Exploring the Performance of Ultrasound Risk Stratification Systems in Thyroid Nodules of Pediatric Patients

SIMPLE SUMMARY: Although pediatric thyroid nodules are uncommon, they need high clinical expertise and alert since they carry a greater risk of malignancy compared with those presenting in adults. Since there are no specific ultrasound (US)-based risk stratification systems (RSSs) for pediatric thyroid nodules, the application of adult-based RSSs in the pediatric population could represent a step forward in the care of children and adolescents with thyroid nodules. We compared the diagnostic performance of the main US-based RSSs *i.e., the American College of Radiology (ACR), European (EU), Korean (K) Thyroid Imaging Reporting and Data Systems (TI-RADSs) and ATA US RSS criteria) for detecting malignant thyroid lesions in pediatric patients. For ACR TI-RADS and EU-TIRADS, we found a sensitivity of 41.7%, and, for K-TIRADS and ATA US RSS, we found a sensitivity of 50%. The four US-based RSSs (i.e., ACR-TIRADS, EU-TIRADS, K-TIRADS, and ATA US RSS) have suboptimal performance in managing pediatric patients with thyroid nodules, with one-half of cancers without indication for FNA according to their recommendations. All thyroidologists, as well as the panelists of next TIRADSs, should be aware of these findings. ABSTRACT: Neck ultrasound (nUS) is the cornerstone of clinical management of thyroid nodules in pediatric patients, as well as adults. The current study was carried out to explore and compare the diagnostic performance of the main US-based risk stratification systems (RSSs) (i.e., the American College of Radiology (ACR), European (EU), Korean (K) TI-RADSs and ATA US RSS criteria) for detecting malignant thyroid lesions in pediatric patients. We conducted a retrospective analysis of consecutive children and adolescents who received a diagnosis of thyroid nodule. We included subjects with age <19 years having thyroid nodules with benign cytology/histology or final histological diagnosis. We excluded subjects with (a) a previous malignancy, (b) a history of radiation exposure, (c) cancer genetic susceptibility syndromes, (d) lymph nodes suspicious for metastases of thyroid cancer at nUS, (e) a family history of thyroid cancer, or (f) cytologically indeterminate nodules without histology and nodules with inadequate cytology. We included 41 nodules in 36 patients with median age 15 years (11–17 years). Of the 41 thyroid nodules, 29 (70.7%) were benign and 12 (29.3%) were malignant. For both ACR TI-RADS and EU-TIRADS, we found a sensitivity of 41.7%. Instead, for both K-TIRADS and ATA US RSS, we found a sensitivity of 50%. The missed malignancy rate for ACR-TIRADS and EU-TIRADS was 58.3%, while that for K-TIRADS and ATA US RSS was 50%. The unnecessary FNA prevalence for ACR TI-RADS and EU-TIRADS was 58.3%, while that for K-TIRADS and ATA US RSS was 76%. Our findings suggest that the four US-based RSSs (i.e., ACR-TIRADS, EU-TIRADS, K-TIRADS, and ATA US RSS) have suboptimal performance in managing pediatric patients with thyroid nodules, with one-half of cancers without indication for FNA according to their recommendations.