In this prospective study, we showed the high accuracy of the HRMA method for detecting two major EGFR mutations, DEL and L858R in patients with NSCLC. The accuracy of HRMA was clearly equal to that of DNA sequencing with LCM for the detection of mutations in surgically resected specimens. On the other hand, the sensitivity and specificity of HRMA were 83.3% (90% confidence interval: 68.9-97.7%) and 100%, respectively, when the small diagnostic samples were analyzed. Although the sensitivity of HRMA which was estimated to be at least 0.80 did not reach statistical significance, we consider HRMA as one of the available methods for the detection of EGFR mutations in clinical practice because the specificity, which is important for clinical decision-making, of HRMA was 100% and the EGFR mutation rate was less than the expected 40% to secure enough statistical power in this study.

Recently, many researchers reported establishing simple and highly sensitive nonsequencing methods for detecting EGFR mutations using small tumor samples (1122), and the results of several mutation analyses were correlated with the clinical outcome of EGFR tyrosine kinase inhibitor treatment (1719). Using serial dilution studies, some researchers have reported methods that are able to detect mutations in samples containing ∼0.1% to 10% mutated DNA (13, 14, 1618, 2022), as opposed to direct DNA sequencing which requires the presence of at least 10% to 30% of mutated DNA in the samples (18, 20). Additionally, several novel methods offered higher sensitivity and specificity than DNA sequencing to identify the mutations in clinical samples. But almost none of the methods were validated for diagnostic accuracy in a prospective study, and we therefore consider these methods to still be unsuitable for routine clinical examination. Although these nonsequencing methods were not mutually compared, based on our previous results of retrospectively verifying the accuracy of HRMA (9, 10), we thought to develop in this prospective study an easy, quick (PCR for ∼1 hour and HRMA for 2 to 3 minutes), and inexpensive (at a running cost per sample of approximately $7.50, which consisted of$5.50 for the DNA extract and less than \$2.00 for PCR using LCGreen I dye) method that might be useful in clinical practice with a great advantage over DNA sequencing, which requires the extraction of high-quality DNA from an adequate amount of pure tumor cells, takes a long time, and is expensive.

In this study, the three patients with L858R detected by DNA pyrosequencing with LCM using the surgically resected specimens were labeled as having the wild-type EGFR in the analyses conducted using the small diagnostic samples. With regard to these false-negative results, the following three points need to be discussed: first, our previous study, conducted using human lung cancer cell lines, showed that HRMA can detect the mutations, even when samples contain only a small proportion (DEL, 10%; L858R, 0.1%) of mutant cells (10). In this study, the sensitivity of HRMA was also considered to be sufficiently high for the detection of EGFR mutations, especially L858R, even when the analysis was conducted using small samples after evaluation by a clinical pathologist to determine if they contained benign or malignant cells. Thus, we assume a higher accuracy of HRMA when using small samples in clinical practice. Although it still needs to be comparatively analyzed with the previously reported nonsequencing methods, HRMA can be considered as one of the sensitive methods available for the detection to EGFR mutations in clinical practice.

Second, high-quality DNA should be preserved in clinical samples to obtain the best results. There always remains the risk of an indeterminate or false-negative result because the DNA might have degenerated during sampling or during the preservation of clinical samples. In a comparison between the cytologic slides and biopsy specimens, better results were obtained from analyses of the first PCR products using the cytologic slides rather than the results obtained using the biopsy specimens, regardless of the amount of tumor cells examined (Table 5). This could probably be explained by the differences in the method of sample fixation between the two types of specimens. It has been suggested by a previous report that DNA is preserved better in the methanol-fixed samples than in the formalin-fixed specimens (26). Therefore, if we used methanol for specimen fixation of biopsy specimens, the results of HRMA using the first PCR products from small biopsy samples might improve. Hereafter, we propose to perform mutation analyses using methanol-fixed specimens, if possible.

Table 5.

Results of HRMA using cytologic slides or biopsy specimens

Cytologic slides (n = 35)
Biopsy specimens (n = 34)
First PCRSecond PCRFirst PCRSecond PCR
Successfully analyzed 29 (83.0%) 35 (100%) 5 (15.0%) 34 (100%)
True-positive 11 10
True-negative 19 21 22
True-negative
False-positive
Sensitivity 70.0% (7/10) 78.6% (11/14) 100% (1/1) 83.3% (10/12)
Specificity 100% (19/19) 100% (21/21) 100% (4/4) 100% (22/22)
NPV 100% (7/7) 100% (11/11) 100% (1/1) 100% (10/10)
PPV 86.4% (19/22) 87.5% (21/24) 100% (4/4) 91.2% (22/24)
Accuracy 89.7% (26/29) 91.4% (32/35) 100% (5/5) 94.1% (32/34)
Cytologic slides (n = 35)
Biopsy specimens (n = 34)
First PCRSecond PCRFirst PCRSecond PCR
Successfully analyzed 29 (83.0%) 35 (100%) 5 (15.0%) 34 (100%)
True-positive 11 10
True-negative 19 21 22
True-negative
False-positive
Sensitivity 70.0% (7/10) 78.6% (11/14) 100% (1/1) 83.3% (10/12)
Specificity 100% (19/19) 100% (21/21) 100% (4/4) 100% (22/22)
NPV 100% (7/7) 100% (11/11) 100% (1/1) 100% (10/10)
PPV 86.4% (19/22) 87.5% (21/24) 100% (4/4) 91.2% (22/24)
Accuracy 89.7% (26/29) 91.4% (32/35) 100% (5/5) 94.1% (32/34)

NOTE: The results of these analyses were compared with those of DNA sequencing with LCM (used as the gold standard in this study). True-positive is defined as the correct detection of DEL in exon 19 or L858R in exon 21.

Abbreviations: NPV, negative predictive value; PPV, positive predictive value.

Finally, we need to consider the possibility of intratumoral heterogeneity, and small diagnostic samples and surgically resected specimens may each represent overlapping but different populations of these tumor cells. A lack of association in the immunohistochemical expression profile between lung biopsy specimens and the corresponding resected tumor specimens has been reported (27). Furthermore, intratumoral heterogeneity was shown not only in terms of microheterogeneity of the tumor cell phenotype (28), but in terms of genetic heterogeneity in cancer (29, 30). In particular, the intratumoral genetic heterogeneity of EGFR mutations may explain the variable clinical response of NSCLC to gefitinib. It is also possible that the small diagnostic samples contain only wild-type cells, even if the tumor, overall, shows mutations, because the small samples yield only small part of the tumor. It is always necessary to consider the possibility of a false-negative result of mutational analyses conducted using the small samples.

In the current prospective study, we showed the feasibility and high accuracy of using HRMA for detecting two major EGFR mutations, DEL and L858R, in patients with NSCLC. Although HRMA showed high accuracy, the possibility of indeterminate or false-negative results, and because of the sensitivity of this method, the quality of DNA preservation in the clinical samples or intratumoral genetic heterogeneity, must be borne in mind to a certain extent when this analysis is conducted using small diagnostic samples. Therefore, HRMA should not be used to exclude patients from EGFR tyrosine kinase inhibitor treatment on the basis of the negative results only. Based on the results of this prospective study, we suggest that this method is very useful for clinical decision-making, especially in patients with a positive result.

No potential conflicts of interest were disclosed.

Grant support: Promotion of Fundamental Studies in Health Sciences of the National Institute of Biomedical Innovation, a Health and Labour Science Research grant from the Ministry of Health, Labour and Welfare, Japan, and a grant-in-aid for Young Scientists from the Ministry of Education, Culture, Sports, Science and Technology, Japan.

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.

We thank Kiyoaki Nomoto, Karin Yokozawa, Chizu Kina, Sachiko Miura, Misuzu Okuyama, Sachiyo Mimaki, and Chie Hirama for their technical support.

1
Lynch TJ, Bell DW, Sordella R, et al. Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib.
N Engl J Med
2004
;
350
:
2129
–39.
2
Paez JG, Janne PA, Lee JC, et al. EGFR mutations in lung cancer: correlation with clinical response to gefitinib therapy.
Science
2004
;
304
:
1497
–500.
3
Pao W, Miller V, Zakowski M, et al. EGF receptor gene mutations are common in lung cancers from “never smokers” and are associated with sensitivity of tumors to gefitinib and erlotinib.
Proc Natl Acad Sci U S A
2004
;
101
:
13306
–11.
4
Janne PA, Engelman JA, Johnson BE. Epidermal growth factor receptor mutations in non-small-cell lung cancer: implications for treatment and tumor biology.
J Clin Oncol
2005
;
23
:
3227
–34.
5
Takano T, Ohe Y, Sakamoto H, et al. Epidermal growth factor receptor gene mutations and increased copy numbers predict gefitinib sensitivity in patients with recurrent non-small-cell lung cancer.
J Clin Oncol
2005
;
23
:
6829
–37.
6
Mitsudomi T, Kosaka T, Endoh H, et al. Mutations of the epidermal growth factor receptor gene predict prolonged survival after gefitinib treatment in patients with non-small-cell lung cancer with postoperative recurrence.
J Clin Oncol
2005
;
23
:
2513
–20.
7
Han SW, Kim TY, Hwang PG, et al. Predictive and prognostic impact of epidermal growth factor receptor mutation in non-small-cell lung cancer patients treated with gefitinib.
J Clin Oncol
2005
;
23
:
2493
–501.
8
Sequist LV, Bell DW, Lynch TJ, Haber DA. Molecular predictors of response to epidermal growth factor receptor antagonists in non-small-cell lung cancer.
J Clin Oncol
2007
;
25
:
587
–95.
9
Takano T, Ohe Y, Tsuta K, et al. Epidermal growth factor receptor mutation detection using high-resolution melting analysis predicts outcomes in patients with advanced non-small cell lung cancer treated with gefitinib.
Clin Cancer Res
2007
;
13
:
5385
–90.
10
Nomoto K, Tsuta K, Takano T, et al. Detection of EGFR mutations in archived cytologic specimens of non-small cell lung cancer using high-resolution melting analysis.
Am J Clin Pathol
2006
;
126
:
608
–15.
11
Marchetti A, Martella C, Felicioni L, et al. EGFR mutations in non-small-cell lung cancer: analysis of a large series of cases and development of a rapid and sensitive method for diagnostic screening with potential implications on pharmacologic treatment.
J Clin Oncol
2005
;
23
:
857
–65.
12
Sasaki H, Endo K, Konishi A, et al. EGFR Mutation status in Japanese lung cancer patients: genotyping analysis using LightCycler.
Clin Cancer Res
2005
;
11
:
2924
–9.
13
Pan Q, Pao W, Ladanyi M. Rapid polymerase chain reaction-based detection of epidermal growth factor receptor gene mutations in lung adenocarcinomas.
J Mol Diagn
2005
;
7
:
396
–403.
14
Nagai Y, Miyazawa H, Huqun, et al. Genetic heterogeneity of the epidermal growth factor receptor in non-small cell lung cancer cell lines revealed by a rapid and sensitive detection system, the peptide nucleic acid-locked nucleic acid PCR clamp.
Cancer Res
2005
;
65
:
7276
–82.
15
Endo K, Konishi A, Sasaki H, et al. Epidermal growth factor receptor gene mutation in non-small cell lung cancer using highly sensitive and fast TaqMan PCR assay.
Lung Cancer
2005
;
50
:
375
–84.
16
Janne PA, Borras AM, Kuang Y, et al. A rapid and sensitive enzymatic method for epidermal growth factor receptor mutation screening.
Clin Cancer Res
2006
;
12
:
751
–8.
17
Yatabe Y, Hida T, Horio Y, Kosaka T, Takahashi T, Mitsudomi T. A rapid, sensitive assay to detect EGFR mutation in small biopsy specimens from lung cancer.
J Mol Diagn
2006
;
8
:
335
–41.
18
Kimura H, Fujiwara Y, Sone T, et al. High sensitivity detection of epidermal growth factor receptor mutations in the pleural effusion of non-small cell lung cancer patients.
Cancer Sci
2006
;
97
:
642
–8.
19
Oshita F, Matsukuma S, Yoshihara M, et al. Novel heteroduplex method using small cytology specimens with a remarkably high success rate for analysing EGFR gene mutations with a significant correlation to gefitinib efficacy in non-small-cell lung cancer.
Br J Cancer
2006
;
95
:
1070
–5.
20
Cohen V, Agulnik JS, Jarry J, et al. Evaluation of denaturing high-performance liquid chromatography as a rapid detection method for identification of epidermal growth factor receptor mutations in nonsmall-cell lung cancer.
Cancer
2006
;
107
:
2858
–65.
21
Asano H, Toyooka S, Tokumo M, et al. Detection of EGFR gene mutation in lung cancer by mutant-enriched polymerase chain reaction assay.
Clin Cancer Res
2006
;
12
:
43
–8.
22
Hoshi K, Takakura H, Mitani Y, et al. Rapid detection of epidermal growth factor receptor mutations in lung cancer by the SMart-Amplification Process.
Clin Cancer Res
2007
;
13
:
4974
–83.
23
Wittwer CT, Reed GH, Gundry CN, Vandersteen JG, Pryor RJ. High-resolution genotyping by amplicon melting analysis using LCGreen.
Clin Chem
2003
;
49
:
853
–60.
24
Emmert-Buck MR, Bonner RF, Smith PD, et al. Laser capture microdissection.
Science
1996
;
274
:
998
–1001.
25
Ronaghi M. Pyrosequencing sheds light on DNA sequencing.
Genome Res
2001
;
11
:
3
–11.
26
Noguchi M, Furuya S, Takeuchi T, Hirohashi S. Modified formalin and methanol fixation methods for molecular biological and morphological analyses.
Pathol Int
1997
;
47
:
685
–91.
27
Taillade L, Penault-Llorca F, Boulet T, et al. Immunohistochemichal expression of biomarkers: a comparative study between diagnostic bronchial biopsies and surgical specimens of non-small-cell lung cancer.
Ann Oncol
2007
;
18
:
1043
–50.
28
Ruffini E, Rena O, Oliaro A, et al. Lung tumors with mixed histologic pattern. Clinico-pathologic characteristics and prognostic significance.
Eur J Cardiothorac Surg
2002
;
22
:
701
–7.
29
Gonzalez-Garcia I, Sole RV, Costa J. Metapopulation dynamics and spatial heterogeneity in cancer.
Proc Natl Acad Sci U S A
2002
;
99
:
13085
–9.
30
Carey FA, Lamb D, Bird CC. Intratumoral heterogeneity of DNA content in lung cancer.
Cancer
1990
;
65
:
2266
–9.