Adoption of ACR TI-RADS and Documentation of Biopsy Criteria in Thyroid Ultrasound Reports: A Seven-Year Analysis of a National Radiology Network.
Oct 29, 2026DOI: 10.1016/j.jacr.2026.10.027Journal ArticleJACR
Written by: Mohadese AhmadzadeGary Lloyd HornMohammad Ghasemi-RadPhil RamisLeelakrishna NallamshettyZachary Nuffer
OBJECTIVE: To evaluate longitudinal adoption of American College of Radiology (ACR) Thyroid Imaging Reporting and Data System (TI-RADS) Terminology, report-level documentation of Fine Needle Aspiration (FNA) criteria, and associated ultrasound-guided FNA practice in a large multisite radiology network.METHODS: In this retrospective, multicenter study across 89 imaging sites, 304,455 diagnostic thyroid ultrasound reports (CPT 76536) and 11,246 ultrasound-guided FNA reports (CPT 10005/10006) from 2017 through 2023 were analyzed. A linked pre-biopsy cohort of 994 diagnostic reports immediately preceding FNA was examined. Automated text mining identified TI-RADS language and documentation of nodules meeting ACR size-based FNA thresholds. Trends were assessed using Cochran-Armitage, Spearman correlation, and χ2 tests.RESULTS: TI-RADS language increased from 1.6% of thyroid ultrasound reports in 2017 to 54.1% in 2023. Among 112,036 TI-RADS-containing reports, 11,845 (10.6%) documented nodules meeting FNA criteria. This proportion declined from 23.3% in 2017 to 8.0% in 2023, an absolute decline of 15.3 percentage points (Z = -47.94, P < 0.001) and was inversely correlated with annual TI-RADS adoption across the seven annual observations (ρ = -1.000, P < 0.001). Among FNA-indicated reports, TR4 predominated (56.9%), followed by TR3 (21.0%) and TR5 (19.4%); category distribution shifted significantly over time, although the effect size was small (χ2 = 169.91, P < 0.001; Cramér's V=0.086). In the pre-biopsy cohort (n=994), 498 (50.1%) contained TI-RADS language. The FNA rate per 1,000 thyroid ultrasound examination remained stable (29.1 in 2017; 27.7 in 2023).CONCLUSION: ACR TI-RADS terminology use increased substantially but remained incomplete. Increasing terminology use was associated with declining report-level FNA eligibility. Biopsy appropriateness could not be assessed from report-level documentation alone.
