MON-525 ATA and ACR TIRADS Classification Systems as Additional Predictors of Malignancy in Afirma GSC Suspicious Thyroid Nodules

Abstract

Listen

Translate article

Objective: Evaluation and management of thyroid nodules with cytologically indeterminate results remain challenging in clinical practice. Despite the implementation of molecular testing in an attempt to avoid surgical intervention, diagnostic thyroidectomy still occurs due to the relatively low positive predictive value of these molecular testing. We conducted a study to analyze whether combining US characteristics and results of molecular testing would better elucidate predicting true positive results. Methods: We retrospectively reviewed thyroid ultrasound images of 172 nodules in 162 patients (mean age, 55 years +/- 14) with indeterminate cytology results (Bethesda III and IV) that underwent Afirma Gene Sequencing Classifier (GSC) testing at a single academic medical center between 2017–2019. All nodules were classified according to 2015 American Thyroid Association (ATA) and 2017 American College of Radiology Thyroid Imaging Reporting and Data System (ACR-TIRADS). Results: A total of 172 with subsequent Afirma GSC molecular testing were included in the study. There were 127 nodules with Bethesda III (AUS/FLUS) (73.8%), and 45 nodules with Bethesda IV (SFN/HCN) (26.2%) results. The mean nodule volume was 5.4 +/- 10 cm3. Afirma GSC identified 129 nodules (75%) as benign and 43 nodules (25%) as suspicious. Per ATA classification, 10.4% (18) of nodules were classified as very low risk, 40.7% (70) as low risk, 36.5% (62) as intermediate and 12.8% (22) as high risk for malignancy. There was a significant association between ATA classification and Afirma benign nodules (P=0.002). All nodules were also classified per TIRADS system with the following distribution: 5 (1.16%) TIRADS 1, 7 (4%) TIRADS 2, 54 (31%) TIRADS 3, 90 (52%) TIRADS 4, and 16 (9.3%) TIRADS 5. We did not observe the similar association between TIRADS system and benign nodules as we did with ATA classification (P=0.4).35 patients (79.5%) with Afirma suspicious results underwent surgery, of which 18 (51.4%) surgical pathology were malignant. 8 patients with Afirma suspicious results decided to proceed with ultrasound surveillance. The malignancy rates of nodules with low, intermediate and high suspicion for malignancy classified by the ATA guidelines were 44% (9 of 18), 33% (56 of 18) and 22% (4 of 18). The malignancy rates of TIRADS category 3, 4 and 5 nodules were 44% (8 of 18 nodules), 44% (8 of 18 nodules) and 11% (2 of 18 nodules). Subset analysis of surgical pathology benign and cancerous nodules did not show significant association between ATA (P=0.5) or TIRADS (P=0.4) classification systems. Conclusion: Our study showed that Afirma benign nodules were associated with a lower risk of malignancy per ATA classification but not with TIRADS system. We did not find a significant association between pathology proven cancer cases and high-risk ATA or TIRADS ultrasound classification systems.