Diagnostic Performance of Ultrasound vs Ultrasound-Guided FNAC in Thyroid Nodules: Data From the ElaTION Trial.

Abstract

ElaTION is a large multicenter pragmatic randomized controlled trial, performed in 18 secondary/tertiary hospitals across England, comparing elastography ultrasound-guided fine needle aspiration cytology (EUS-FNAC) with ultrasound-guided FNAC (US-FNAC) alone in the diagnostic assessment of thyroid nodules. Secondary trial outcomes, reported here, assessed the accuracy of ultrasound alone (US) compared with US-FNAC to inform and update current practice guidelines. Adults with single or multiple thyroid nodules who had not undergone previous FNAC were eligible. Radiologists assessed all thyroid nodules using US alone, thereby enabling assessment of its accuracy (sensitivity and specificity) vs US-FNAC. Of the 982 participants, a final definitive diagnosis was obtained in 688, who were included in the final analyses. The sensitivity of US alone was the same as US-FNAC (0.91 [95% CI, 0.85-0.97] vs 0.87 [95% CI, 0.80-0.95] P = .37). US alone had statistically significant lower specificity than US-FNAC alone (0.48 vs 0.67 respectively, P < .0001). The malignancy rate on histology in a nodule classified as benign on ultrasound (U2) was 9/263 (3.42%) and on cytology (Thy2) was 15/353 (4.25%), whereas the malignancy rate in a nodule that was benign on both (U2, Thy2) was 3/210 (1.43%). Malignancy risk for U3, U4, and U5 nodules was 68/304 (22.4%), 43/83 (51.8%), and 29/38 (76.3%), respectively (P < .0001). Yet 80/982 (8%) patients were discharged despite having U3-U5 scans with Thy1 (nondiagnostic) FNAC and no definitive diagnosis.Malignancy risk was higher in smaller nodules: < 10 mm 23/60 (38.3%), 10-20 mm 46/162 (28.4%), and >20 mm 80/466 (17.2%) (P < .0001). Nodules with indeterminate cytology with atypical features (Thy3a) carried a similar malignancy risk to those with indeterminate cytology (Thy3/3f): 27/95 (28.4%) vs 42/113 (37.2%) respectively (P = .18). Ultrasound alone appears to be an effective diagnostic modality in thyroid nodules, confirming the recommendations of recent guidelines and the British Thyroid Association classification. However, findings also suggest caution regarding existing recommendations for conservative management of nondiagnostic (Thy1/Bethesda I) and atypical (Thy3a/Bethesda III) nodules. In those cases, ultrasound (U3-U5) features may help identify high-risk subgroups for more proactive management.