BACKGROUND. Risk stratification systems for evaluating thyroid nodules on ultrasound use varying approaches to classify levels of suspicion for malignancy, leading to variable performance. OBJECTIVE. The purpose of this study was to perform a network meta-analysis comparing six risk stratification systems used to evaluate thyroid nodules on ultrasound in terms of their diagnostic performance for the detection of thyroid cancer. EVIDENCE ACQUISITION. Five bibliometric databases were searched for studies published through August 31, 2022, that compared at least two of six ultrasound risk stratification systems (the American Association of Clinical Endocrinologists, American College of Endocrinology, and Associazione Medici Endocrinologi [AACE/ACE/AME] system; American College of Radiology Thyroid Imaging Reporting and Data System [ACR TI-RADS]; the American Thyroid Association [ATA] risk stratification system; European Thyroid Association Thyroid Imaging Reporting and Data System [EU-TIRADS]; the Korean Thyroid Imaging Reporting and Data System [K-TIRADS] endorsed by the Korean Thyroid Association and the Korean Society of Thyroid Radiology; and the Thyroid Imaging Reporting and Data System developed by Kwak et al. [Kwak TIRADS]) in terms of their diagnostic performance for the detection of thyroid cancer, with cytologic or histologic evaluation used as a reference standard. The studies' risk of bias was evaluated using the Newcastle-Ottawa Scale. A meta-analysis of each system was performed to identify the risk category threshold that had the highest accuracy as well as the highest sensitivity and specificity at this threshold. Network meta-analysis was used to perform hierarchic ranking and identify the systems having the highest sensitivities and specificities at each system's most accurate threshold. EVIDENCE SYNTHESIS. The analysis included 39 studies with 49,661 patients. All studies were of fair (n = 17) or good (n = 22) quality. The most accurate risk category thresholds were class 3 (high risk) for the AACE/ACE/AME system, TR5 (highly suspicious) for ACR TI-RADS, EU-TIRADS 5 (high risk) for EU-TIRADS, 4c (moderate concern but not classic for malignancy) for Kwak TIRADS, K-TIRADS 5 (high suspicion) for K-TIRADS, and high suspicion for the ATA system. At these thresholds, the systems had sensitivity of 64-77% and specificity of 82-90%. Network meta-analysis identified the highest sensitivity and highest specificity for ACR TI-RADS, followed by K-TIRADS. CONCLUSION. Of six risk stratification systems, ACR TI-RADS had the highest diagnostic performance for the detection of thyroid nodules on ultrasound. CLINICAL IMPACT. This network meta-analysis can inform decisions regarding implementation of the risk stratification systems and can aid future system updates.
Keywords: biopsy; diagnostic imaging; fine-needle; network meta-analysis; thyroid neoplasms; thyroid nodule.