Missed Malignancies in Unbiopsied Thyroid Nodules: Diagnostic Performance of Combined TI-RADS and Bethesda Systems

Purpose Thyroid nodules are a widespread pathology. While only a small percentage of cases prove to be malignant, careful preoperative evaluation is essential to identify patients who need intervention. We aimed to evaluate the diagnostic performance of preoperative ultrasound and fine-needle aspiration cytology in predicting thyroid malignancy. Methods This cross-sectional study included 100 adult patients with 145 suspicious thyroid nodules who were scheduled for hemithyroidectomy or total thyroidectomy. Postoperative histopathology was considered the reference standard. Ultrasound findings and FNAC results were compared with the final histopathology. Results Final histopathology revealed 63.5% benign and 36.5% malignant nodules, with 58% PTC being the most common malignancy; 27% of the unbiopsied nodules were found to be malignant. Malignant nodules had significantly higher TI-RADS and Bethesda scores (p-value < 0.001, 95% CI 5–5; p-value < 0.001, 95% CI 4–5, respectively), while nodule size and number showed no significant association. ROC analysis demonstrated good diagnostic performance for the Bethesda scores (0.819 AUC, 95% CI: 0.731–0.908, p < 0.001 at score > 2) and the TI-RADS score (0.763 AUC, 95% CI: 0.668–0.858, p < 0.001 at score > 4). Univariate analysis showed hypoechogenicity, taller-than-wide shape, irregular or lobulated margins, and microcalcifications were strong predictors of malignancy (p < 0.001). Conclusion This study confirms that the combination of TI-RADS and Bethesda systems in preoperative thyroid nodule evaluation enhances diagnostic confidence and optimizes clinical decision-making. Ultrasound features, like taller-than-wide shape, irregular margin, hypoechogenicity, and microcalcifications, are significant indicators of thyroid malignancy, particularly PTC. Nondominant nodules in MNG may harbor neoplasms, necessitating multiple FNACs in specific cases.