Comparison of immunohistochemistry and direct sequencing methods for identification of the BRAFV600E mutation in papillary thyroid carcinoma

Abstract

Background: BRAFV600E mutation is the most common somatic variant in papillary thyroid carcinoma (PTC) and is associated with aggressive prognostic factors. The conventional detection methods for BRAF mutations is polymerase chain reaction followed by Sanger sequencing. Recently, an immunohistochemistry (IHC) method-using a BRAFV600E specific antibody (VE1) has been developed and widely adopted in the clinics. However, there is lack of evidence regarding the comparability of the IHC and Sanger sequencing methods.

Methods: We retrospectively reviewed the clinical data of patients who underwent thyroidectomy in Seoul National University Hospital from January 2013 to October 2016. Among a total of 3584 patients, samples from 886 cases had been analyzed by both IHC and direct sequencing. We excluded cases in which the above tests were performed with tissues obtained lymph node metastasis or histology of other PTC subtypes. Finally, patients with classic PTC, where both tests had been performed in primary tumor tissue, were included in the study (n=697).

For IHC staining, we used a commercially available BRAF V600E mutation-specific antibody (clone VE1, Spring Bioscience, Pleasanton, CA, USA). IHC was performed using paraffin-embedded tissue microarrays (TMAs). Briefly, TMA sections were moved to adhesive slides and placed in a drying oven at 60℃ for 30 min. Using the Benchmark XT system, specimens were deparaffinized with EZ prep solution (Ventana Medical Systems, Inc, Tucson, AZ, USA). During the pretreatment step, cell-conditioning solutions containing EDTA (Ventana) were applied for 64 min. Subsequently, antibody was applied to the slides, and incubated for 32 min. And a positive control was attached to the same slide. Finally, stained tissue sections were examined by a microscopy.

PCR amplification and direct Sanger sequencing were performed to detect the BRAFV600E mutation in 697 PTC samples. In cases with multifocal PTC, the largest tumor was selected for sequencing analysis. The target region was marked on hematoxylin and eosin-stained slides. We extracted DNA with Maxwell instrument. Briefly, BRAF exon 15, potentially containing the T1799A transversion (encoding BRAFV600E) was amplified by PCR from genomic DNA. After purification of the PCR products, direct bidirectional sequencing was performed using a 3730 DNA Analyzer (Applied Biosystems, Foster City, CA, USA). We performed IHC and direct sequencing methods simultaneously on the sample.

Results: BRAF mutation was detected in 90.0% (627/697) of samples using IHC and 83.4% (581/697) by direct sequencing. The diagnostic parameters of IHC compared with Sanger sequencing were as follows: 100% sensitivity (581/581), 60.3% specificity (70/116), 92.7% positive predictive value (581/627), and 100% negative predictive value (70/70). No false negative results were recorded using IHC. The overall concordance rate between the two methods was 93.4% (651/697). Discordant results were found in 46 samples (6.6%), 29 of which were from cases with small tumors (< 6mm), eight were from cases with low tumor cellularity, and nine were specimens yielding low quality DNA.

To investigate clinical and pathological differences among patients with discordant results, we split patients into three groups: Group 1 contained patients with positive results in both IHC and direct sequencing (n = 581)

group 2 included patients with positive results by IHC but negative results using direct sequencing (n = 46)

and group 3 consisted of patients with negative results from both IHC and direct sequencing (n = 70). We compared the sex ratio, age, gross ETE, tumor size, lymph node metastasis, tumor multifocality, thyroiditis, tumor lymphatic invasion, and tumor angio-invasion among the three subgroups. When these three groups were compared, group 3 patients had larger tumors than those in group 1, and group 1 patients had larger tumors than those in group 2 (p < 0.001). An increased frequency of lymph node metastasis was observed in group 1 compared with group 2 (p = 0.007), however, more group 2 than group 1 patients had thyroiditis (p = 0.001). Comparisons of tumor lymphatic invasion, and tumor angio-invasion that group 1 and 3 patients exhibited a higher incidence of these phenomena than those in group 2 (p = 0.005 and p = 0.006 respectively).

Univariate and multivariate logistic regression analysis showed that tumor size and thyroiditis were independent predictors for discordance between IHC and direct sequencing. Of these factors, tumor size is the most powerful predictor, decreasing the odds ratio (OR) of discordance between IHC and sequencing to 0.249 (95% confidence interval [CI], 0.113-0.548) in multivariate analysis. Thyroiditis is also predictor, increasing the OR of discordance between these methods to 2.825 (95% CI, 1.516-5.266) in multivariate analysis.

Conclusions: IHC using the VE1 antibody is a reliable and highly sensitive method to detect the BRAFV600E mutation in classic PTC. When we analyzed the results according to the discordancy of the IHC and sequencing, the smaller tumor size and the presence of thyroiditis did not match the results of the two tests. Because the results of this study suggest that the sequencing results were likely to be false negative if the two test results were inconsistent, IHC is an acceptable test method in patients with small sized thyroid cancer or thyroiditis.