Abstract
PURPOSE: Thyroid nodules are common clinical findings, with a subset requiring evaluation for malignancy. TIR 3B nodules, classified as indeterminate with high risk, present a diagnostic challenge. This study aims to identify clinical, ultrasonographic, and cytological features predictive of malignancy in TIR 3B thyroid nodules. METHODS: We retrospectively analyzed 175 patients with TIR 3B nodules who underwent surgery, evaluating their clinical, ultrasound, and cytological characteristics. RESULTS: The malignancy rate was 43.4%, with papillary thyroid carcinoma being the most prevalent. Multivariable analysis identified solid hypoechoic nodules, irregular margins, nuclear grooves, and micro- and macrocalcifications as independent predictors of malignancy. Based on these features, we developed a predictive nomogram model with an area under the curve (AUC) of 0.790, demonstrating good diagnostic performance. The model identified a malignancy risk cut-off score of 8, with high specificity (94%) for distinguishing benign from malignant nodules. CONCLUSIONS: This study highlighted several findings regarding malignancy risk and identified ultrasonographic and cytological features associated with malignant outcomes to help determine which patients should be candidates for surgery. The proposed predictive model could be a valuable clinical tool and may facilitate more personalized treatment approaches.