ACR’s TI-RADS criteria improve radiologist recommendations for thyroid nodule management

The American College of Radiology (ACR) Thyroid Imaging Reporting and Data System (TI-RADS) criteria for thyroid nodule evaluation can reduce the number of thyroid nodules radiologists recommend for biopsy, according to a new study published in Radiology. The guidelines were also associated with improving the overall accuracy of radiologist recommendations for nodule management.

“The aim of this study was to compare the biopsy rate and diagnostic accuracy before and after applying ACR TI-RADS risk stratification,” wrote Jenny K. Hoang, MBBS, MHS, with the department of radiology at Duke University Medical Center in Durham, North Carolina, and colleagues. “We hypothesized that there would be a reduction in recommendations for biopsy among radiologists with application of ACR TI-RADS compared with their current practice patterns.”

The authors studied how eight radiologists with varying years of experience reviewed ultrasound (US) features of 100 nodules in 92 patients from December 2016. Radiologists relied on their own existing practice patterns. The ACR TI-RADS criteria were then applied retrospectively to the US features, with the researchers comparing biopsy rates, sensitivity, specificity and accuracy.

Overall, the mean number of nodules radiologists recommended for biopsy was 80 when using their own practice patterns. That number dropped to 57 when applying ACR TI-RADS criteria, a reduction of 29 percent.

Also, readers had an overall sensitivity of 95 percent, an overall specificity of 20 percent and an overall accuracy of 28 percent when using their own practice patterns. After applying ACR TI-RADS criteria, sensitivity dropped slightly to 92 percent, but specificity (44 percent) and accuracy (52 percent) increased significantly.

“Although the sensitivity of test readers was slightly lower with ACR TI-RADS guidelines (92 percent) than with readers’ own practice patterns (95 percent), an important consideration is that many higher-suspicion nodules that did not meet the criteria for biopsy with ACR TI-RADS guidelines were recommended for follow-up US,” the authors wrote. “ACR TI-RADS provides specific follow-up guidelines for TR3, TR4, and TR5 nodules based on a lower size threshold than that used for biopsy. Given the indolent behavior of small thyroid cancers, observation of small suspicious nodules is a safe strategy.”

Hoang and colleagues noted that all but two malignancies were recommended for biopsy, and those were missed due to reading errors. “Two malignancies were incorrectly interpreted as spongiform and mixed cystic and solid nodules by two of the eight test readers but were regarded to have solid composition by all other readers and by the experts,” the authors wrote. “This suggests that further education should be directed toward the distinction between solid nodules (including those with a few cysts), spongiform nodules and mixed cystic and solid nodules.”