Evaluation of a Two-Year Routine Application of Molecular Testing of Thyroid Fine-Needle Aspirations Using a Seven-Gene Panel in a Primary Referral Setting in Germany

Patients and samples

Within a time frame of two years (June 2012–June 2014), 9460 thyroid FNAs were cytologically graded by an experienced thyroid cytopathologist (I.R.) according to the TBSRTC guideline (17). At variance with other data mostly originating from tertiary centers, these data demonstrate low percentages for malignant (MAL), suspicious for malignancy (SFM), FN/SFN, and AUS/FLUS cases, and high percentages for cysts/cystic nodules in this primary care setting in Germany. Of the 9460 FNAs, mostly performed with ultrasound (US) guidance by 64 different physicians, 564 were selected for molecular diagnostics according to the following criteria: FNAs with a AUS/FLUS cytology if additional signs of follicular proliferation, including suspicion for a follicular variant of papillary thyroid carcinoma (fvPTC), were present or if the AUS/FLUS diagnosis was based on a low number of cells, unless molecular diagnostics was not ordered by the referring physician (n = 195); most cytologies with the diagnosis FN/SFN, unless molecular diagnostics was not ordered by the referring physician (n = 278); cytologies with the diagnosis SFM if ordered by the referring physician (n = 55); and cytologies with a MAL diagnosis if ordered by the referring physician, except for those with cytological evidence for anaplastic thyroid carcinoma or a primary tumor outside the thyroid (n = 36). Current guidelines do not recommend molecular diagnosis for SFM and MAL cytologies. For the 564 patients, the standardized collection and analysis of the following data was performed: thyrotropin (TSH), thyroid autoantibody values, calcitonin, thyroid ultrasound, nodule size, sonographic malignancy criteria, Tc scintiscan results, recommendation for surgery by the treating physician, surgery report, histology and/or follow-up ultrasound. The median age of the 564 patients was 55 years (range 13–88 years), and the median size of the nodules was 2.4 cm (range 0.6–13 cm). No histologic or follow-up data could be obtained for 142/564 patients (Fig. 1). While 348/564 patients with molecular diagnostics underwent thyroid surgery, 74 patients had follow-up. For 322/348 patients with thyroid surgery, histology results could be obtained from the local routine pathologists who were blinded to the molecular test results. For 33/74 patients who had follow-up, follow-up data of at least one year could be obtained revealing no evidence of malignancy.

OPEN IN VIEWER

FIG. 1.

Composition of the set of thyroid nodules receiving presurgical molecular fine-needle aspiration (FNA) diagnosis with regard to thyroid surgery, follow-up, and available histology and follow-up data, respectively.

Because mutation testing on FNA samples is the standard of care in the routine diagnostic setting, the molecular analysis did not require approval of the ethics committee of the University of Leipzig. The retrospective analysis of cytological, histological, clinical, and molecular data was approved by the ethics committee of the University of Leipzig.

DNA and RNA extraction from FNA smears

DNA and RNA were extracted from the routine air-dried FNA smears using the AllPrep FFPE Kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol. In brief, the FNA smears were incubated in xylene for four to five days to remove the cover slips. Thereafter, 150 μL of PKD buffer was added to the slide, and the cells on the slide were detached within the PKD buffer using a cell scraper. The detached cells were transferred to a new tube, and DNA and RNA were extracted according to the manufacturer's protocol.

cDNA synthesis

cDNA was synthesized using the miScript Reverse Transcription Kit (Qiagen) according to the manufacturer's protocol. In brief, 7.5 μL of template RNA was added to a master mix consisting of 2 μL of 5 × miScript RT buffer and 0.5 μL of miScript Reverse Transcriptase Mix and incubated for 60 min at 37°C. Subsequently, the miScript Reverse Transcriptase Mix was inactivated for 5 min at 95°C.

Detection of rearrangements by quantitative polymerase chain reaction

PAX8/PPARG, RET/PTC1, and RET/PTC3 rearrangements were detected in multiplex-quantitative polymerase chain reactions (PCR) on a LightCycler 480 (Roche, Mannheim, Germany) using the QuantiTect SYBR Green PCR kit (Qiagen), with one primer pair amplifying an intron spanning 125 bp control fragment of PAX8 mRNA (exon 5–6; for cDNA quality control), and one rearrangement specific primer pair, as described previously (14,15). Following the PCR, a fluorescence melting curve analysis was performed to detect and discriminate the control fragment and the rearrangement specific amplicons by their specific melting points (14).

Detection of point mutations by pyrosequencing

BRAF (codons 600/601), KRAS (codons 12/13), HRAS (codon 61), and NRAS (codon 61) point mutations were detected by pyrosequencing. In brief, 3 μL of genomic DNA was amplified using self-designed template specific PCR primers (Supplementary Table S1; Supplementary Data are available online at www.liebertpub.com/thy) (14,15), the PyroMark PCR kit (Qiagen), and template-specific PCR conditions (Supplementary Table S2 (14,15)) according to the manufacturer's protocol. Afterwards, the PCR products were immobilized to streptavidin sepharose beads, and single-stranded DNA was prepared allowing subsequent annealing of the sequencing primer to the template DNA. Then, the primed single-stranded DNA was released from the streptavidin surface and transferred to a PyroMark Q24 (Qiagen) for pyrosequencing. Based on a recent retrospective study comparing the allelic percentage in FNA samples and corresponding histology samples (15), a cutoff value of 10% for the allelic percentage of point mutations was applied.

Histologic re-evaluation of mutation-positive histological benign samples

Mutation-positive nodules, which were surgically removed and were histologically diagnosed as adenomatous nodule (AN), follicular adenoma (FA), or oncocytic subtype of FA (oFA) by the local pathologists, were, in case of available blocks, re-reviewed according to the World Health Organization criteria (18) and Piana et al. (19) by an expert referral pathologist (U.S.), who was aware of the molecular testing results. Statistical analysis was based on these revised histological diagnoses. While one follicular thyroid carcinoma (FTC) case was considered as malignant, follicular tumors of uncertain malignant potential (FT-UMP) were considered as benign. The histologic re-evaluation did not significantly influence the statistical outcome of the data (p = 1).

Statistics

Sensitivities, specificities, PPV, and negative predictive values (NPV) were calculated as follows: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}{ \rm { sensitivity } } = { \frac { { \rm { true \ positives } } } { { \rm { true \ positives } } + { \rm { false \ negatives } } } } \end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}{ \rm { specificity } } = { \frac { { \rm { true \ negatives } } } { { \rm { true \ negatives } } + { \rm { false \ positives } } } } \end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}{ \rm { PPV } } = { \frac { { \rm { true \ positives } } } { { \rm { true \ positives } } + { \rm { false \ positives } } } } \end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}{ \rm { NPV } } = { \frac { { \rm { true \ negatives } } } { { \rm { true \ negatives } } + { \rm { false \ negatives } } } } \end{align*} \end{document}

The prevalence of malignancy based PPV and NPV were calculated as follows: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}{ \rm { PPV } } = { \frac { { \rm { sensitivity \times prevalence }} } { { \rm { sensitivity \times prevalence } } + {(1 - \rm {specificity}) \times ({ 1 - \rm {prevalence } } )} }} \end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}{\rm {NPV}} = { \frac { \rm { sensitivity \times {(1 - prevalence) } } } {{\rm {(1 - sensitivity) \times prevalence } } + { \rm specificity} \times ( 1 - \rm {prevalence} ) } } \end{align*} \end{document}

In the comparisons of cytology versus histology, and molecular FNA analysis (mFNA) versus histology, histology is the gold standard. In the comparison mFNA versus histology, a mutation-negative, histologically benign sample is considered as “true negative”; a mutation-positive, histologically malignant sample is considered “true positive”; a mutation-positive, histologically benign sample is considered as “false positive”; and a mutation-negative, histologically malignant sample is considered “false negative.”

Inadequate PCR product

In total, molecular testing revealed inadequate results in 30/355 analyzed FNAs (8%). While in 31 FNAs the RNA-based detection of one or more rearrangements failed, the DNA-based detection of point mutations failed in only two FNAs. In one sample, both RNA and DNA were insufficient for molecular testing. Of the 355 FNAs with available histology or follow-up of at least one year, 30 were classified “not evaluable” (NE), since they were characterized by at least one inadequate PCR and a mutation-negative test outcome in all the other tests (Table 1). The percentage of NE-FNAs was highest in the AUS/FLUS-FNAs (15.2% NEs), followed by 7.1%, 6.6%, and 3% NE-FNAs in the SFM-FNAs, FN/SFN-FNAs, and MAL-FNAs, respectively. FNAs classified as NE were not considered in the calculation of sensitivity and specificity.

Performance of cytology in comparison to histology

Cytological evaluation of the 355 FNAs with histology or follow-up of at least one year revealed 81 AUS/FLUS, 199 FN/SFN, 42 SFM, and 33 malignant results (Table 1). All 33 cytologically malignant samples were carcinomas at histology (30 classic papillary thyroid carcinomas (cPTCs), three fvPTC). The final histological classification of the 42 SFM-FNAs revealed 26 cPTCs, eight fvPTCs, one oncocytic subtype of FTC (oFTC), one FTC, two oFA, three FA, and one thyroiditis. The final histological classification of the 199 FN/SFN-FNAs revealed eight cPTCs, 12 fvPTCs, eight FTCs, six oFTCs, 68 FAs, 36 oFAs, 40 ANs, three thyroiditis, and 18 samples without surgery but follow-up of at least a year. The final histological classification of the 81 AUS/FLUS-FNAs revealed four cPTCs, four fvPTCs, three FTCs, one oFTC, 31 FAs, two oFAs, 20 ANs, one thyroiditis, and 15 samples without surgery but follow-up of at least a year. As shown in Table 2, these data result in a cytology-based risk of malignancy (ROM) of 15%, 17%, 86%, and 100% for the AUS/FLUS, FN/SFN, SFM, and malignant categories, respectively.

Overall performance of molecular FNA testing in comparison to histology

Molecular FNA (mFNA) revealed 85 mutations in total (Table 1). Forty-four BRAF mutations were detected in 36 cPTCs, seven fvPTCs, and one oFA. Unfortunately, although the detection of a BRAF mutation and the cytological diagnosis of SFM question the histological diagnosis of an oFA, the histological diagnosis could not be validated, since the histology block of this tumor could not be obtained because the patient did not consent. Furthermore, 22 NRAS mutations were detected in the FNAs, representing two cPTC, six fvPTCs, one FTC, seven FAs, two oFAs, one AN, one thyroiditis at final histology, and two FNAs with follow-up of at least a year. In addition, nine HRAS mutations could be detected in the FNAs (one fvPTC, five FAs, one oFA, and two ANs). One KRAS mutation was detected in an FNA representing an oFA at histology. Furthermore, the following rearrangements were detected: three PAX8/PPARG rearrangements in three FAs, four RET/PTC1 rearrangements in three cPTCs and one fvPTC, and two RET/PTC3 rearrangements in two cPTCs. A total of 240 FNAs were mutation negative, representing 21 PTCs, 10 fvPTCs, 11 FTCs, six oFTCs, 81 FAs, 33 oFAs, 47 ANs, four thyroiditis, and 27 FNAs with follow-up of at least a year. Overall, these data result in a sensitivity of 55%, a specificity of 88%, and a PPV and NPV of 69% and 80%, respectively, for mFNA (Table 3).

Performance of mFNA in the AUS/FLUS and FN/SFN group in comparison to histology

mFNA detected 17 mutations (two BRAF, seven NRAS, four HRAS, three PAX8/PPARG, and one RET/PTC3) in the 81 AUS/FLUS samples with histology or follow-up of at least a year, representing three cPTCs, four fvPTCs, six FAs, one oFA, two ANs, and one nodule with follow-up of at least a year (Table 1). Fifty-one AUS/FLUS FNAs were mutation negative, representing one cPTC, three FTCs, one oFTC, 21 FAs, 13 ANs, one thyroiditis, and 11 nodules with follow-up of at least a year. Based on these data, mFNA allowed the detection of 7/12 carcinomas in the AUS/FLUS category, while 10/56 benign samples revealed a false-positive test outcome, resulting in a sensitivity of 58%, a specificity of 82%, and a PPV and NPV of 41% and 90%, respectively (Table 3).

In the 199 FN/SFN samples, mFNA detected 22 mutations (two BRAF, 13 NRAS, five HRAS, one KRAS, and one RET/PTC3), with histology or follow-up of at least a year, representing four cPTCs, three fvPTCs, one FTC, nine FAs, three oFAs, one AN, and one nodule with follow-up of at least a year (Table 1). A total of 164 FN/SFN FNAs were mutation negative, representing two cPTC, eight fvPTC, seven FTCs, five oFTCs, 57 FAs, 32 oFAs, 34 ANs, three thyroiditis, and 16 nodules with follow-up of at least a year. Based on these data, mFNA allowed the detection of 8/30 carcinomas in the FN/SFN category, while 14/156 benign samples revealed a false-positive test outcome, resulting in a sensitivity of 27%, a specificity of 91%, and a PPV and NPV of 36% and 87%, respectively (Table 3).

Histologic re-evaluation of mutation-positive histological benign samples

For 16 mutation-positive histologically benign samples, histology blocks could be obtained for a histologic re-evaluation by a second pathologist who was aware of the mutation testing results. The re-evaluation was based on the criteria described by Piana et al. (19), and resulted in the reclassification of one FA into FTC, while five nodules where re-classified as FT-UMP (Table 4).