Detection and comparison of EGFR mutations from supernatants that contain cell-free DNA and cell pellets from FNA non–small cell lung cancer specimens
Abstract
Background
Epidermal growth factor receptor (EGFR) is an important marker for targeted therapy in patients with advanced non–small cell lung cancer (NSCLC). The samples obtained with minimally invasive biopsy techniques are usually small, and this limits their application in tissue subtyping or molecular profiling. The supernatants obtained after centrifugation of fine-needle aspiration (FNA) samples are typically discarded. However, these fractions might contain cell-free DNA that could be tested for EGFR mutations by genotyping methods that are normally used for plasma analysis.
Methods
In this study, 214 patients with known or suspected NSCLC who underwent FNA were enrolled. The workflow of the supernatants before molecular detection was as follows. The discarded FNA samples (15 mL) were stored in CytoLyt, a cleaning, fixation solution, and 10 mL of each sample was placed in a preservation solution for separation by low-speed centrifugation. The primary supernatants (8 mL) were then separated by high-speed centrifugation to obtain secondary supernatants. DNA was extracted from the supernatants with QIAamp circulating nucleic acid kits (Qiagen) and circulating DNA kits (AmoyDx), and EGFR mutations were assessed with Super-ARMS EGFR detection kits (AmoyDx). The DNA was then extracted from corresponding cell pellets with tissue DNA kits (AmoyDx), and the EGFR status was analyzed with the amplification refractory mutation system and next-generation sequencing methods.
Results
All 214 samples yielded an adequate amount of cell-free DNA for EGFR detection. The use of different DNA commercial extraction kits and the DNA contents of tumor cells did not affect the yield of DNA from the supernatants. The external controlled cycle threshold value of the EGFR test was affected by the concentration of the DNA in the supernatants (P < .05). However, the difference in the concentrations of the DNA in the supernatants did not affect the EGFR mutation status. The EGFR-positive rate was 57.5% (123 of 214) in both the supernatants and the pellets from the 214 FNA samples. The concordance between EGFR variants in the supernatants and the corresponding pellets was 97.2%. EGFR mutations were also detected in 3 pellets but not in their corresponding supernatants and in 3 supernatants but not in their corresponding pellets. The supernatants of FNA biopsy samples might represent a new source for gaining information regarding the molecular characteristics of patients for targeted therapy.
Conclusions
Discarded supernatants provided an adequate amount of cell-free DNA for EGFR detection, and this means that the pellets can be reserved for additional morphological and molecular analyses or to avoid repeat biopsies. Analyzing the EGFR status in cell supernatants and pellets might improve detection sensitivity and confer benefits to patients.
Introduction
Lung cancer is the most prevalent cause of cancer-related deaths worldwide among men and is the second-most prevalent cause among women.1 Patients with non–small cell lung cancer (NSCLC) with specific genomic aberrations benefit from molecular targeted therapies.2 Epidermal growth factor receptor (EGFR) mutations are important markers for guiding treatment options.3 The incidence of EGFR mutations in NSCLC tumors among East Asian populations reaches approximately 50%, and 95% of such mutations have been identified in adenocarcinomas.4, 5 Thus, EGFR mutation tests are essential for patients with NSCLC.
The current molecular testing guidelines consider cell blocks or other cytological preparations as suitable specimens for lung cancer biomarker molecular testing except for tissue samples.6 Thus, cytology samples from fine-needle aspiration (FNA) play an important role in the classification of NSCLC.7 The sample quality and quantity from certain needle biopsies are very limited in terms of their further use for tissue subtyping or molecular profiling. Therefore, finding ways to optimize these samples in cytopathology laboratories is important for patient management. During the routine preparation of liquid-based cytology (LBC) samples derived from FNA biopsies, the supernatant is typically discarded after centrifugation. However, we hypothesize that it might contain a sufficient amount of cell-free DNA for characterizing the molecular changes in the whole sample. Adequate quantities of amplifiable cell-free DNA are extracted from the supernatants obtained from solid tumor FNA samples for gene testing using sensitive assays.8-11 This process has the potential to save patient samples by allowing the cell pellets to be reserved for additional morphological and molecular analyses. Super-ARMS is a modification of the amplification refractory mutation system (ARMS), which is a highly sensitive and specific approach for detecting free plasma DNA and is widely applied to routine gene testing. However, whether Super-ARMS is useful for detecting DNA in FNA samples remains unknown.
Therefore, in this study, we aimed to determine the feasibility of detecting EGFR mutations in discarded supernatants from FNA biopsy specimens of NSCLC with CytoLyt (Hologic Inc) and Super-ARMS and to compare the usefulness of such supernatants and their corresponding cell pellets.
Materials and Methods
Sample Collection and Processing
We collected the cell supernatants and corresponding cell pellets from 214 FNA cytology samples of NSCLC between December 2018 and June 2019 at the Shanghai Pulmonary Hospital under institutional review board approval. The samples were obtained by computed tomography (CT)–FNA or endoscopic transbronchial ultrasound–guided FNA using 20- to 25-gauge needles. Figure 1 shows how the tissues from the needle aspirates were processed. Briefly, direct smears were immediately prepared at the time of collection and were stained with hematoxylin-eosin for rapid onsite evaluation, whereas the residual tissue in the needle was rinsed into 15 mL of CytoLyt (Hologic Inc), which is a methanol-based fixative solution. When the cytopathologists found that the cells in the direct smears might be malignant, the corresponding cellular material was initially separated by centrifugation at 3000 rpm for 5 minutes. The resulting supernatants (the first supernatants; 10 mL) were transferred into 15-mL centrifuge tubes and were stored at –20 °C until DNA extraction, and the remaining 5 mL of the cell pellets was processed according to the LBC workflow. After further diagnosis of adenocarcinoma, squamous cell carcinoma, or NSCLC by the LBC smear, cell block, or immunocytochemistry, 10 mL from the upper liquid was recentrifuged at 10,000 rpm for 10 minutes and was divided into 8 mL of supernatants for cell-free DNA extraction and a 2-mL solution with the pellets for tissue DNA extraction.
DNA Extraction
We extracted DNA from the second supernatants (8 mL each) of 144 samples with the QIAamp circulating nucleic acid kit (Qiagen) according to the manufacturer's protocol. We also extracted DNA from another 70 supernatants with circulating DNA kits (AmoyDx). The volume of DNA eluted from the supernatants was 80 µL. Tissue DNA kits (AmoyDx) were then used to extract the DNA from the cell pellets and stained smears. The DNA concentrations were measured with the Qubit DNA high-sensitivity assay kit (Thermo Fisher Scientific, Waltham, Massachusetts).
EGFR Mutation Detection
The status of the EGFR mutations in the supernatants was analyzed with Super-ARMS according to the manufacturer's instructions. The samples with DNA concentrations ≥ 1 ng/µL were diluted to 1 ng/µL for the subsequent steps. When the sample concentrations were <1 ng/µL, the original concentration was used in the subsequent steps. The Super-ARMS assay was conducted with the SLAN-96S real-time polymerase chain reaction system (Shanghai Hongshi Medical Technology Co, Ltd, Shanghai, China). The detection sensitivity of Super-ARMS was 0.5%. The EGFR mutation status of pellets was analyzed with the ARMS method according to the manufacturer's instructions. The ARMS assay was conducted on the Mx3000P system (Agilent Technologies, Santa Clara, California). The detection sensitivity of ARMS was 1%.
Next-Generation Sequencing Detection
DNA sequencing was performed with a capture-based sequencing panel (Amoy Diagnostics Co, Ltd, Xiamen, China). The kit covered the targeted drug-related hot-spot mutation regions of 10 genes, including EGFR. The DNA from the stained smears was sequenced with a NextSeq 500 (Illumina, Inc) with pair-end reads with a reading length of PE150 and a sequencing depth of >10,000×. The sequence data were analyzed with the AmoyDx next-generation sequencing (NGS) data analysis system ADXLC10 module (Amoy Diagnostics).
Data and Statistical Analysis
The differences in the yields obtained from the supernatants with the 2 circulating DNA extraction methods were compared with Mann-Whitney U tests. The rate of a consistent EGFR mutation status between the supernatants and pellets from the FNA specimens was compared with Fisher's exact test, and the frequency of the EGFR status, derived from OriginLab 2019b, was displayed with Sankey plots. All the data were statistically analyzed with IBM-SPSS V25.0 software. P < .05 was considered statistically significant.
Results
Characteristics of the Patients and Samples
Table 1 summarizes the demographic and pathological characteristics of the 214 enrolled patients (120 males and 94 females; median age, 63 years; age range, 38-90 years). In total, the diagnoses were adenocarcinoma (n = 184), squamous cell carcinoma (n = 9), and NSCLC not otherwise specified (n = 21). We obtained 176 and 38 tumor samples from the lungs and lymph nodes of the patients, respectively. The TNM stages of these tumors were I (n = 16 [7.5%]), II (n = 10 [4.7%]), III (n = 28 [13.1%]), and IV (n = 160 [74.7%]). Two senior cytopathologists diagnosed all the specimens.
| Factor | Value |
|---|---|
| Sex, No. (%) | |
| Male | 120 (56.1) |
| Female | 94 (43.9) |
| Age, median, y | 63 |
| Age, range, y | 38-90 |
| Pathology, No. (%) | |
| Adenocarcinoma | 184 (86.0) |
| Squamous cell | 9 (4.2) |
| NSCLC NOS | 21 (9.8) |
| Tumor stage, No. (%) | |
| I | 16 (7.5) |
| II | 10 (4.7) |
| III | 28 (13.1) |
| IV | 160 (74.7) |
| Cytological samples, No. (%) | |
| FNA (lungs) | 176 (82.2) |
| FNA (lymph nodes) | 38 (17.8) |
- Abbreviations: FNA, fine-needle aspiration; NOS, not otherwise specified; NSCLC, non–small cell lung cancer.
Supernatant DNA Yield
The DNA concentrations from the supernatants were measured with Qubit. We initially extracted DNA from the supernatants and pellets to evaluate the differences between these materials. The yield of DNA was higher in the pellets in comparison with that of the cell-free DNA in the supernatants. Table 2 shows the median concentrations of DNA extracted from the supernatants with the QIAamp circulating nucleic acid kit (Qiagen) and a circulating DNA kit (AmoyDx). The DNA concentrations were 0.77 ng/µL (range, low to 120 ng/µL) and 0.48 ng/µL (range, low to 120 ng/µL), respectively, and they did not significantly differ (P = .1). The numbers of samples with low DNA concentrations extracted with the AmoyDx and Qiagen kits were 10 and 2, respectively.
| QIAamp Circulating Nucleic Acid Kit (Qiagen) | Circulating DNA Kit (AmoyDx) | P | |
|---|---|---|---|
| Samples, no. | 144 | 70 | |
| DNA concentration, median, ng/µL | 0.77 | 0.48 | .1 |
| DNA concentration, range, ng/µLa | <0.05 to 120 | <0.05 to 120 | |
| Samples with low concentration, No. | 2 | 10 | |
| EGFR mutation rate, % (n/N) | 60.4 (87/144) | 51.4 (36/70) | .212 |
- Abbreviation: EGFR, epidermal growth factor receptor.
- a The lower limit of detection was 0.05 ng/µL.
The EGFR gene could be detected with Super-ARMS from the DNA obtained with the Qiagen kits. Table 3 shows the relationship between the external control CT values and the DNA concentrations in the supernatants. The external control CT values significantly differed and were 0.5<C≤1 ng/µL group and C≤0.5 ng/µL group (P < .05). However, the difference did not affect the EGFR mutation rate (60.2% vs 54.5%; P > .05).
| Concentration Range | External Control Ct Value, Median | P | EGFR Mutation Rate, % (n/N) | P |
|---|---|---|---|---|
| 0.5<C≤1 ng/µL (n = 113) | 12.26 | .001 | 60.2 (68/113) | .398 |
| C≤0.5 ng/µL (n = 101) | 15.415 | 54.5 (55/101) |
- Abbreviations: Ct, cycle threshold; EGFR, epidermal growth factor receptor.
The relationship between the yield from the supernatant DNA and the content of tumor cells was further analyzed. The samples were divided according to the median tumor cell concentrations as follows: 0.591 (≤200), 0.798 (≥500), and 0.886 ng/µL (200-500). The DNA yield did not significantly differ among the 3 groups (P > .05; Table 4).
| Tumor Cells (n) | No. of Cases | DNA Concentration, Median, ng/μL | P |
|---|---|---|---|
| n≥500 group | 99 | 0.798 | .9498 |
| 200<n<500 group | 54 | 0.886 | |
| n≤200 group | 61 | 0.591 |
EGFR Mutation Analysis
The EGFR-positive rate was 57.5% (123 of 214) for both the supernatants and the pellets from the 214 FNA samples. The EGFR-positive rate for 144 samples with the Qiagen kit was 60.4% (87 of 144), and that for 70 samples with the AmoyDx kit was 51.4% (36 of 70). There was no difference between the 2 extraction kits for EGFR detection (P > .05; Table 2). Table 5 shows a comparison of the EGFR mutations detected in the supernatants and pellets. The consistency rate of the two was 97.2% (208 of 214; 95% CI, 95.0%-99.4%). The consistency of the mutation status detected in the supernatants and pellets from the same sample was excellent (Fig. 2). Thus, the EGFR mutation status in 6 nonconcordant cases from the supernatants and pellets is shown in Table 6. Mutations in EGFR were detected in pellets, and wild-type mutations were detected in supernatants from 3 samples. Notably, EGFR mutations were also detected in supernatants from 3 samples, whereas wild-type mutations were also detected in their corresponding pellets. The EGFR mutation status of the pellets from these inconsistent samples was verified with Super-ARMS, and the results were consistent with those obtained with the ARMS method. At the same time, we further validated the EGFR mutation status in 6 nonconcordant cases from the supernatants and pellets by using the NGS method in the corresponding smears. There were no additional hematoxylin-eosin smears for NGS in 2 cases, and the EGFR mutation status in 4 cases was consistent with the corresponding supernatants and pellets. Interestingly, 1 patient with a G719X mutation, as detected in the supernatant, was under erlotinib-targeted therapy, and the effects were assessed as a partial response until this article was submitted.
| Pellets | Total | ||
|---|---|---|---|
| Mutated | Wild Type | ||
| Supernatants, No. | |||
| Mutated | 120 | 3 | 123 |
| Wild type | 3 | 88 | 91 |
| Total | 123 | 91 | 214 |
| PCR | 97.6% (120/123); 95% CI, 94.8%-100.0% | ||
| NCR | 96.7% (88/91); 95% CI, 93.0%-100.0% | ||
| Overall agreement | 97.2% (208/214); 95% CI, 95.0%-99.4% | ||
- Abbreviations: CI, confidence interval; EGFR, epidermal growth factor receptor; NCR, negative coincidence rate; PCR, positive coincidence rate.
| Case No. | Supernatants (Super-ARMS) | Pellets (ARMS) | Pellets (Super-ARMS) | Smears (NGS) |
|---|---|---|---|---|
| 6 | Wild type | L858R | L858R | Exon21 c.2573T>G p.L858R |
| 36 | G719X | Wild type | Wild type | Exon18 c.2155G>A p.G719S |
| 41 | 20ins | Wild type | Wild type | Exon20 c.2316_2317ins9 p.P772_H773insYNP |
| 72 | Wild type | 19del+T790M | 19del+T790M | NA |
| 155 | L858R | Wild type | Wild type | NA |
| 175 | Wild type | 19del | 19del | Exon19 c.2235_2249delGGAATTAAGAGAAGC p.E746_A750del |
- Abbreviations: ARMS, amplification refractory mutation system; EGFR, epidermal growth factor receptor; NA, not available; NGS, next-generation sequencing.
Discussion
Liquid biopsy is widely used in the clinical diagnosis and treatment of cancer. Tumor fluid biopsy mainly includes the molecular detection of free nucleic acids from body fluids, such as blood, pleural fluid, ascites, urine, and cerebrospinal fluid.12-17 After nearly 20 years of clinical practice, liquid biopsies now provide valuable information for the diagnosis of various tumors. In terms of a molecular diagnosis for NSCLC, the mutation rate of EGFR has been highly consistent for both liquid biopsies and paired histological samples (70%-90%).18-20 Liquid biopsy samples are collected noninvasively, and molecular testing guidelines recommend using these samples as supplementary diagnostic material. Discarded supernatants from FNA samples are a specific type of tumor-associated liquid. Finkelstein et al9 first proposed and verified the hypothesis that supernatants derived from cytobrushed specimens of obstructed pancreatic or bile ducts adjacent to cancer would contain cell-free DNA. Several recent studies have revealed that supernatants from patients with NSCLC contain enough tumor DNA to serve as complementary material for gene testing using highly sensitivity methods, such as digital polymerase chain reaction and NGS.21, 22 The concordance of gene variants between supernatants and their corresponding pellets has been high overall. However, few studies have investigated this topic, and the data as well as the samples are insufficient. Therefore, such samples appeal more to research that focuses on evaluating the feasibility of supernatants as molecular diagnostic materials. The current study found that cell-free DNA supernatants from FNA samples could be extracted and used to determine the EGFR gene status. We detected and compared the EGFR gene status with the free DNA from discarded supernatants and the corresponding cell pellets from FNA samples. The consistency rate of the findings was 97.2%.
Many factors affect the cell-free DNA found in the bloodstream of patients, including diet, time of blood collection, selection of blood collection tubes, transportation, preservation, and DNA extraction kits used to analyze the collected samples.23-28 Cell-free DNA is often extracted from supernatants with QIAamp circulating nucleic acid kits. Here, we compared the amount of cell-free DNA extracted with QIAamp (Qiagen) and circulating DNA (AmoyDx) kits. The results indicated that the Qiagen kit performed slightly better than the AmoyDx kit. However, the amount of cell-free DNA extracted by both kits was sufficient to detect EGFR. Janaki et al21 found no correlation between the total yield of cell-free DNA and how long cytology samples are preserved. We further investigated whether or not the amount of supernatant DNA was associated with the tumor cell content of the samples. Our results showed that the content of the tumor cells did not affect the amount of extracted free DNA. In other studies, needle aspiration samples were treated with 10 mL of Roswell Park Memorial Institute 1640 medium or normal saline.29, 30 In our study, we used CytoLyt solution, which provided cell protection and prevented DNA degradation. To make full use of the supernatant samples, more studies are needed to verify whether the amount of cell-free DNA is affected by different treatment solutions or volumes or by the time and temperature at which the cell-free DNA is released.
Previous studies have assayed EGFR mutations with supernatants derived from NSCLC cytology samples from small patient cohorts.31, 32 Here, we extracted cell-free DNA from waste supernatants and determined the EGFR mutation status by using Super-ARMS in samples from 214 patients, and we found that the results obtained with the supernatants and cell pellets were consistent. Some investigators have simultaneously assessed the EGFR gene status in the cell pellets and supernatants derived from pleural and bronchial lavage fluids and found that the detection sensitivity increases and that mutations are detectable in the supernatants even when tumor cells are not present.33, 34 Our data support these results. The reasons that more EGFR mutation sites can be detected in the supernatant are as follows: the cell-free DNA in the supernatant is homogeneous, and nucleic acid extraction from cell pellets requires only a very small portion of the precipitate, which can include no or a few mutated tumor cells.
In conclusion, FNA is a well-established technique for the diagnosis of NSCLC. The discarded supernatants from FNA preparations provide a source of cell-free DNA that can be used for EGFR detection. Super-ARMS would be useful for detecting cell-free DNA in discarded supernatants from FNA samples. Thus, the precious cell pellets can be preserved and used to detect other genetic alterations and immune-related indicators. The application of discarded supernatants for molecular detection has changed the workflow of cytology samples. Therefore, previously discarded supernatants generated from FNA samples offer a new approach to molecular detection.
Funding Support
This study was sponsored by the Science and Technology Commission of Shanghai Municipality (18411962900).
Conflict of Interest Disclosures
The authors made no disclosures.
Author Contributions
Wei Wu: Study design, performance of the experiments, writing–original draft, and approval of the final manuscript. Yan Huang: Study design, performance of the experiments, writing–editing and review, and approval of the final manuscript. Junhong Guo: Analysis and interpretation of the patient data and approval of the final manuscript. Xiaofeng Xie: Analysis and interpretation of the patient data and approval of the final manuscript. Hui Li: Performance of the experiments and approval of the final manuscript. Ziyang Cao: Performance of the experiments and approval of the final manuscript. Haiting Wei: Analysis and interpretation of the patient data and approval of the final manuscript. Chunyan Wu: Study design, writing–editing and review, and approval of the final manuscript.
Open Research
Data Availability
The raw data are available upon request.
