LAKE BUENA VISTA, FLA. – A scoring system for ultrasound evaluation of thyroid nodules is useful in predicting malignancy in cases of atypia of unknown significance but is less useful for nodules categorized as follicular lesions of undetermined significance.
Dr. Jung Hyun Soon, a radiologist at Severance Hospital, Yonsei University, Seoul, presented these findings at the International Thyroid Congress.
The atypia of unknown significance (AUS) and follicular lesion of undetermined significance (FLUS) subcategories within the Bethesda cytopathology system carry a 5%-15% risk of malignancy. The Bethesda system for assessing fine-needle aspirate (FNA) helps to identify more precisely “the different patterns of atypia associated with higher malignancy risk, and to enable adequate malignancy risk stratification of thyroid nodules,” said Dr. Yoon.
Ultrasonography also contributes to the prediction of malignancy, and the Thyroid Imaging Reporting and Data System (TIRADS) provides a framework to report and score imaging assessment of thyroid nodules. The number of suspicious features is tallied and a TIRADS score is assigned according to the sum: possible scores are 3, 4a, 4b, 4c, or 5, according to number of abnormalities. Categories of abnormalities that are scored include composition, echogenicity, margin characteristics, calcifications, and shape.
To determine whether TIRADS helps predict which patients with AUS and FLUS will have malignancy, Dr. Yoon and her colleagues enrolled 188 patients with a total 192 AUS or FLUS thyroid nodules whose diagnoses had been confirmed by surgery, sequential ultrasound, or repeat FNA. The patients’ mean age was 50.2 years, and the mean nodule size was 14.7 mm, though nodules ranged from 5 to 60 mm.
Of the 192 AUS/FLUS nodules, 82 (42.7%) were malignant and 110 (57.3%) were benign. Of the 149 nodules (77.6%) that were characterized as AUS, 45.6% were assessed as malignant, while the 43 FLUS nodules (22.4%) had a malignancy rate of 32.6%. This difference in malignancy rates was not statistically significant.
In applying TIRADS scoring to the AUS cytological subcategory, significant differences were seen in TIRADS ratings between malignant and benign nodules (P less than .001). As TIRADS scores increased, so did the likelihood of malignancy: only 15.4% of TIRADS 3 AUS nodules were malignant, while, for TIRADS 5 AUS nodules, the probability of malignancy was 80%. Intermediate values carried increasing risk. These differences were not seen, however, for FLUS nodules (P = .414).
“Suspicious ultrasound features are useful in predicting malignancy among the AUS subcategory but not in the FLUS subcategory,” Dr. Yoon said in summarizing her findings at the meeting held by the American Thyroid Association, Asia-Oceania Thyroid Association, European Thyroid Association, and Latin American Thyroid Society.
She noted several study limitations, including the inherent selection bias in the study’s retrospective nature and the fact that AUS and FLUS are not universally accepted subcategories. Additionally, uniform diagnosis could not be assured since five cytopathologists made the original diagnoses from the FNA samples.
Dr. Yoon and her investigators reported no relevant disclosures.
On Twitter @karioakes