Context: Ultrasound (US) risk stratification systems (RSSs) have been developed to reduce the number of unnecessary fine-needle aspiration procedures (FNA) in patients with thyroid nodules.Objective: We conducted a systematic review and meta-analysis evaluating the ability of the 5 most common US RSSs for the appropriate selection of thyroid nodules for FNA.Data sources: This systematic review and meta-analysis was registered on PROSPERO (CRD42019131771). PubMed, CENTRAL, Scopus, and Web of Science were searched until March 2019.Study selection: Original articles reporting data on the performance of AACE/ACE/AME, ACR TI-RADS, ATA, EU-TIRADS, and K-TIRADS were included.Data extraction: The number of nodules classified as true negative, true positive, false negative, and false positive was extracted. Summary operating points were estimated using a random-effects model. Interobserver agreement was also assessed.Data synthesis: Twelve studies evaluating 18 750 thyroid nodules were included. Participants were adult outpatients with thyroid nodules submitted to either FNA or core-needle biopsy or surgery and with available US images. The final diagnosis for malignant nodules was generally based on histology, while cytology was used for benign nodules. Diagnostic odds ratio (DOR) ranged from 2.2 to 4.9. A head-to-head comparison showed a higher relative DOR for ACR-TIRADS versus ATA (P = .002) or K-TIRADS (P = .002), due to a higher relative likelihood ratio for positive results.Conclusions: The present meta-analysis found a higher performance of ACR TI-RADS in selecting thyroid nodules for FNA. However, the comparison across the most common US RSSs was limited by the data available. Further studies are needed to confirm this finding.

Thyroid imaging reporting and data systems (TIRADS) are used to stratify the malignancy risk of thyroid nodule by ultrasound (US) examination. We conducted a meta-analysis to evaluate the pooled cancer prevalence and the relative prevalence of papillary, medullary, follicular thyroid cancer (PTC, MTC, and FTC) and other malignancies among nodules included in studies evaluating their performance. Four databases were searched until February 2020. Original articles with at least 1000 nodules, evaluating the performance of at least one TIRADS among AACE/ACE/AME, ACR-TIRADS, ATA, EU-TIRADS, or K-TIRADS, and reporting data on the histological diagnosis of malignant lesions were included. The number of malignant nodules, PTC, FTC, MTC and other malignancies in each study was extracted. For statistical pooling of data, a random-effects model was used. Nine studies were included, evaluating 19,494 thyroid nodules. The overall prevalence of malignancy was 34% (95%CI 21 to 49). Among 6162 histologically proven malignancies, the prevalence of PTC, FTC, MTC and other malignancies was 95%, 2%, 1%, and 1%, respectively. A high heterogeneity was found for all the outcomes. A limited number of studies generally conducted using a retrospective design was found, with possible selection bias. Acknowledging this limitation, TIRADSs should be regarded as accurate tools to diagnose PTC only. Proposed patterns and/or cut-offs should be revised and other strategies considered to improve their performance in the assessment of FTC, MTC and other malignancies.