BRAF V 600 E Immunohistochemistry in Papillary Thyroid Carcinomas : Relationship Between Clinical and Morphological Parameters

Material and Method: We applied BRAFV600E immunohistochemistry in a cohort of 107 papillary carcinomas, 19 adenomas and 13 normal thyroid tissues that was chosen retrospectively between 2011 and 2015. Statistical analysis was based on semiquantitative immunohistochemistry findings. We also applied BRAF RT-PCR in a subgroup of 14 papillary carcinomas, 13 metastatic lymph nodes and 4 adenomas that was chosen randomly.


INTRODUCTION
Papillary carcinoma of the thyroid (PTC) is one of the most common cancers with an incidence rate of 14.42 per 100 000 person-years in 2010-2013 (1). PTCs have been known to have a better prognosis than other malignant tumours of the human body, although around 10% exhibit a worse clinical course than expected in PTCs. Several immunohistochemical stains have been used, such as cytokeratin 19 (CK19), Hector Battifora mesothelial cell-1 (HBME-1) and Galectin-3, to diagnose and detect this small group. The rapid development in molecular pathology has led to deeper efforts in coming up with the privileges of targeted therapy options.
The most relevant genetic alterations in PTC are generally mutually exclusive and in the vast majority of cases cause activation of the MAPK pathway; such as BRAF, RET and RAS mutations with BRAF mutations in the centre of attention (2). Mutations affecting the BRAF protooncogene are point mutations, small in-frame deletions, insertions and chromosomal rearrangements; the most frequent of which is the BRAFV600E point mutation. As a group, BRAF mutations activate BRAF kinase and lead to chronic stimulation of the mitogen-activated protein kinase pathway. In PTCs, BRAF mutations have been postulated as a cause of tumour recurrence (3) and worse prognosis (4,5), along with initial tumour pathogenesis. Polymerase • If a tumour extends directly into the vessel lumen, it should form a polypoid mass protruding into the lumen or exhibit thrombus formation in association with the tumour and not just bulging into the lumen.
• The cell aggregates within the lumen should be histologically identical to the tumour cells and be composed of epithelial cells and not of reactive endothelial cells.
• The intravascular tumour aggregate should be attached to the wall of the blood vessel and covered by a layer of endothelial cells.
Extrathyroidal extension was defined as tumour penetration through the thyroid pseudocapsule into the adjacent skeletal muscles or other organs.
Intrathyroidal spread was defined as an intraglandular dissemination of a tumour via lymphatic channels, and multiple small or larger satellite foci in the vicinity or remotely from the main tumour mass.
Information regarding the gender and age of the patients was obtained from the automation system of our hospital. Clinical follow-up was provided by the general surgery department. PTC cases were classified as either 'good prognosis (GP)' or 'poor prognosis (PP)' upon the clinical occurrence of lymph node metastasis, local recurrence and/ or distant metastasis. Clinicopathologic features are shown in Table I. Serial sections (4 µm thick) were obtained from the paraffin-embedded blocks of the selected preparations and fixed on positively charged slides to perform IHC. BRAFV600E IHC was performed manually using the Novolink ® Polymer Detection System (Leica, Australia). Additional information on IHC is summarised in Table II. Positive and negative control slides were also stained. All 31 cases (14 primary tumours (PT), 13 metastatic lymph nodes and 4 adenomas) were selected randomly, and BRAF mutation analysis was performed on these cases using the Cobas ® 4800 RT-PCR System (Roche Diagnostics, USA).

Assessment of Immunostaining
The assessment of BRAFV600E IHC in PTCs, melanomas and colonic adenocarcinomas is still under debate (9). However, in our study, nuclear with or without cytoplasmic staining was considered positive as in most of the literature. A semiquantitative approach was used based on the staining intensity of positively stained cells: negative, 1+ (weak staining), 2+ (moderate staining) and 3+ (strong staining) of any proportion of tumour cells (Figures 1-4). Any proportion of tumour cells with membranous and cytoplasmic staining with CK19, apical chain reaction (RT-PCR) and Sanger sequencing are the gold standard techniques to detect BRAFV600E mutation, whereas immunohistochemistry (IHC) needs more scientific evidence of high specificity and sensitivity. Some morphological findings like multicentricity, lymph node metastasis, tumour extension beyond the thyroid parenchyma and Psammoma bodies (6) may also predict the BRAFV600E mutation.
CK19 is a low molecular weight cytokeratin found in simple and complex epithelia, as well as in some carcinomas. An increased intensity of CK19 immunostaining is used for the diagnosis of PTC. HBME-1 is a marker of the apical surface of the mesothelium. An apical membranous staining of HBME-1 is also seen in PTCs. Galectin-3 is a β-galactoside binding lectin in charge of cell adhesion. Nuclear and cytoplasmic immunostaining is seen in PTCs.
Ki-67 detects the nuclei of cells in late G1, S, G2 and M phases. Proliferation index in PTCs is no more than 5% in general (7).
The aim of this study is to evaluate the immunoexpression of BRAFV600E, and its correlation with clinicopathologic parameters.

MATERIALS and METHODS
The study involved the use of formalin-fixed paraffinembedded tissue sections of histopathologically diagnosed cases of PTC (n = 107) and adenoma (n = 19) from the archives of the Department of Pathology. Of the PTC cases, 23 were microcarcinomas: 2 were follicular variants, 1 was oncocytic and 20 were classic. Twenty-three of the PTC group had metastatic lymph nodes available. These lymph nodes were assessed similarly. The slides that had been routinely stained with hematoxylin and eosin, CK19, HBME-1, Galectin-3 and Ki-67 were re-evaluated. The pT stage, necrosis, calcification, lymphatic invasion, vascular invasion, perineural invasion, tumour capsulation and capsule invasion, extrathyroidal extension, multicentricity, intrathyroidal spread and surgical margin status were assessed as prognostic parameters. To avoid controversy, certain criteria were used (8) as elaborated below.
Vascular invasion was defined as a direct tumour extension into the blood vessel lumen or a tumour aggregate within the vessel lumen. The criteria for vascular invasion are as follows: • The affected vessel must be located within the capsule or immediately beyond the capsule but not within the tumour nodule itself.
• The vessel should have a clearly identifiable wall with endothelial lining.
membranous staining with HBME-1 and nuclear and/ or cytoplasmic staining with Galectin-3 was considered positive. The eyeballing technique was used for the Ki-67 labelling index, and 100 tumour cells were counted in hot spot areas. A proportion of nuclear-stained cells with Ki-67 was recorded and divided into groups with a threshold of 5%.

Statistical Analysis
Statistical assessments were performed using the SPSS software (SPSS version 15, SPSS Inc., Chicago, IL, USA). Continuous variables were expressed as mean ± standard deviation together with a range (minimum-maximum). The comparison of categorical variables was performed using the chi-square test. Fisher's exact test was used to compare BRAFV600E IHC and RT-PCR. A P-value of less than 0.05 was accepted as statistically significant.
Twelve of the PTC cases were negative for HBME-1, one was negative for CK19 and six were negative for Galectin-3.
In BRAFV600E IHC, 31 of the PTC cases were negative. Twelve cases showed a Ki-67 proliferation index higher than 5%. For BRAFV600E IHC, positive cases exhibited varying percentages of staining as shown in Table III.
The results of BRAF RT-PCR of randomly selected cases are shown in Table IV            In cases where a metastatic lymph node (MLN) is present, the BRAFV600E positivity was 11% in PT. Two of those showed no staining in MLN, whereas PT was positive.
Another two cases showed a lower percentage of positive tumour cells in MLN (Table VI). However, PTs and conjugate MLNs were statistically correlated upon the BRAFV600E mutation (Phi = 83.7%, p = 0.0001).
The clinical prognosis was assessed in two separate groups as described earlier. The poor prognostic group was found to be statistically correlated with a higher pT stage (p = 0.005), classic morphology (p = 0.011), calcification (p = 0.017), lymphatic invasion (p = 0.008), vascular invasion (p = 0.0001), lack of tumour capsule (p = 0.004), extrathyroidal extension (p = 0.0001), intrathyroidal spread (p = 0.001) and positive surgical margin (p = 0.002). No association was found with the patient's age, sex, necrosis, mitotic index, perineural invasion, tumour capsule invasion and multicentricity. HBME-1 positive cases were found to be correlated with PP (p = 0.049), whereas CK19 expression, Galectin-3 expression and Ki-67 proliferation index were irrelevant.  associated with a classic variant, tall cell variant and poorly differentiated/anaplastic carcinomas that arise from welldifferentiated PTCs (15). The mutation rate is much lower in follicular carcinomas (16), which is similar to our results.
Intratumoural heterogeneity is a substantial phenomenon for understanding pathogenesis and its clinicopathologic role. As in other BRAF-harbouring tumours such as malignant melanomas and colorectal and pulmonary adenocarcinomas, PTCs have been shown to exhibit heterogeneously mutated tumour cells. Guerra  To start with associations between the BRAFV600E mutation and clinicopathologic parameters, we found noassociation with the patient's age and sex, as in the largescale meta-analysis of Wang et al. (19) and series of Shin et al. (20). In our experience, BRAFV600 mutation was found to be correlated with a higher pT stage, lymphatic invasion, perineural invasion, lack of tumour capsule, extrathyroidal extension and intrathyroidal spread. Several studies have stated various morphological findings, and their combinations are correlated with the BRAFV600E mutation, interestingly having extrathyroidal extension in common (6,21,22).
Surprisingly, the BRAFV600E mutation rate was higher in tumours with a lower mitotic index, as in tumours with a lower Ki-67 proliferation index, despite its incoherency. No effort has been found in the English literature that addresses theassociation between BRAF mutations and mitotic index or the Ki-67 proliferation index. Nevertheless, welldifferentiated PTCs are known to have a lower proliferation index than other malignancies. We observed that the Ki-67 In cases where the BRAFV600E mutation was present, poor prognostic incidents such as lymph node metastasis, local recurrence and distant metastasis were more frequent. However, we could not reveal any statistical significance (p = 0.255). On the other hand, BRAFV600E positive cases constituted a minority of 40% in the poor prognostic group.

DISCUSSION
PTCs are malignant tumours with an increasing incidence rate (1). The widespread use of radiological techniques and fine-needle aspiration biopsy has been partly responsible for this increasing incidence, although the vast majority of new cases are microcarcinomas (10). Lobectomy or total thyroidectomy with subsequent radioactive iodine ablation in selected patients is usually more effective; however, approximately 10% of PTC patients suffers from recurrence, lymph node or distant metastasis and hence require further intervention. BRAF mutations are the major promising development for PTC recently, providing a highly efficient anti-tyrosine kinase therapy. Using the criteria 3+ nuclear and cytoplasmic staining in the PTC group, the BRAFV600E mutation rate was 31.8%. This rate increased up to 59.5% in classic variant cases, but it decreased to 26% in papillary microcarcinomas (PMCs) and 9% in oncocytic variant cases. In follicular variant and adenoma cases, no mutation was detected. In the recent literature, the BRAFV600E mutation rate has been reported between 35% and 70%, and the mutations were more often proliferation index is higher than 5% in 19.6% of the PTC cases, reaching up to 15%. In addition, we did not find any significant association between mitotic/Ki-67 index and worse clinical and/or pathologic prognostic parameters. Guerra et al. (23) showed a higher rate of CK19 expression in BRAF-mutated tumours, whereas Galectin-3 was not associated with BRAF. In terms of HBME-1 expression and BRAF, our effort needs to be published first. However, in our series, no significantassociation was found between the BRAFV600E mutation and expression of CK19, Galectin-3 and HBME-1.
In cases where follow-up data are available, a survival analysis could not be made because there was no death by disease. The cases were assessed in two separate groups: GP and PP, as described earlier. The patient's age and sex were not found to be correlated with the prognosis. This is despite the fact that Howell et al. (24)  In our series, PP was found to be associated with a higher pT stage, classic morphology, calcification, lymphatic invasion, vascular invasion, lack of tumour capsule, intrathyroidal spread, extrathyroidal extension, positive surgical margin and loss of HMBE-1 expression. Likewise, Rossi et al. (26) have associated PP in poorly differentiated and anaplastic thyroid carcinomas with loss of HBME-1 expression.
The association between the BRAFV600E mutation and PP can be properly summarised by the meta-analysis of Wang et al. (19). In contrast to what has been reported recently. Pelttari et al. (27), with their lengthy follow-up duration, have shown that the BRAFV600E mutation is not correlated with lymph node metastasis and/or recurrence. Zheng et al. (28) have revealed that the BRAFV600E mutation and the recurrence within PMCs are not related. Nam et al. (21) have also shown that the BRAFV600E mutation is not significantly associated with lymph node metastasis. In these series, despite the overall concern, some morphological findings such as extrathyroidal extension are interestingly correlated with the BRAFV600E mutation. For instance, Shin et al. (20) revealed that the BRAFV600E mutation does not seem to be associated with the overall prognosis but morphological parameters are associated solely and together with aggressive behaviour. We also did not find any association between the BRAFV600E mutation and the overall prognosis but with such morphologic parameters.
In conclusion, BRAFV600E IHC with VE1 clone can be accepted as a reliable technique for detecting the BRAFV600E mutation. Our series of well-differentiated PTCs has exhibited a rate of BRAFV600E mutation similar to recent literature. With our effort, morphological findings may be considered keener than the BRAFV600E mutation in predicting aggressive behaviour. However, demographic, clinical and morphological findings and genetic alterations should be assessed together to estimate a more precise prognosis. Although further therapeutic interventions are needed, it is better to look for the BRAFV600E mutation in PT, lymph node metastasis, recurrent tumour and distant metastasis, if available.

ACKNOWLEDGEMENT
We would like to express our sincere gratitude to Prof. Nural Bekiroğlu (Marmara University, School of Medicine, Department of Biostatistics) for her supervision in our statistical assessments.