Clinical-Ultrasound Nomogram for Malignancy Prediction in Thyroid Nodules with Follicular Neoplasm-Like Sonographic Features: Comparison with Four International TIRADS Systems

Background Thyroid nodules with follicular neoplasm-like ultrasound features (FNLUF-TNs) are frequently misdiagnosed due to their inconspicuous malignant manifestations. We aimed to develop and validate a clinical-ultrasound nomogram for predicting malignancy risk in FNLUF-TNs, and to compare its diagnostic performance with conventional TIRADS. Methods This retrospective study included patients with 458 FNLUF-TNs (defined as oval, solid/predominantly solid, well-circumscribed nodules) who underwent preoperative ultrasound and surgical resection. Univariate and multivariate analyses were conducted to screen out independent predictors of benign and malignant nodules, and a nomogram model based on clinical and ultrasonic features was constructed. The diagnostic performance of the nomogram was assessed by calculating sensitivity, specificity, accuracy, and AUC, and further compared with the corresponding performance of four TIRADS. Calibration, clinical utility, and interobserver agreement were also evaluated. Results Malignancy was confirmed in 127 of 458 nodules (27.7%). Independent predictors of malignancy included younger age, male gender, smaller transverse diameter, hypoechogenicity, calcification, trabecular formation, and irregular interrupted halo (all p < 0.05). The nomogram achieved an AUC of 0.824 (95% CI 0.786–0.858), accuracy of 87.55%, specificity of 93.96%, significantly outperforming all four TIRADS systems (all p < 0.001). Calibration curves showed good consistency between predicted and actual malignancy risk. Decision curve analysis demonstrated favorable net clinical benefit. Interobserver agreement for key ultrasound features was excellent (all κ > 0.81). Conclusion This clinical ultrasound nomogram provides accurate and individualized malignancy risk stratification for FNLUF-TNs, outperforming conventional TIRADS in diagnostic performance while showing good reproducibility and clinical value, which may help optimize management and minimize unnecessary interventions for this challenging subgroup. Trial registration: Not applicable.