Modified Bethesda criteria for thyroid aspirates significantly decrease nondiagnostic rates without decreasing sensitivity

Abstract

Previous studies suggest that the adequacy rate of thyroid aspirates can be improved by altering the adequacy criteria of the Bethesda System. We sought to measure the performance of these altered criteria in a prospective fashion. Over a 6-year period, cases with 1 to 59 follicular cells were prospectively classified as "nondiagnostic, favor benign" or "scant but adequate". "Scant but adequate" cases were classified as either benign (Bethesda category 2) or atypia of undetermined significance (AUS) (Bethesda category 3). Bethesda category 3 cases were referred for Afirma testing (Veracyte, San Francisco, CA). Of 5147 cases, 131 (3%) were classified as "nondiagnostic, favor benign"; 45 (65%) of these had follow-up with a risk of malignancy of 2.6%. Additionally, 436 (8%) of all 5147 cases were classified as "scant but adequate" and "benign"; 49 (11%) of these had follow-up with a risk of malignancy of 0%. Lastly, 197 (4%) of all 5147 cases were classified as "scant but adequate" with AUS; 177 (90%) of these 197 cases had an adequate Afirma result. The "suspicious" rate was not significantly different than that of cases classified as "adequate" and AUS (Bethesda category 3 and 4) (35 of 197 [18%] versus 140 of 848 [17%] P = 0.67), and there was no significant difference in the risk of malignancy for these 2 categories ("scant but adequate" 9 of 18, "adequate" 50% versus 27 of 85, 32%, P = 0.10). Overall, the modified Bethesda criteria reduced the nondiagnostic rate from 22% to 10% (P <0.001) without lowering the sensitivity of the test. Modified Bethesda adequacy criteria can significantly lower nondiagnostic rates without lowering sensitivity.