181 Diagnostic utility of molecular and imaging biomarkers 2 the above is debatable, as the exact extent of the influence of the GEC test results, clinical suspicions or the patients’ wishes on surgical management decisions is unknown. The current meta-analysis is not a sound reflection of the GEC’s true performance, but a mere realistic display of the current clinical situation in the USA where the GEC already seems to be an accepted standard. We constructed bestcase and worst-case scenarios. Nondiagnostic GEC results were disregarded. In the hypothetical bestcase scenario (i.e., the test was always right), the estimated pooled sensitivity, specificity, positive and negative LR would be 98.2% (95% CI: 95.5%-99.3%), 56.5% (95% CI: 45.4%-67.0%), 2.26 (95% CI: 1.75-2.91) and 0.03 (95% CI: 0.01-0.09), respectively (Table 38, Figure 4, Figure 45, Figure 46). In the hypothetical worst-case scenario, the estimated pooled sensitivity, specificity, positive and negative LR would be 42.5% (95% CI: 27.9%-58.4%), 10.4% (95% CI: 6.3%-16.6%), 0.47 (95% CI: 0.32-0.71) and 5.54 (95% CI: 2.86-10.76), respectively (Table 39, Figure 4, Figure 47 and Figure 48). Due to a very high rate of unavailable histopathology for the GEC, both scenarios are extreme. By exception we subsequently also constructed a more realistic, ‘second-best-case’ scenario. We hypothesized that the population without histopathological confirmation – assumed benign in most studies regardless of the GEC result – harboured a frequency of missed malignancies similar to at least the false-negative rate in Alexander et al.’s first study with full histopathological validation: 6% [164]. We recalculated the summarized patient data, assuming carcinoma existed in 6% of the unoperated GEC-positive nodules (truepositive) and 6% the unoperated GEC-negative nodules (false-negative). All other nodules without histopathological correlation were considered benign regardless of the Afirma® GEC test result. In this second-best-case scenario, estimated pooled sensitivity, specificity, positive and negative LR would be 89.3% (95% CI: 84.7%-92.6%), 46.7% (95% CI: 37.1%-56.4%), 1.67 (95% CI: 1.42-1.97) and 0.23 (95% CI: 0.17-0.31), respectively (Table 40, Figure 49 and Figure 50). Figure 44. SROC curve of the Afirma® GEC – Bethesda IV Summary receiver operating characteristic plot showing sensitivity versus 1-specificity of the Afirma® GEC in Bethesda IV thyroid nodules with available histopathology. AUC = 0.92 (95% CI: 0.89-0.94). AUC, area under the curve; HSROC, hierarchical summary receiver operating characteristic.
RkJQdWJsaXNoZXIy MTk4NDMw