Diagnostic Performance of Practice Guidelines for Thyroid Nodules: Thyroid Nodule Size versus Biopsy Rates

Abstract

Background The diagnostic performance and unnecessary biopsy rates for evaluation of thyroid nodules varies among the current society guidelines. We hypothesized that diagnostic performance and unnecessary biopsy rates of different guidelines are largely determined by different thyroid size thresholds used as cutoff values that determine when to biopsy a thyroid nodule. Purpose To compare the malignancy risk, diagnostic performance, and unnecessary biopsy rates of three guidelines and evaluate the changes according to the nodule size cutoff for biopsy. Materials and Methods This retrospective study included 3323 consecutive thyroid nodules from 3190 patients between January 2013 and December 2013. US features were reviewed according to the American Thyroid Association (ATA), Korean Thyroid Association/Korean Society of Thyroid Radiology (KTA/KSThR), and American College of Radiology (ACR) guidelines. The diagnostic performance and unnecessary biopsy rates were calculated. The size cutoff of ATA and KTA/KSThR guidelines was varied and diagnostic performance and unnecessary biopsy rates were calculated and compared by using the generalized estimating equations method. Results A total of 856 (25.8%) thyroid nodules were diagnosed as malignant after surgical excision or biopsy. The highest sensitivity was observed with the KTA/KSThR guideline (93.5% [300 of 321]), followed by the ATA (92.5% [297 of 321]) and ACR guidelines (60.1% [193 of 321]). The highest specificity (75.2% [1216 of 1617], P < .001) with lowest unnecessary biopsy rate (20.7% [401 of 1938]) was observed with the ACR guideline. When the low-suspicion category in the ATA guideline modeled 2.5 cm instead of 1.5 cm as the size for biopsy, the unnecessary biopsy rate decreased to 43.6% (845 of 1938, 2.5-cm cutoff) compared with 55.1% (1068 of 1938, 1.5-cm cutoff) (P < .001). For the KTA/KSThR guidelines, the unnecessary biopsy rate decreased to 36.4% (706 of 1938, 2.5-cm cutoff) compared with 59.5% (1153 of 1938, 1.5-cm cutoff) (P < .001). Conclusion The diagnostic performance and unnecessary biopsy rates of the American Thyroid Association and the Korean Thyroid Association/Korean Society of Thyroid Radiology guidelines are strongly influenced by nodule size cutoff. © RSNA, 2019 Online supplemental material is available for this article. See also the editorial by Foley in this issue.