Cytomorphological Categorization of Thyroid Lesions according to The Bethesda System for Reporting Thyroid Cytology and Correlation with their Histological Outcome: An Indian Oncology Centre Experience.

Objective: The objective of this study was to report the experience of an Indian premiere tertiary care oncology center in reporting fine needle aspiration cytology of thyroid lesions according to the Bethesda system of reporting thyroid cytopathology (TBSRTC) given by National Cancer Institute (NCI). These were then correlated with their histopathological outcome, analyzing the level of specificity and sensitivity of the procedure. Material and
Methods: Aspiration cytology of thyroid lesions, presented during a 5.5-year duration, was reported retrospectively and prospectively, according TBSRTC, and correlated with their histopathologic diagnosis.
Results: A total of 431 patients were evaluated comprising 289 females and 142 males, with a median age of 52 years. Among the cytological categories 80 (18.6%) were non-diagnostic (ND), 131 (30.2%) benign, 45 (10.4%) follicular lesion of undetermined significance (FLUS), 27 (6.3%) follicular neoplasm, 33 (7.9%) suspicious for malignancy (SM), and 115 (26.7%) malignant. Histopathology reports were available in 142 of these cases. Final malignant diagnosis was reported in 11 of 14 ND (78.6%), 5 of 18 benign cases (27.7%); 9 of 17 FLUS (52.9%), 7 of 13 FLUS (53.89%), 19 of 20 SM (95%), and 58 of 60 malignant cases (96.7%). The procedure had sensitivity of 94.4%, specificity of 61.9%, positive predictive value of 90.3% and negative predictive value of 72.22%.
Conclusion: TBSRTC provides uniform categorization of thyroid cytology, which also helps in further management. This valid system has helped to streamline the reporting terminologies as well as the clinical management. Copyright:

INTRODUCTION

Fine needle aspiration cytology (FNAC) is a widely accepted, cost-effective, and minimally invasive procedure for screening and triaging patients with thyroid nodules. The Bethesda System for reporting thyroid cytology (TBSRTC) was introduced in October 2007 by National Cancer Institute (NCI) to standardize the cytopathological terminologies used in reporting thyroid nodule aspirates, and also provide with risk stratification, leading to consistent management guidelines.[1] Patients with suspicious and malignant cytology undergo surgery, avoiding unnecessary intervention and its complications in patients with benign disease. However, the procedure has its own limitations including false negative and false positive results, and the indeterminate findings. Inter-observer variations occur between pathologists with different diagnostic thresholds, experience, and sample preparation methods.

TBSRTC is an effective and robust classification system to guide clinical treatment of patients with thyroid nodules, which also provides an estimate of malignant potential of each individual category.

MATERIAL AND METHODS

The present study was designed with retrospective and prospective data of fine needle aspiration cytology (FNAC) of thyroid lesions, performed during the period 2011 to 2016 (5.5 years duration), in an Indian premiere tertiary care oncology center.

A total of 431 cases were collected and evaluated with complete demographic details of patients. The inclusion criteria comprised treatment naïve thyroid lesions, who underwent FNAC, with availability of slides (for reviewing). Both ultrasound-guided and non-guided aspirations were included in the study with patient's written consent. Any cases with history of previous thyroid malignancy, and treatment were excluded. A general informed consent was taken from the patient regarding sharing of clinical data for research purpose. All the patients’ information in the manuscript is anonymized and only de-identified data is used.

Smears were stained using modified Giemsa Stain for rapid on-site evaluation (ROSE) of adequacy; smears fixed in 95% ethanol were stained with Papanicolaou stain, and the air-dried smears were stained by standard Giemsa stain for the final cytopathologic examination. Material from lesions which yielded fluid on aspiration, were subjected to centrifugation to prepare cytospin smears.

The reporting and reviewing of archived slides was done by AM, SP and MK independently, and classified as per The Bethesda system for reporting thyroid cytopathology (TBSRTC),[1] [Table 1]. The recommended criterion for specimen adequacy was strictly followed, requiring at least a minimum of 6 groups of well-preserved follicular epithelial cells. The cytopathologic results were compared with the histopathological diagnoses in the cases which were followed by surgery [Table 2]. The results have been compared with those provided by TBSRTC and discussed with the data from other studies [Table 3].

Table 1

Summary of Cytologic Diagnoses according to TBSRTC and Distribution of Malignancies on Surgical Resection in the present study (n=431)

Category	No. of Cases (%) (n=431)	Mean Age (years)	Cases with Histopathology (n=142)	Malignant histopathology	Overall Risk of Malignancy (%) (lower bound limit)	Risk of Malignancy (%) (upper bound estimate)	Risk of Neoplasia (%)
Non-diagnostic	80 (18.6%)	55.3	14	11	13.8%	78.6%	78.6%
Benign	131 (30.2%)	48.58	18	5	3.9%	27.7%	27.7%
Follicular lesion of undetermined significance	45 (10.4%)	46.28	17	9	20%	52.9%	44.4%
Follicular neoplasm	27 (6.3%)	52.88	13	7	25.9%	53.8%	42.8%
Suspicious for malignancy	33 (7.9%)	45.78	20	19	57.6%	95%	95%
Malignant	115 (26.7%)	51.36	60	58	50.4%	96.7%	96.7%

Table 2

Details of cases with their available final histopathologic diagnosis

Diagnostic TBSRTC Category	Malignant on Surgery (total 109)	Final Diagnosis as Malignant lesions	Final Diagnosis as Benign lesions
Non-Diagnostic	11 of 14	PTC 2; FVPTC 2; Medullary Ca 2; PDCa 1; Undifferentiated (anaplastic Ca) 1; DLBCL 1; Lymphoproliferative neoplasm 1; Hodgkin Lymphoma 1	Colloid goiter with adenomatous change 2; Nodular goiter with lymphocytic thyroiditis 1
Benign	5 of 18	FVPTC 2; PTC 1; Papillary microcarcinoma 1; PTC (diagnosed with evidence of metastasis)-1
Follicular lesion of undetermined significance	9 of 17	PTC 5; FVPTC 2; Hürthle Cell Ca 1; WDC 1	Follicular adenoma 3; Hürthle cell adenoma 1; Adenomatous nodule in MNG 3; Ch lymphocytic thyroiditis 1
Follicular neoplasm	7 of 13	Follicular Ca 2; Hürthle cell Ca 2; PTC 1; FVPTC 1; PDCa 1	Hashimoto’s 1; Follicular adenoma 3; MNG 1; Inadequate 1
Suspicious for malignancy	19 of 20	PTC-8; Follicular Ca 2; Medullary Ca 2; FVPTC 1; PTC microcarcinoma 1; WDC 1; Follicular neoplasm 1; Anaplastic Ca 1; Undifferentiated Ca 1; DLBCL 1	Ch lymphocytic thyroiditis-1
Malignant	58 of 60	PTC 32; Anaplastic Ca 6; PDCa 3; Sarcoma 3; Medullary 2; Hürthle cell Ca 2; FVPTC 1; Undifferentiated Ca 1; NEC 1; Follicular Ca 1; Sarcomatoid Ca 1; Metastatic Ca 1; DLBCL 3; PTCL 1	Benign nodule 1; MNG with paraganglioma 1

Ca: Carcinoma; PTC: Papillary thyroid carcinoma; FVPTC: Follicular variant of PTC; PDCa: Poorly differentiated carcinoma; WDC: well differentiated carcinoma; DLBCL: Diffuse large B-cell lymphoma; NEC: neuroendocrine carcinoma; PTCL: Peripheral T-cell lymphoma; MNG: Multinodulat goiter

Table 3

Comparison of present study data with literature from other similar studies

Authors	Total Cases	Non-Diagnostic		Benign		Follicular lesion of undetermined significance		Follicular Neoplasm		Suspicious for Malignancy		Malignant
		Cases	ROM	Cases	ROM	Cases	ROM	Cases	ROM	Cases	ROM	Cases	ROM
Present Study	431	18.6%	78.6% (11/15)	30.2%	27.7% (5/18)	10.4%	52.9% (9/17)	6.3%	53.8% (7/13)	7.9%	95% (19/20)	26.7%	96.7% (58/60)
Mahajan et al.[2]	4562	3.5%	50% (2/4)	79.6%	7.8% (13/166)	2.5%	50% (6/12)	3.9%	23.6% (9/38)	0.5%	75% (3/4)	9.8%	85.4% (94/110)
Ratour et al.[3]	2210	14.3%	-	65.5%	-	11%	22.5%	4.9%	-	2.3%	-	2%	-
Bhasin et al.[4]	80	1.3%	-	61.3%	-	10%	-	20%	-	3.8%	-	3.8%	-
Theoharis et al.[5]	3207	11.1%	-	73.8%	-	3%	-	5.5%	-	1.3%	-	5.2%	-
Hershman et al.[6]	25,445	13%	16.8%	59%	3.7%	9.6%	15.9%	10.1%	26.1%	2.6%	75.2%	5.4%	98.6%
Jo et al.[7]	3080	18.6%	8.9% (12/135)	59%	1.1% (20/1792)	3.4%	17% (9/53)	9.7%	25.4% (45/177)	2.3%	70% (39/56)	7%	98.1% (151/154)
Tepeoglu et al.[8]	1021	11.9%	-	68.3%	0	9.8%	12.7%	4%	35%	3.5%	91.4%	2.4%	100%
Melo-Uribe et al.[9]	196	4.1%		23.5%		2%		16.8%		37.2%		16.3%
Oncology hospital	104	6.7%	71.4% (5/7)	6.7%	14.3% (1/7)	2.9%	66.7% (2/3)	15.4%	56.3% (9/16)	37.5%	97.4% (38/39)	30.8%	100% (32/32)
Non-Oncology Hospital	92	1.1%	100% (1/1)	42.4%	12.8% (5/39)	1.1%	100% (1/1)	18.5%	23.5% (4/17)	36.9%	82.3% (28/34)	0	0

ROM: Risk of Malignancy

RESULTS

A total number of 431 thyroid cytology cases were evaluated. The age range of patients varied from 10 to 87 years, with a median age of 52 years. Among these, 289 were females (67.1%) and 142 were males (32.9%), with female to male ratio of ~ 2:1.

The breakdown of the cytological diagnostic category based on the current TBSRTC was as follows: 80 (18.6%) non-diagnostic (ND), 131 (30.2%) benign, 45 (10.4%) follicular lesion of undetermined significance (FLUS), 27 (6.3%) follicular neoplasm (FN), 33 (7.9%) suspicious for malignancy (SM), and 115 (26.7%) malignant. [Table 1] We calculated risk of malignancy (ROM) and risk of neoplasia (RON) based on cases with histologic follow-up (upper bound estimate), as has been done by Mahajan et al.[2] in their study, instead of determining overall ROM using total number of cases reported on cytology (lower bound estimate). Difference between these two type calculations suggest a selection bias for surgery in cases with higher suspicion for malignancy on radiology and clinical findings.

A repeat procedure was performed in 16 of 431 cases, with change of diagnosis in 14 cases considering the diagnosis with higher malignant potential, thus affecting the patient management and proving the importance of adequate sampling. Repeat aspirate was performed in 9 ND cases, yielding classifiable material in 7 cases; 2 of these were reported benign, 2 malignant, and 3 FLUS. One of the two cases that remained ND on repeat smear turned out to be a medullary carcinoma on final histopathology; probable cause of inadequacy being the non-yielding amorphous amyloid material.

A total of 97 aspiration of 431 cases was performed in known case of malignancies of other sites. These included cases of head and neck carcinoma (17), breast carcinoma (22), upper and lower GIT (13), lymphomas (13), lung cancer (10), renal cell carcinoma (5), female genital tract carcinomas (7), gall bladder cancer (2), prostate cancer (2), and one case each of urinary bladder carcinoma, melanoma, GIST, mesothelioma, osteosarcoma and Warthin's tumor.

Surgical resection (lobectomy or total thyroidectomy) was performed in 142 of these cases, and 109 of these were malignant. The data has been summarized in Table 1 and 2, along with patient's demographic details.

Among the 80 cases of ND category, 14 patients with a high clinical suspicion of malignancy underwent surgery, and 11 (78.6%) of them were malignant. One of the cases showed scattered inflammatory cells, thus being inadequate according to criteria, but later was diagnosed as Hodgkin lymphoma on histopathology (cytology did not show any RS-like cell).

Of the 131 cases classified as benign, 5 of 18 surgically resected thyroid nodules were malignant (27.7%). Among these 131 cases, 63 were reported as colloid nodule, 38 benign follicular nodule, 13 Hashimoto's or chronic lymphocytic thyroiditis, 12 adenomatous nodule, 2 thyroglossal cyst, and 1 each as necrotizing granulomatous thyroiditis, inflammatory lesion, and autoimmune thyroiditis. None of the benign thyroid case showed acute inflammation. One of the cases with benign cytology of thyroid yielded metastatic papillary carcinoma in upper deep cervical lymph node, proving the case as malignant, with probable inadequate sampling from the nodule.

In the FLUS category, 17 of 45 cases were followed up with surgery, with 9 (ROM 52.9%) diagnosed as malignant; and in FN category 13 of 27 cases underwent surgery, with 7 (ROM 53.8%) diagnosed as malignant. In one of the cases diagnosed as FN, presence of rib deposit with similar follicular epithelial cells proved it to be a follicular carcinoma. One case remained as follicular neoplasm on histopathology, as only trucut biopsy was performed, which cannot rule out capsular or vascular invasion for a follicular carcinoma. Risk of neoplasia (RON) calculated in FLUS and FN was 44.4% and 42.8%.

Among the cases suspicious for malignancy (SM), 20 of 33 underwent surgery, of which 19 (95%) were confirmed to be malignant. One of the cases SM was finally diagnosed as Hürthle cell adenoma with a focus of papillary microcarcinoma, with possible aspiration done from the adenoma part of nodule.

Of the 115 malignant cytology cases, 2 cases were reported on repeat aspirate, changing diagnosis from benign to malignant. Surgery was done in 60 cases, and 58 of these were confirmed as malignant on histopathology (96.7%). Of the 2 non-malignant cases on histopathology, one was a benign nodule and other a multinodular goiter with co-existent paraganglioma. Of the 115 malignant cases, 6 cases were metastatic carcinoma in a previously known case, and only 1 of these underwent histopathologic confirmation. Lymphomatous process was reported in 4 of these cases.

Of these 97 FNAC done in known case of malignancy at other site, 22 (22.7%) were malignant (6 metastatic carcinoma, and 15 primary carcinomas of thyroid, and one could not be categorized as primary versus secondary on cytology).

In our study, the sensitivity (true positive rate) of procedure was 94.4% and specificity (true negative rate) was 61.9%. Positive predictive value for suspicious and malignant was 90.3% and negative predictive value 72.22%.

DISCUSSION

Thyroid nodules commonly occur in about 4–10% of the general population, and the incidence has increased to 10 times higher (~11-50%) with the aid of ultrasonography.[34] Thyroid malignancies constitute 0.1-0.2% of all cancers in India.[45] The clinical prevalence in females (5.3% to 6.4%) is more than that in men (0.8% to 1.6%), especially the benign lesions.[3] The presented study included patients with median age of 52 years, 67.1% being women. A similar study of 196 patients comprised of 88.8% women, with a mean age of 50.8 years.[6]

It is recommended to evaluate all thyroid nodules >1 cm in size, both palpable and radiologically detected.[5] Less than 5% of these thyroid nodules are malignant requiring operative and medical therapy, and it is critical to differentiate between benign and malignant lesions, so as to avoid unnecessary surgery and the related complications.[3] By using FNAC 70-80% of thyroid lesions can be accurately classified as benign or malignant with a 92% negative predictive value (NPV) for benign and 100% positive predictive value (PPV) for malignancy.[7]

Percentage of surgically removed malignant thyroid nodules has increased from 14% to 50% with the use of FNAC.[8] With a high sensitivity of more than 90%, it is the recommended preliminary test to detect thyroid cancer. However, it has a low specificity of 50% to 65% which leads to many unnecessary surgical procedures.[4]

Accuracy of reporting is essential for management; however, lack of reporting consistency leads to confusion among clinicians regarding management. The reasons include inadequate sampling, varied categories used for reporting in different laboratories (especially for cases of atypical, suspicious and follicular neoplasm), and difference in expertise of reporting pathologists.[3] Among benign category also the descriptive terminologies used like colloid goiter, multinodular goiter (MNG), adenomatous hyperplasia, adenomatous goiter with hyperplastic nodules, hyperplastic nodular goiter and colloid cyst adds to the confusion.[9]

Categorization of cytological features into defined classes helps provide uniformity and quality control to the reporting across different laboratories and reduces inter-observer variation.[10] In 2006 and 2009, the American Thyroid Association and the Italian Society of Pathology and Cytology published a 4-tiered classification for thyroid lesions: nondiagnostic, benign, malignant, and suspicious or indeterminate. In 2010, the British Thyroid Association had proposed a 5- to 8- tiered terminology.[4] The Bethesda System for Reporting Thyroid Cytopathology (TBSRTC) was instituted in a multidisciplinary conference held at the National Cancer Institute in Bethesda, Maryland, in 2007, with the goal of standardizing the practice of thyroid cytopathology.[110] Revised guidelines were published in 2018, including role of molecular testing in cytopathology examination.[1]

TBSRTC is a uniform, 6-tiered reporting system, with defined diagnostic categories, standardized nomenclature, predictive risk of malignant outcomes for each category and recommendations for management; thus, minimizing the number of inconclusive cases, providing uniformity between laboratories and clinicians, and elucidating a uniform treatment.[410]

The 6 categories of TBSRTC are: nondiagnostic (ND), benign, follicular lesion of undetermined significance/atypia of undetermined significance (FLUS/AUS), follicular neoplasm (FN), suspicious for malignancy (SM) and malignant.[1] It is widely used in USA and Europe, but still has relatively reduced acceptability in Asia and Africa.[9] A retrospective study conducted by comparing previous classification with TBSRTC, and calculating the kappa score, indicated a significant reduction in number of inconclusive cases (P < 0.001). It also showed a strong inter-observer agreement with TBSRTC (Fleiss kappa score = 0.6561).[11]

Smears are considered as non-diagnostic (category 1) if they fail to fulfill the adequacy criteria (as described above), and are cystic lesions with insufficient or no follicular cells.[1] Predominant causes for ND smears include inadequate sampling, hemorrhage, fibrosis, small nodule not amenable to needling and inexperienced pathologist.

In our study, 18.6% of FNAC were non-diagnostic. Most of the other studies have shown lower rate than ours, ranging from 11.1% to 14.3%.[3412] Overall ROM among our ND cases was higher (13.8%) as compared to 5-10% recommended by TBSRTC.[1] Upper bound ROM was 78.6% as compared to 50% reported by Mahajan et al.[2]

Aspirates classified as benign/negative for malignancy (category 2) account for the majority of the thyroid FNAC samples. This category comprises a morphologically varied group of benign histological lesions, including colloid nodules, hyperplastic (adenomatoid) nodules, multinodular goiter, lymphocytic or granulomatous thyroiditis, and aspirates with only few benign follicular cells. Cellular smears show monolayered sheets of follicular cells without significant nuclear overlapping or atypia. The ROM in this category is 0–3% and most of these patients are followed up clinically.[13]

In our data, 30.2% were benign, much lower than other comparable studies having benign cases ranging from 61.3% to 73.8%.[34912] Reason for this is our center being a tertiary cancer institute, receiving mostly referred cases with high suspicion of malignancy. Many of our benign these cases were incidental thyroid nodules found during work-up for malignancies at another site. High upper bound ROM seen our ND (78.6%) and benign (27.7%) cases underlines the significance clinician's discretion for therapy, in the clinically and radiologically suspicious lesions.

’’Atypia of undetermined significance or follicular lesion of undetermined significance’’ (AUS/FLUS), category 3, is the most controversial. These are defined as smears with cytological and architectural atypia, but not convincing enough for a follicular neoplasm.[4] The features include focal to mild cytological atypia, atypical cast lining cells, histiocytoid cells, sparse cellularity with microfollicular architecture and minimal nuclear atypia, Hürthle cell aspirates, psammomatous calcification in absence of nuclear features, and atypical lymphoid cells.[1] Such cases are recommended to be kept on follow-up with repeat aspirate whenever possible, or a surgical intervention preceded by frozen section.[1101314]

The reported ROM for FLUS is 10% to 30%.[1410] In studies reporting reclassification according to TBSRTC, least inter-observer agreement has been seen in FLUS.[8] Different researchers have exhibited markedly variable incidences of FLUS (0.7-18%) and their malignant outcome (6-48%) in resections.[15]

We reported 10.4% cases as FLUS, which was similar to data of 9.8% to 11% in few other studies[34913]; whereas lower incidence of 3%[12] has also been reported. The ROM in our data (52.9%) was higher than TBSRTC recommended range, and those reported by others as 12.7% to 37.8%.[31415161718]

Adequate cellularity and additional training may help better distinguish true benign and atypical cases, with better reproducibility, leading to prudent use of category FLUS.[419]

Incidence of AUS is inversely correlated to ND (ND decrease by 1.1% for every 3.5% increase in AUS), while is directly correlated to malignant (M) incidence. This is because M lesions are more likely to be reported as AUS than as ND or benign.[20] Recommended AUS: M ratio is 1-3. A ratio of >3 is seen with overdiagnosis of AUS or underdiagnosis of M, and <1 ratio with low AUS rates given at expense of sensitivity.[13] Overcalling of benign cases as AUS gives a low M rate; whereas under calling SM and FN as AUS will yield a higher M rate.[19]

Many studies on reclassification have reported >50% of AUS as benign.[2122] In TBSRTC study, only 33% of FLUS cases underwent surgery, from which ROM was calculated; however accurate outcome would be known only if 100% of cases undergo surgery, which is not feasible.[19]

Mahajan et al.[2] discussed a large study on indeterminate thyroid nodule categories (category 3, 4, 5), reporting ROM and RON in each category. They demonstrated an increase in percent of FLUS/AUS and SM category from 0.2% to 4.9%, and 0.2 to 1.0%, respectively, over a period of 6 years. They subclassified AUS/FLUS based on cytomorphologic features like cellularity, microfollicles, Hürthle cell change, PTC like changes, nuclear atypia, hyperplastic changes and macrofolicles.

Follicular neoplasm (FN) (category 4) is diagnosed in smears having moderate to high follicular cellularity with predominant microfollicular or trabecular configuration of cells, scant colloid, and those with predominance of Hürthle cells.[7] The purpose of FN category is to identify nodules that may be follicular carcinoma and to triage them for surgery. According to TBSRTC, 25-40% of FN cases prove to be malignant.[1]

Among our cases 6.3% cases were classified as FN, which was similar to that of 5.5% reported by Theoharis[12] and more than Mahajan et al.[2] 3.9%. ROM for FN in our data is 53.8%, and overall ROM (lower bound) is 25.9%, which is within range stated by TBSRTC. Other similar studies have reported higher rates of 32.2% to 36%,[122324] as well as lower rates of 12.7% in their results.[3]

Subclassifying indeterminate FNA as follicular lesion or follicular neoplasm is clinically relevant, and incidence of malignancy has been reported as higher in follicular neoplasm compared with lesion (21.4% versus 7.0%, respectively; P = 0.0005).[25] Our data had almost similar ROM among FN and FLUS (53.8% and 52.9%, respectively).

Smears with cytological features of papillary, medullary or metastatic carcinomas, or lymphomas, were classified as suspicious for malignancy (SM) (category 5). Smears suspicious for PTC show nuclear overlapping with rare nuclear grooves, and/or pseudoinclusions along with thick colloid. These patients carry a high ROM (50–75%), and should undergo either a total thyroidectomy or surgical lobectomy.[1] In a study comparing data from the pre- and post-TBSRTC periods, ROM in SM lesions increased from 44% to 65% (P = 0.07). This highlights the improved diagnostic accuracy of the category.[17]

We reported 7.9% cases as SM, which was higher than some other data with 1.3% to 3.5%.[3412] The final histopathologic ROM rate was 95%, higher than others with rate of 70% to 91.4%.[31416]

The lesions with unequivocal neoplastic cytological features are classified as malignant (category 6) with type specification if possible, and are expected to account for 4–8% of all thyroid FNACs. The management of malignant cases is total thyroidectomy.[3]

Our center being a tertiary care center for oncology, receives filtered cases from the peripheral regions, thus an increased incidence of 26.7% of malignant cytology was seen. This was much higher than all other studies with rates ranging from 2% to 5.2%.[3412] The ROM on final histopathology was 96.7%, similar to that recommended by Bethesda (97–99%)[1] and other studies as well, with 98% to 100% rate.[31416]

A cross-sectional study performed on 196 FNA results, which included patients from oncology hospitals (OH) and non-oncology hospitals (NOH), showed higher incidence of SM and malignant cases in OH (68.3%), and more of benign cases in NOH (42.4%) with none being malignant. The final ROM was 83.6% in OH vs 42.4% in NOH, which was statistically significant (P = 0.03).[6]

Most of the cases in the ’malignant’ category consist of papillary thyroid carcinoma (PTC), reported rates as 59.7% to 67%.[26] In our data PTC formed 46.8% of all histologically proven malignant cases. The pitfalls in reporting PTC on cytology include specimen inadequacy, skill of the aspirant, and the experience of interpreting pathologist. False negative results can be seen in PTC with cystic degeneration, and lymphomas reported as hashimoto's thyroiditis due to presence of lymphocytes. A false negative rate of 4% has been reported by Bakhos et al.[26]

False positive results can be seen in cases of follicular and Hürthle cell adenomas diagnosed as PTC or suspicious for PTC. Various features like pseudopapillae, syncytial sheets, nuclear grooves and pinpoint nucleoli (in chronic lymphocytic thyroiditis and Hürthle cell neoplasms), intranuclear inclusions (in parathyroid adenoma and mesenchymal repair) can lead to false positive reports.[27]

On final histopathology, we reported the follicular variant of PTC (FVPTC) in 11.1% of cytological malignant cases, and formed 8.3% of all malignancies. FVPTC is a diagnostic dilemma with subtle papillary carcinoma like nuclear changes with overlapping benign and neoplastic features. In a retrospective study including 152 cases of FVPTC, malignant cytology was reported in only 27% of these.[28]

The malignant lesion with sarcomatoid morphology would need biopsy along with immunohistochemistry to reach an accurate diagnosis; differentials include medullary carcinoma, anaplastic carcinoma, undifferentiated malignancies, sarcomatoid squamous cell carcinoma (with metastasis from head and neck tumors), and leiomyosarcoma.

Our data largely reflects the results from an oncology center, and the data would vary with the experience and knowledge of pathologists, as well as the nature of laboratory or hospital setup with differing population cohort.

Ethical consideration

As the study required only routine cytology slides and FFPE tissue material with no additional financial requirements, and consent being taken from every patient at the time of hospital admission regarding utilization of data for research purpose, the protocol did not need formal ethical clearance from review board.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.