Abstract
<b><i>Introduction:</i></b> The 6 categories of the Bethesda System for Reporting Thyroid Cytology (TBSRTC) with associated risk of malignancy (ROM) provide evidence-based clinical management guidelines. This study aimed to determine the ROM and accuracy of FNAB in South Africa (SA). <b><i>Methods:</i></b> Thyroid specimens from 3 pathology laboratories registered between January 2015 and December 2019 were considered for inclusion. ROM was obtained per TBSRTC category by cytohistological correlation and dividing the total number of specimens with malignant histology by the total number of cases operated. Accuracy was calculated based on the Bethesda category and eventual malignant histology. <b><i>Results:</i></b> Seventeen thousand seven hundred and seventy-three histology and 4,791 cytology cases were identified. Of the 4,791 cytology cases, 931 (19%) underwent surgery. More than a third (333, 35.8%) of cases were confirmed as malignant following histological assessment, with the majority being benign (584, 62.7%). The ROM for the nondiagnostic and benign categories was 24.3% and 20.5%. The highest ROM was for category VI (91.5%), followed by categories V (69.5%), IV (51.9%), and III (38.8%). Thyroid FNAB had a sensitivity of 73%, specificity of 74%, and overall accuracy of 74%. <b><i>Conclusion:</i></b> Bethesda categories II and IV have a relatively higher ROM in SA compared to findings from other developed countries. The diagnostic accuracy of thyroid FNAB in SA and the high rate of nondiagnostic diagnoses (38%) require further investigation. A national thyroid registry could provide location-specific data to aid the implementation of appropriate local policies and national guidelines for practicing thyroid surgeons.