Discussion

This prospective study evaluated 178 patients with thyroid swellings through a structured multimodal pathway, with histopathological correlation in 68 operated cases. The findings are relevant across three domains: cohort epidemiology, pathological spectrum, and the comparative performance of FNAC and USG TI-RADS against histopathology.

The age distribution and pronounced female predominance (F:M = 1.54:1) mirror large population-based series [1,2]. Although hypothyroidism was the most common functional abnormality, thyroid functional status showed no association with malignancy, and mean TSH did not differ between benign and malignant groups, consistent with the broader view that TSH, while essential to overall management, lacks the discriminatory power to be used alone in operative decision-making [7]. Practically, this means thyroid function testing should continue to inform management, but biopsy and surgical decisions should rest on cytological and sonographic risk stratification rather than biochemical status.

USG TI-RADS achieved a sensitivity of 75.00%, specificity of 78.12%, and accuracy of 76.47%. The malignant/borderline histopathology rate increased progressively across TI-RADS categories and reached 22 of 27 (81.48%) among TR5 nodules. This graded pattern is consistent with the behavior expected of a well-calibrated risk stratification system and with contemporary ACR TI-RADS validation data [6]. Its false positives, comprising complex adenomatous goiter, densely calcified nodules, and Hashimoto pseudonodules, are well-documented sonographic mimics of malignancy [11]. This suggests that radiologists should apply particular caution when scoring nodules in a heterogeneous or pseudonodular gland, especially where autoimmune thyroiditis is common.

In this surgically verified cohort, Bethesda V/VI cytology showed 100% specificity and PPV, although the confidence intervals were wide; these findings support its high-risk clinical significance but require cautious interpretation [3]. Its more modest sensitivity (58.33%) is the more clinically important limitation, explained largely by the pathological spectrum encountered: follicular-patterned lesions such as FTC and follicular-variant PTC are poorly discriminated by cytology because their defining feature, capsular or vascular invasion, is an architectural finding visible only on histological sections [4]. NIFTP poses a similar diagnostic problem, since it may show cytological features overlapping with follicular-patterned lesions. In the present study, NIFTP was included within the malignant/borderline reference outcome because it is a clinically relevant lesion requiring definitive histopathological classification; this outcome definition should be considered when interpreting diagnostic-performance estimates [5,8]. These two entities, together with the three unexpected Bethesda II malignancies, account for most of the false negatives observed.

Malignant/borderline histopathology was identified in 9 of 10 (90.00%) Bethesda IV cases. Several factors plausibly contribute: tertiary referral centers accumulate a case mix enriched for complex follicular neoplasms already triaged away from primary care. The inclusion of NIFTP within the malignant/borderline outcome category may have increased the observed rate of clinically significant histopathology [8]. The small absolute number of Bethesda IV cases here makes the proportion inherently less stable. The high observed malignant/borderline histopathology rate among Bethesda IV lesions warrants confirmation in larger cohorts and should be interpreted alongside ultrasound findings, patient factors, and available molecular testing.

FTC outnumbering classical PTC as the leading malignancy represents a departure from most published Indian and international series, where PTC usually predominates. This may reflect regional iodine nutritional status, which is known to influence the follicular-to-papillary carcinoma ratio, together with a referral pattern favoring diagnostically ambiguous follicular neoplasms [1]. Whatever the explanation, it reinforces the broader theme of this study: follicular-patterned disease represents the principal blind spot of cytology-based screening.

The central, actionable finding is the complementary behavior of FNAC and USG TI-RADS: FNAC's high specificity and PPV make it the preferred tool for confirming malignancy, while USG's higher sensitivity makes it more useful for flagging structural concern in cytologically reassuring or indeterminate lesions [3,6]. The differing sensitivity-specificity profiles suggest that FNAC and USG may provide complementary information; however, the diagnostic performance of a prespecified combined testing strategy was not evaluated in this study. Molecular adjuncts such as BRAF, RAS family, and TERT promoter testing represent a promising next step for refining risk stratification specifically within indeterminate Bethesda III/IV categories [3,12].

Limitations

As a single tertiary-center study, this cohort is subject to referral bias toward complex, higher-risk lesions. Verification bias is a major limitation because histopathological confirmation was available for only 68 of 178 enrolled patients (38.20%), as surgery was undertaken according to routine clinical indications rather than a study-mandated protocol. The surgically verified subgroup was therefore likely enriched for patients with suspicious clinical, cytological, or sonographic findings, as reflected by the malignant/borderline histopathology prevalence of 36 of 68 (52.94%) participants. Consequently, PPV, NPV, overall accuracy, and observed disease prevalence should be interpreted as estimates for the operated subgroup rather than for the full cohort or for unselected patients with thyroid nodules. Occult malignancies among unoperated patients could not be excluded.

FNAC’s inherent limitation in follicular-patterned neoplasms, where the adenoma-carcinoma distinction depends on histological features unavailable to aspiration cytology, reflects a constraint of the technique itself rather than a limitation of this study; however, it materially affects the sensitivity reported here. A complete thyroid profile, including T3 and T4, was not available for the entire cohort. NIFTP was included within the malignant/borderline outcome category because of its diagnostic and management relevance; this classification choice may have influenced the reported diagnostic performance estimates. Next, histopathologists were not fully blinded to prior clinical or imaging findings because histopathological assessment was performed as part of routine clinical practice, which may have introduced interpretation bias. Finally, because ultrasonographic examinations were interpreted by a single radiologist, inter-observer reproducibility could not be assessed. Residual variability in cytological and radiological interpretation cannot be excluded, and the single-center design may limit generalizability