Predicting malignancy in thyroid AUS: ultrasonographic and clinical risk factors beyond ACR TI-RADS

Thyroid cancers often present as nodules, with fine-needle aspiration biopsy (FNAB) and ultrasonography as the primary diagnostic tools. Given the variability in clinical approaches and the lack of consensus guidelines, ultrasonography has become increasingly important, particularly for indeterminate cytology. We aim to assess the clinical utility and diagnostic performance of the ACR TI-RADS (American College of Radiology Thyroid Imaging Reporting and Data System), together with cytopathological subtypes, in predicting malignancy in atypia of undetermined significance (AUS) cases.

In this prospective, single-centre study conducted over 30 months, all AUS cases underwent preoperative ultrasonographic evaluation using the ACR TI-RADS, and had postoperative histopathological results.

Seventy-one nodules from 67 patients were included. The median age was 51 years (range: 12-71). The overall malignancy rate was 20%. Among the nodules, 26 (37%) were classified as the nuclear atypia, and 45 (63%) as other atypia. The ACR TI-RADS score had a strong association with malignancy (p = 0.025). The presence and type of echogenic foci were associated with malignancy (p = 0.001). The malignancy rates for TR3, TR4, and TR5 nodules were 21%, 8%, and 55% respectively.

The ACR TI-RADS was significantly associated with malignancy in both the overall AUS cohort and the nuclear atypia subgroup. The relatively high number of follicular thyroid carcinoma cases may have masked the predictive value of ultrasonography. The presence of peripheral calcifications was significantly associated with malignancy. Although the sample size was adequate by power analysis, larger cohorts would further strengthen these findings.