169 Diagnostic utility of molecular and imaging biomarkers 2 Next Generation Sequencing Our systematic literature search yielded three articles on NGS in indeterminate thyroid nodules, including a total of 639 indeterminate thyroid nodules with full availability of histopathology. Baseline characteristics are presented in Table 31. The average malignancy rate was 10.6% (range 5%-27%). All three studies applied a slightly different multi-gene NGS panel. Consequently, meta-analysis was deemed inappropriate and not performed. Individual study results are reported in Table 32. Afirma® GEC Our systematic literature search yielded 13 original studies on the Afirma GEC. All studies applied the BSRTC and were conducted in the USA between 2014 and 2016 in community as well as tertiary hospitals [50, 164, 165, 167-173, 175, 180, 182] (Table 33). None of the studies published after Alexander et al. had complete histopathologic follow-up; histopathological confirmation ranged between 35-82% of specimens (Table 34) [164]. We included 13 studies with a total of 1,746 indeterminate thyroid nodules in our meta-analysis [50, 164, 165, 167-173, 175, 180, 182]. The prevalence of malignancy was 20.3% (355/1,746), including at least 86 PTC (24%), 57 FVPTC (16%), 11 FTC (3.1%), 11 FTC-OV (3.1%), two MTC (0.6%), three mPTC (0.8%) and two other types of thyroid malignancy (0.6%). Four studies did not report the subtype of a total of 196 thyroid carcinoma [164, 167, 173, 175]. Afirma® GEC test results were positive (‘GEC suspicious’) in 996 (57.0%) and negative (‘GEC benign’) in 669 (38.3%). The remaining 81 nodules (4.6%) had a nondiagnostic GEC test result, or no result due to issues with storage or shipment of the samples, for example, and were excluded from our analysis (Table 34). Histopathology was available in 57.5% (958/1,665, range 38%-100%) of the nodules with a conclusive GEC result: 80.8% (805/996, range 66%-100%) of the GEC-positive and 22.9% (153/669, range 0%-100%) of the GECnegative indeterminate thyroid nodules (Pearson chi-squared, p<0.0001). A false-negative GEC result was found in 13 nodules: five PTC [164, 173, 180], two FVPTC [164, 173], one FTC [168], one FTC-OV [164], two mPTC [165, 180], and two unreported malignancies [175]. The individual study sensitivities and specificities of the Afirma® GEC in indeterminate thyroid nodules ranged from 83% to 100% and 0% to 52%, respectively. Visual inspection of the forest plots suggests between study heterogeneity regarding test specificity. I2 is 25.0% for sensitivity and 92.6% for specificity. Pooled sensitivity, specificity, positive and negative LR are 97.3% (95% CI: 94.2%-98.8%), 15.0% (95% CI: 9.5%-22.9%), 1.15 (95% CI: 1.06-1.23) and 0.18 (95% CI: 0.09-0.37), respectively. The AUC is 0.85 (95% CI: 0.82-0.88) (Table 35, Figure 38 and Figure 39). For a given prevalence of malignancy of 15%, 25% or 40%, these results correspond to an estimated PPV and NPV of 16.8% (95% CI: 15.8%-17.9%) and 96.9% (95% CI: 93.8%-98.5%), 27.6% (95% CI: 26.2%-29.1%) and 94.3% (95% CI: 88.9%-97.2%), or 43.3% (95% CI: 41.5%-45.1%) and 89.3% (95% CI: 80.0%-94.5%), respectively (Figure 40).
RkJQdWJsaXNoZXIy MTk4NDMw