Thyroid swelling, or goiter, represents one of the most common clinical presentations encountered in endocrine practice. The vast majority of thyroid nodules are benign, yet thyroid cancer is the most frequent malignancy of the endocrine system, with a globally rising incidence over recent decades [1][2]. This increasing prevalence, largely attributed to the widespread use of high-resolution ultrasonography and incidental detection, has created a critical clinical dilemma: how to accurately identify the small subset of nodules that are malignant while avoiding unnecessary invasive procedures in the majority of patients with benign disease [3]. The cornerstone of current preoperative evaluation rests on two pillars: clinical assessment and cytopathological analysis. Clinical predictors—including patient age, sex, family history of thyroid cancer, history of head and neck irradiation, rapid nodule growth, voice hoarseness, cervical lymphadenopathy, and suspicious ultrasound features such as microcalcifications, irregular margins, and taller-than-wide shape—provide essential initial risk stratification [4][5]. However, clinical features alone lack sufficient sensitivity and specificity to guide definitive management [6]. Fine-needle aspiration cytology (FNAC), guided by ultrasound, has emerged as the gold standard diagnostic tool, boasting high sensitivity and specificity [7]. The Bethesda System for Reporting Thyroid Cytopathology (TBSRTC) has standardized reporting, categorizing results from nondiagnostic to malignant [8]. Despite its utility, FNAC carries inherent limitations. Up to 20% - 30% of aspirations yield indeterminate cytology (Bethesda categories III, IV, and V), a diagnostic gray zone where the risk of malignancy is variable but not negligible, often leading to diagnostic lobectomy or thyroidectomy for histopathological confirmation [9][10]. Consequently, many patients with indeterminate nodules undergo unnecessary surgery for what ultimately proves to be benign disease [11]. Therefore, there is a pressing need to refine preoperative risk prediction beyond cytology alone. Identifying robust clinical and cytological predictors that can distinguish benign from malignant thyroid swellings would enable more personalized decision-making [12]. Such predictors could help avoid surgery in low-risk patients, guide the extent of surgery (lobectomy vs. total thyroidectomy) in intermediate-risk cases, and expedite definitive treatment in high-risk scenarios [13]. This study aims to evaluate clinical and cytological predictors of malignancy in patients presenting with thyroid swelling, with the goal of developing a practical risk-stratification model to optimize patient management and reduce unnecessary surgical interventions.

This hospital-based, observational, cross-sectional study was conducted in the Department of Ear, Nose, and Throat (ENT) and Head-Neck Surgery at Rajshahi Medical College Hospital (RMCH), Rajshahi, Bangladesh, over a 6-month period from June 2014 to December 2014. The study population comprised all consecutive patients admitted to the department with a clinical and ultrasonographic diagnosis of thyroid swelling who subsequently underwent thyroidectomy. Of 52 patients screened, 2 were excluded (1 declined participation, 1 had incomplete data), and 50 were enrolled. A total of fifty patients were enrolled during the study period. Although the calculated sample size using the formula n = (Z²*pq)/d² (assuming a prevalence of malignancy of 10%, a margin of error of 5%, and a 95% confidence interval) was 136, the sample size was limited to fifty due to time constraints of the dissertation. Patients of any age, sex, or race with a clinically diagnosed goiter who voluntarily provided informed consent were included; those who declined to participate were excluded. All patients underwent a standardized evaluation, including detailed history-taking—covering age, sex, duration of thyroid swelling, presenting complaints, history of radiation exposure, family history of thyroid malignancy, and medication history—and a thorough physical examination. This encompassed general and systemic examinations, inspection and palpation of the thyroid gland (assessing size, consistency, nodularity, mobility, tracheal position, and fixity), assessment of cervical lymph nodes, and indirect laryngoscopy to evaluate vocal cord mobility. High-resolution ultrasonography of the thyroid was performed in all patients to characterize nodules, specifically to determine whether they were solid or cystic. Fine-needle aspiration cytology (FNAC) was conducted using a 22-gauge needle under sterile conditions; smears were air-dried and alcohol-fixed, then stained with Giemsa and Papanicolaou stains. Local cytological diagnoses were assigned to the Bethesda System as follows: colloid goiter was classified as Bethesda II (benign); follicular neoplasm as Bethesda IV (follicular neoplasm); and papillary carcinoma as Bethesda VI (malignant). For the purposes of calculating diagnostic accuracy, Bethesda II was considered negative for malignancy, Bethesda VI was positive for malignancy, and Bethesda IV was also treated as positive for malignancy—assuming the worst-case scenario for sensitivity analysis. Preoperative assessments included complete blood count, random blood sugar, blood urea, serum creatinine, urine routine and microscopy, chest X-ray (posterior-anterior view), and electrocardiography when indicated. Surgical procedures—ranging from hemithyroidectomy, subtotal thyroidectomy, to total thyroidectomy with or without neck dissection—were performed based on preoperative clinical and cytological findings, with the specific type documented. Resected thyroid specimens were sent for histopathological examination by the Department of Pathology at RMCH, serving as the gold standard for the final diagnosis (benign or malignant, including specific histological types). Data were recorded on a pre-designed data collection sheet, compiled, and analyzed using descriptive statistics (frequencies, percentages, and rates). Diagnostic accuracy measures—sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy—were calculated for FNAC against histopathology, accompanied by 95% confidence intervals. Malignancy rates were compared across age groups, sexes, types of swelling, symptom duration, and ultrasound findings using odds ratios and Fisher’s exact test, with p < 0.05 considered statistically significant. Non-significant findings are reported as descriptive only. Additionally, an exploratory risk stratification model integrating FNAC categories and patient age was developed for indeterminate (Bethesda IV) nodules. Ethical approval was obtained from the Institutional Ethics Committee of Rajshahi Medical College Hospital. All patients provided written informed consent in Bengali after a comprehensive explanation of the study; they were assured of their right to withdraw at any stage without affecting their medical care, and all data were kept confidential and used solely for research purposes.

Fifty patients with thyroid swelling who had thyroidectomy at Rajshahi Medical College Hospital from June to December 2014 were included. All 50 eligible patients who underwent thyroidectomy during the study period were enrolled consecutively; none were excluded. All underwent clinical assessment, ultrasonography, FNAC, and histopathology.

The study evaluated 50 patients with thyroid swelling undergoing thyroidectomy to identify clinical and cytological predictors of malignancy and assess FNAC accuracy. The malignancy rate was 14.0%, similar to previous reports: Htwe etal. reported 10% in goitrous thyroids [14], Tarrar etal. reported 13.33% [15], Islam etal. reported 18.65% in solitary nodules [16], and Huque etal. reported 21.11% [17]. The lower rate may be due to the inclusion of multinodular goitres (68%), which have a lower malignancy risk than solitary nodules [18]. Despite this, a 14% malignancy rate highlights the importance of preoperative risk assessment.

Age was a descriptively strong predictor of malignancy in this study. Patients at the extremes of age (<20 years and >50 years) had a malignancy rate of 33.3% each, compared to only 9.1% in those aged 21 - 50 years. This finding is consistent with the wellknown bimodal distribution of thyroid cancer risk [4][19]. Cady etal. demonstrated that younger age (<20 years) and older age (>60 years) are independent risk factors for aggressive thyroid cancer [20]. Similarly, the American Thyroid Association guidelines note that age <20 years and >70 years increases the suspicion of malignancy [4]. In our series, the youngest patient (11 years) had papillary carcinoma, and the oldest (59 years) had follicular carcinoma, reinforcing the importance of age in clinical decisionmaking.

Although thyroid swelling was more common in females (84%), the malignancy rate was higher in males (25.0% vs. 11.9%). This male predominance of malignancy, despite lower overall nodule prevalence, is well documented [5]. Tarrar etal. reported a 25% malignancy rate in males compared to 10.5% in females [15]. The odds ratio of 2.47 in our study (though not statistically significant, 95% CI: 0.41 - 14.8, p = 0.31 by Fisher’s exact test due to small sample size) aligns with the literature, indicating that on gender basis male confers high risk of thyroid cancer than female. Therefore, male patients with thyroid swelling warrant more aggressive evaluation.

Solitary thyroid nodules had a 26.7% malignancy rate, over three times higher than multinodular goiter (8.8%), aligning with studies showing higher cancer risk in solitary nodules [18][21]. Notably, 3 of 7 malignant cases also arose in multinodular goiters, indicating cancer can occur there. Short symptom duration (<2 years) correlated with higher malignancy (16.7%) than longer duration (≥2 years) (11.5%), likely because rapidly growing or recent nodules are more suspicious, as clinical guidelines emphasize [4][5].

Ultrasonography was valuable: all 7 malignant nodules were solid, and no malignancy was found in purely cystic lesions. The malignancy rate among solid nodules was 20.0%, aligning with a meta-analysis showing a sensitivity of 69% and specificity of 56% for solid composition [6][22]. Cystic nodules are almost always benign, but complex cystic nodules with solid parts can still be cancerous. Our finding that all malignant nodules were solid highlights ultrasound’s role in triaging for FNAC.

FNAC showed excellent diagnostic performance with a sensitivity 85.7%, a specificity 100%, and an accuracy 98.0%. These results are comparable to published data, such as Chandanwale etal. (sensitivity 90%, specificity 100%) [23] and Basharat etal. (sensitivity 80%, specificity 97.7%) [24]. The high specificity and positive predictive value (100%) suggest that a cytological diagnosis of malignancy is reliable. The negative predictive value (97.7%) indicates that benign FNAC effectively excludes malignancy, reducing unnecessary surgery. One case was misclassified—a follicular neoplasm later identified as follicular carcinoma—highlighting FNAC’s limitation in differentiating follicular adenoma from carcinoma [10][25].

The histopathological subtype distribution in this study showed that, among malignant cases, papillary carcinoma was the most common (6; 85.7%), followed by follicular carcinoma (1; 14.3%). This predominance of papillary carcinoma is consistent with global epidemiological data, which reports that papillary thyroid cancer accounts for 80 - 85% of all thyroid malignancies [1][2]. Similarly, a large Bangladeshi series reported papillary carcinoma in 76.9% of malignant thyroid specimens [26]. Among benign cases, multinodular goiter was the most frequent diagnosis (79.1%), followed by follicular adenoma (11.6%) and colloid goiter (9.3%). This distribution mirrors findings from other South Asian studies, where multinodular goiter remains the predominant benign lesion in patients undergoing thyroidectomy [14][17].

Six patients (12%) had a cytological diagnosis of follicular neoplasm (Bethesda IV), an indeterminate category. The overall malignancy rate was 16.7% (1/6), within the reported 15% - 30% range for Bethesda IV nodules [8][10]. In an exploratory analysis, adding the clinical predictor of extreme age (<20 or >50 years), we further stratified risk: patients with follicular neoplasm plus extreme age had a 50% malignancy rate, while those aged 21 - 50 had 0%. However, given that there are only six Bethesda IV cases and two patients in the extreme-age subgroup, this finding remains purely hypothesis-generating and does not substantiate direct operative recommendations. External validation involving larger cohorts is imperative prior to clinical implementation. If this approach is validated, it could potentially decrease the number of unnecessary total thyroidectomies performed for benign conditions, thereby reducing the risk of complications such as hypocalcemia and nerve injury [13]. The Bethesda System estimates malignancy risks of 0% - 3% for Bethesda II, 10% - 30% for Bethesda IV, and 97% - 99% for Bethesda VI [8]. Our observed rates (0%, 16.7%, and 100%) fall within these ranges, validating the applicability of the categories. The absence of Bethesda III and V cases reflects our small sample size and cytological practice.

Several limitations must be acknowledged: small sample size (n=50) limiting statistical power; singlecenter design with potential selection bias; short study period (6 months); absence of full Bethesda categories (III, V) and molecular markers (BRAF, TERT); limited ultrasound data (only solid/cystic consistency, lacking detailed features like microcalcifications). The proposed risk model requires external validation in larger, prospective cohorts. The proposed exploratory risk model requires external validation in larger, prospective cohorts.

This study confirms malignancy in 14% of patients undergoing thyroidectomy for thyroid swelling. Key associated predictors include age (<20 or >50), male sex, solitary nodule, short symptom duration, and solid ultrasound. FNAC has high accuracy, with a negative predictive value of 97.7%. For Bethesda IV nodules, an exploratory analysis suggests that extreme age enhances risk assessment, showing a 50% malignancy risk in a very small subgroup. A combined FNAC and age algorithm requires external validation before clinical use. Larger studies are needed to validate this model and add ultrasound and molecular predictors.