Comparison of the Clinical Validity of Droplet Digital PCR to ARMS‐PCR for BRAF V600E Mutation Detection in Thyroid Nodules

Abstract Objectives Droplet digital PCR (ddPCR) has been reported to have a superior validity over PCR with amplification‐refractory mutation system (ARMS‐PCR) for detecting the BRAF V600E mutation in thyroid nodule fine‐needle aspiration (FNA) samples using cytological diagnosis as the reference. However, the added value of ddPCR on surgical decision‐making remains to be illustrated when the technique is combined with FNA cytology. Methods A total of 277 consecutive patients with thyroid nodules were subjected to FNA cytology and BRAF V600E testing with ARMS‐PCR. Within this patient cohort, 90 patients underwent surgical intervention with pathological diagnosis available. BRAF V600E testing with ddPCR was performed retrospectively using FNA frozen DNA specimens. The clinical validity and utility of ddPCR in comparison with ARMS‐PCR were compared using surgical pathology as the reference. Results Overall, 101 BRAF V600E mutations were detected by ddPCR, including five ARMS negative patients, four of whom were confirmed to have papillary thyroid cancer (PTC) by surgical pathology. Of the 90 patients with surgical pathology, which is considered the gold standard, ddPCR BRAF V600E testing yielded a sensitivity of 91.3% and specificity of 100% for PTC diagnosis, higher than that of ARMS (sensitivity 83.1%, specificity 100%). However, ddPCR only identified one more candidate patient for surgical intervention than ARMS when the techniques were combined with cytology. Conclusions This study highlighted the superior performance of ddPCR over ARMS in BRAF V600E detection from thyroid nodule FNA samples. Further studies are needed to evaluate the cost‐effectiveness of replacing ARMS‐PCR with ddPCR for surgical decision‐making.


| INTRODUC TI ON
Thyroid nodules are common, with an incident rate as high as 50%-70% in the adult population and are especially prevalent in women. The majority of thyroid nodules are benign, yet a small proportion become cancerous. 1 Currently, ultrasound-guided fine-needle aspirate (FNA) cytopathology is the major method for the diagnosis of cancerous thyroid nodules. 2 Unfortunately, the accuracy of FNA cytopathology remains unsatisfactory, with one-third of cases categorized as being diagnostic challenging. 3 Therefore, many patients undertook unnecessary surgery or experienced a false-negative diagnosis due to inaccurate cytological testing.
Papillary thyroid cancer (PTC) accounts for over 80% of all thyroid cancers and has shown a dramatic increase in prevalence in recent years. 4,5 The BRAF V600E mutation occurs in 50%-89% of PTC cases and serves as an important diagnostic and prognostic biomarker. [6][7][8][9][10] The incorporation of BRAF V600E testing has been shown to substantially improve the diagnostic accuracy of FNA. 11,12 One of the most widely used methods for BRAF V600E detection in thyroid nodules is PCR with amplification-refractory mutation system (ARMS); this is a well-established technique that has been widely used for rapid detection of nucleic acid mutations in a wide variety of biological samples. However, ARMS-PCR may not be sensitive enough due to the fact that FNA samples usually have few mutant cells. 13 Therefore, the development of a more sensitive and accurate detection method is warranted.
Droplet digital PCR (ddPCR) is a novel technology characterized by high sensitivity and absolute quantification of nucleic acid targets. 13,14 The superior sensitivity renders ddPCR as a promising detection technique in samples with trace amounts of nucleic acids, such as in liquid biopsy. 15 The Bio-rad QX200 TM ddPCR was shown to improve diagnostic accuracy for detecting the BRAF V600E mutation by 17% compared to cytopathology alone in thyroid nodules FNA samples. 12 A previous study has also reported the superior sensitivity of ddPCR over ARMS-PCR in PTC detection using the FNA cytopathology as the reference. 13 However, as aforementioned, FNA cytological diagnosis is not accurate as surgical pathology, which is considered to be the gold standard; it remains unclear whether ddPCR has a better value than ARMS-PCR on surgical decision-making when in combination with FNA cytology. Therefore, in this study, we validated and compared the clinical utility of ddPCR to ARMS-PCR for the detection of the BRAF V600E mutation in FNA specimens of thyroid nodules, using pathological diagnosis following surgery as the gold standard.

| Study subjects and FNA samples collection
This is a retrospective cohort study involved consecutive patients who underwent FNA cytology and BRAF V600E testing with ARMS for thyroid nodules in The Guangzhou First People's Hospital between October 2018 and July 2019. Based on the thyroid ultrasound findings, high-risk individuals were referred to ultrasound-guided FNA biopsy performed by trained physicians according to the recommended guideline. 16 In addition to cytopathological examination, FNA specimens were also collected for BRAF V600E mutation testing. The pathologists who made the cytopathological diagnosis were blind to the ARMS-PCR and ddPCR results. The study was approved by the institutional ethical review board at The Guangzhou First People's Hospital, and patients gave informed consent.

| Nucleic acid extraction from FNA samples
The Tissue DNA Extration Kit (AmoyDx, China) was used for DNA extraction from thyroid nodule FNA samples. After collection, FNA specimens were immediately transferred to a 1.5 mL microcentrifuge tube with 180 μL lysis buffer and DNA was extracted according to the manufacture's protocol. OD 260/280 was used to measure the DNA concentration.

| Plasmid preparation
The wild-type BRAF plasmid and the V600E mutant plasmid were first synthesized and then had their sequences confirmed by Generay Biotech Co., Ltd (Shanghai). Mutant and wild-type plasmids were prepared as the positive and negative controls. The QIAamp DNA mini kit (QIAGEN) was used for DNA extraction from plasmids. Qubit 4.0 (Thermo Fisher) was used to determine the concentration of plasmid DNA. Mutant plasmid DNA was diluted with wild-type plasmid DNA in a series of concentrations (0.05%, 0.1%, 0.5%, 1%, 5%, 10%, and 50%) to determine the sensitivity of ddPCR for BRAF V600E detection.

| BRAF V600E mutation detection with ARMS-PCR and ddPCR
The ARMS-PCR assay was performed on an ABI 7500 Real-time PCR system (Life Technologist) with a BRAF V600E Diagnostic Kit (AmoyDx) according to the manufacturer's protocol. The details of this testing system have been published elsewhere. 13 The MicroDrop-100 TM ddPCR system (Forevergen), which is based on water-emulsion droplet technology, was used for the ddPCR assay. The primers and probes used for BRAF V600E detection with ddPCR are as follows: The ARMS-PCR assay and the ddPCR assay were performed separately by two researchers to avoid mutual influences.

| Statistical analysis
Analysis was performed using the SPSS 20.0 software (IBM, USA).
We used chi-square tests for comparisons between groups for categorical variables and two-tailed t tests for continuous variables. A P value < 0.05 was considered statistically significant.

| Cytological findings of FNA specimens
A total of 277 patients with both FNA cytology and ARMS-PCR BRAF V600E testing available were enrolled in this study. The mean age (±standard deviation) was 44.9 ± 13.2 years (range: 13-80), and 79.1% of the cohort (n = 219) was female. The cytopathology of FNA indicated that 86 cases (31.0%) were classified as malignant tumor (category VI) and 18 (6.5%) cases were suspicious for malignancy (category V), while 18 (6.5%) cases were classified as category III/IV with diagnostic challenge, and 43 (15.5%) cases were nondiagnostic (category I). The detailed description of the Bethesda classification is shown in Table 1.

| ddPCR vs. ARMS-PCR for BRAF V600E mutation detection
Using ARMS, we detected 96 (34.7%) cases with the BRAF V600E mutation, including 14 cases classified as Bethesda I-III, and 82 cases classified as Bethesda V-VI. Compared to ARMS, ddPCR detected five more mutant cases, of whom four cases were subjected to surgical intervention and PTC was confirmed by surgical pathological diagnosis, suggesting ddPCR has a higher sensitivity than ARMS on FNA BRAF V600E detection (Tables 2-3). Additionally, ddPCR detected the absolute quantity of BRAF V600E copies, which ranged from 6 to 5720 (0.05% to 43.4%) in the 20ul reaction system.

| FNA cytopathology, ARMS, and ddPCR vs. surgical pathology in a subgroup of patients
Of the 277 participants, 90 high-risk cases who had a positive finding on ARMS-PCR BRAF V600E and/or were classified as Bethesda category V/VI undertook surgical intervention. However, 35 highrisk cases with positive ARMS findings and/or Bethesda categories V/VI were lost to follow-up for surgical intervention. Nonetheless, the demographic and clinical characteristics were similar in the highrisk patients with or without surgical intervention (Table 1).

| D ISCUSS I ON
The BRAF V600E mutation testing on FNA specimens from thyroid nodules greatly improves diagnostic accuracy. 11,12 However, detecting trace amounts of DNA molecules in FNA specimens remains challenging. DdPCR is a cutting-edge technique that enables the sensitive and accurate detection of molecular markers from samples with a limited amount of target DNA. 18 In this study, we confirmed the clinical validity of the MicroDrop-100 TM ddPCR system.
It showed a better performance than ARMS-PCR in BRAF V600E mutation detection on thyroid nodule FNA specimens, which is in line with a previous study. 13 The combination of ddPCR and FNA cytology may serve as a better option for the diagnosis of thyroid nodules.
In this study, ddPCR detected five more mutant specimens than ARMS-PCR, three of which displayed mutation rates of less than 0.1%, illustrating the ultra-sensitivity of ddPCR. In support of this  Note: Bethesda classification: I, specimens nondiagnostic/unsatisfactory (ND/UNS); II, benign; III, atypia of undetermined significance/follicular lesion of undetermined significance (AUS/FLUS); IV, follicular neoplasm/suspicious for a follicular neoplasia (FN/SFN); V, suspicious for malignancy (SM); VI, malignancy.

ACK N OWLED G M ENT
We deeply acknowledge the contributions of physicians, nurses, and patients in this study.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

AUTH O R CO NTR I B UTI O N S
X.L, S.X, and Y.Z conceived the study and its design. X.L, H.D, and W.D were involved in the sample collection, reviewed, and edited the study. X.L, J.L, and J.H contributed to the testing, data analysis, and reviewed the study. X.L and Y.Z wrote the study. S.Z and YZ contributed to the discussion, and reviewed, edited, and finalized the study. All the authors read and approved the final study.