Investigation of correlation between cytological and histological findings in suspected carcinoma of thyroid

Introduction: The thyroid cytopathologic classification based on The Bethesda System for Reporting Thyroid Cytopathology (TBSRTC) standardized results derived from fine-needle aspiration (FNA). It has facilitated communication among clinicians, seems to reduce the number of unnecessary thyroidectomies and also improves the quality of malignancy detection. Objective: To evaluate the correlation between cytopathological samples classified according to TBSRTC and their histological finding after specimen examination following surgical procedure. Methods: Cytopathological samples classified according to TBSRTC were correlated with the respective anatomopathological finding after thyroidectomy. There were retrospectively included all cases of thyroidectomy with previous FNA in the same lateral position of the organ that occurred within two years. Results: The 200 cases of thyroid FNA analyzed were distributed in 25 nondiagnostic or unsatisfactory cases (I; 12.5%), 45 benign cases (II; 22.5%), 48 atypia or follicular lesions of undetermined significance [(AUS/FLUS), III; 24%], 23 suspected cases for follicular neoplasia or follicular neoplasias [(SFN/FN), IV; 11.5%], 40 cases suspected for malignancy (V; 20%), and 19 malignant cases (VI; 9.5%). The malignancy rates observed for categories I to VI were 12%, 13.3%, 29.2%, 43.5%, 85%, and 100%, respectively. Conclusion: The prevalence of Bethesda II category was lower than predicted by TBSRTC (60%-70%), as well as Bethesda III was higher than recommended (7%). Nevertheless, the corresponding malignancy rates were within the expected ranges.


introDuCtion
Thyroid carcinoma is the commonest malignant neoplasm of the endocrine system, accounting for 1%-2% of all malignant neoplasms in humans, being more frequent in women in an approximate ratio of 3:1 (1,2) .Studies have shown a worldwide increasing incidence trend, perhaps due to improvements in the detection of small tumors (1) .Thyroid nodules are very common in the general population and, although suspected for carcinoma, only 5% are actually malignant (2,3) .
Nodules smaller than 1 cm of maximum diameter are usually not biopsied unless exhibiting suspected ultrasound characteristics, such as microcalcifications, or if accompanied by other risk factors that increase suspicion of malignancy, such as history of head and neck radiation, family history of thyroid disease, presence of lymphadenopathy, female patient, and Asian ancestry (2,3) .
Fine-needle aspiration (FNA) of nodules in the thyroid gland followed by cytological evaluation, especially ultrasound guided one, is considered the gold standard for the differential diagnosis due to its simplicity of execution, safety, cost-effectiveness and diagnostic accuracy (4) .According to Huang et al. (2015) (1) , FNA shows sensitivity and specificity of approximately 83% and 92%, with false negatives varying from 1% to 21%, emphasizing that these rates may be technician-dependent.Therefore, diagnostic discrepancies are not uncommon.
The thyroid cytopathologic classification based on The Bethesda System for Reporting Thyroid Cytopathology (TBSRTC), introduced in 2007, standardized results derived from FNA and has facilitated communication among clinicians, reducing the interobserver variability of this procedure (4) .It establishes six diagnostic categories and attributes risk of malignancy, adding recommendations for each category as well (4) .Studies have shown that the use of TBSRTC reduces the number of unnecessary thyroidectomies and also improves the quality of malignancy detection (3) .
The objective of this study was to evaluate the correlation of cytopathological interpretation classified according to TBSRTC with histological findings of the specimen examination after surgical procedure.

MEthoDS
This cross-sectional retrospective study was performed based on database from XXX, located in Joinville, southern Brazil.There were collected cytological and histological data from all patients with previous FNA to thyroidectomy between May 2013 and May 2015.There were excluded cases of thyroidectomy without previous FNA or with FNA in a lobe not corresponding to that of thyroidectomy.The cytological result of FNA, according to TBSRTC, was correlated with the histopathological report after thyroidectomy.Data were analyzed using descriptive statistics.
This study was approved by the Research Ethics Committee of the Hospital Municipal São José under protocol no.2171978.

rESuLtS
There were analyzed 200 cases of thyroid FNA classified according to TBSRTC.Most of the results were contained in Bethesda III category (24%) (Table 1).
The correlation of the diagnoses derived from FNA with the histopathological examination of the resected specimens is shown in Table 2.Among the cases classified as Bethesda II, six (13.3%) were malignant lesions at the histological examination, being all of these diagnosed as papillary carcinomas, and one corresponding to the follicular variant.Among the 83 patients with malignancy diagnosis confirmed at the histopathological examination, 77 had been classified as Bethesda III or higher category on previous cytological examination.Therefore, the sensitivity and specificity of this test was 93% and 42%, respectively.

DiSCuSSion
In the present study, the prevalence of cases classified in diagnostic category III (24%) was higher than that predicted by TBSRTC (7%), as well as higher than prevalence rates (0.8% to 18.8%) identified in other studies (5)(6)(7) .This category has generated controversy due to the inconsistent use among different institutions and should be used as a last resort (6) .The malignancy rate observed Investigation of correlation between cytological and histological findings in suspected carcinoma of thyroid in our study (29.2%) was in the range reported in other studies (5% to 69%) (7,8) .It is recognized a great variation in the malignancy rates in the Bethesda III category among different institutions and among cytopathologists of the same institution, depending on where they have received training in cytopathology (9) .In our study, most of the malignant lesions in this category were diagnosed as papillary carcinoma (12/14; 85.7%), an equivalent result (86.8%) as described by Ho et al. (2014) (6) in patients with thyroid nodules classified according to TBSRTC between 2008 and 2011 at the Memorial Sloan-Kettering Cancer Center, in New York.A second opinion is indicated for the accurate diagnosis through FNA in cases classified in this category, since it is considered subjective and not homogeneous (5,9) .It is suggested that cases with focal nuclear atypia have a higher risk of malignancy and should be classified as Bethesda V (9) .
(7)(8) .Park et al. (2014) (10) , who also observed a lower than expected percentage in this category, state that a possible reason is the fact that many patients already presenting clinical suspicion of thyroid carcinoma were referred from other health care services for confirmation purposes of the diagnostic hypothesis by FNA.All malignant cases confirmed in the histological analysis corresponded to papilliferous carcinomas, which was also observed by Park et al.Other studies have shown histopathological diagnoses ranging from no malignant cases to diagnoses such as papillary carcinoma, follicular carcinoma, medullary carcinoma, and lymphoma (4,5,7,8,11) .Günes et al. (2015) (12) reported that, in false-negative cases in FNA, the interpretive diagnostic error was particularly severe in "gray-zone" of follicular adenomas, follicular variant of papillary carcinomas, and papillary carcinomas with cystic change. (7)and Arul et al. (2015) (8) compared their findings with different studies regarding the distribution of cases and malignancy rates according to categories of the TBSRTC.These results were compared with ours and are shown in Table 3.The malignancy rates identified in the present study were similar to those found in other studies across all diagnostic categories.

Garg et al. (2015)
The cases corresponding to category Bethesda I are more likely to present benign than malignant histology, as was identified in our study (Table 1) and described by Yang et al. (2007) (13) and Yassa et al. (2007) (14) (10.7% and 10% of malignancy rates, respectively) (8,9) .However, as reported by Arul et al. (2015) (8) , the prevalence of malignancy observed in Bethesda I may vary considerably (from 0% to 35.3%) and also a large number of patients in this category are not submitted to surgery, which may influence the observed rates (8) .
Among the malignant cases belonging to Bethesda IV category in our study, 80% corresponded to papillary carcinomas and 20% to follicular carcinomas on subsequent histopathological examination, which are results similar to those reported by Guo et al. (2017) (5) (75% papillary carcinoma and 25% follicular carcinoma).
In regard to the Bethesda V category in our study, 94.3% were papillary carcinomas.Differently, Garg et al. (2015) (7) reported that their three cases in this category corresponded to medullary carcinoma, follicular adenoma, and follicular carcinoma at the histological examination, with no papillary carcinoma.Finally, all Bethesda VI cases were confirmed as malignant lesions, as expected.The majority corresponded to papillary carcinomas (89.5%), which was also described by Garg et al.
The sensitivity observed in the present study (93%) was also similar to that of Garg et al. (88.9%), while the specificity was different (42% vs. 84.3%).When evaluating other studies, they showed the sensitivity of thyroid FNA ranged from 76.92% to 100% and the corresponding specificity between 64.6% and 98.5% (7) .The retrospective design and relatively small number of cases, as well as the fact that the diagnoses, both cytopathological and histopathological, were performed by different pathologists, may also have contributed as relevant limitations of this study.

ConCLuSion
We concluded that the cyto-histological correlation of the diagnoses evaluated in this study is in accordance with the current literature, although the prevalence of Bethesda II category was lower and Bethesda III category was higher than that recommended by TBSRTC.It is believed that the presence of pathologists while the FNA is being done may positively influence the diagnostic accuracy allowing a better correlation among clinical, imaging and cytological data.

taBLE 1 −
Distribution of diagnoses derived from thyroid FNA according to TBSRTC categories and corresponding malignancy rates observed in the histological analysis FNA: fine-needle aspiration.TBSRTC: The Bethesda System for Reporting Thyroid Cytopathology.FNA: fine-needle aspiration; TBSRTC: The Bethesda System for Reporting Thyroid Cytopathology.Clarissa A. Brites; Ludmila B. S. Balsimelli; Karina M. P. A. Coelho; Hercílio Fronza-Júnior; Jaqueline Stall; Paulo Henrique C. França

taBLE 3 −
Comparison between the rates of malignancy and prevalences in each diagnostic category recommended by TBRSTC observed in the present study and described by other authors TBSRTC: The Bethesda System for Reporting Thyroid Cytopathology; N/A: not applicable; * Garg et al. (2015) and Arul et al. (2015)(7, 8).