254 chapter 2 89 known non-oncocytic indeterminate thyroid nodules, which we were able to extract from two studies. Prevalence of malignancy in this subgroup was 29.2% (26/89). Twenty-two nodules (24.7%) were [99mTc]Tc-MIBI positive. Estimated pooled sensitivity, specificity, positive and negative LR are 70% (95% CI: 52%-87%), 99% (95% CI: 96%-100%), 10.73 (95% CI: 0.71-162.31) and 0.36 (95% CI: 0.20-0.63), respectively (Table 61). These pooled results confirm that excluding Hürthle cell lesions improves the benign call rate, specificity and accuracy of [99mTc]Tc-MIBI. FDG-PET Our systematic literature search yielded eleven good-quality studies on the diagnostic value of [18F] FDG-PET [37, 58, 59, 303, 305-310, 312] (Table 62). Five studies investigated [18F]FDG-PET only [37, 305, 306, 310, 312], whereas the other six used combined [18F]FDG-PET/CT scans [58, 59, 303, 307-309]. Meta-analysis was performed of the ten studies that performed visual analysis of the [18F]FDG-PET scans, including a total of 434 indeterminate thyroid nodules. The study by Smith et al. was excluded because it reported results only by the means of ROC curve analysis for the AUC of the SUV value and summarized patient data could not be extracted [310]. We defined test positivity as visual focal uptake of [18F]FDG in the thyroid nodule, and disregarded any SUV thresholds, thereby reclassifying three nodules from one study: a 4.0 cm FVPTC with a SUV of 1.7 was reclassified as true positive; a focal thyroiditis (which included a coincidental finding of a 1 mm mPTC) and a hyperplastic nodule with SUVs of 1.8 and 1.7, respectively, were reclassified as false positive [308]. A negative index test was reported in 39.9% (173/434) of patients. I2 is 27.1% for sensitivity and 73.6% for specificity, suggesting the presence of between-study variability. The estimated pooled sensitivity, specificity, positive LR and negative LR are 94.2% (95% CI: 81.1%-98.4%), 46.5% (95% CI: 33.5%-59.9%), 1.76 (95% CI: 1.40-2.21) and 0.12 (95% CI: 0.04-0.40), respectively. The AUC is 0.80 (95% CI: 0.76-0.83) (Table 63 and Figure 84 and 85). For a given prevalence of malignancy of 15%, 25% or 40%, these results correspond to an estimated PPV and NPV of 23.7% (95% CI: 19.8%-28.1%) and 97.8% (95% CI: 93.5%-99.3%), 37.0% (95% CI: 31.8%-42.5%) and 96.0% (95% CI: 88.3%-98.7%), or 54.0% (95% CI: 48.3%-59.6%) and 92.3% (95% CI: 79.1%-97.5%), respectively (Figure 86: Bayesian probability plot). Subgroup analyses for Bethesda III and IV nodules could not be performed, as most studies included indeterminate cytology without further specification. Results of best-case and worst-case scenarios are identical to the presented results as histopathology was available in all cases. FDG-PET in thyroid nodules with Hürthle cell cytology Indeterminate nodules with Hürthle cell cytology are mostly [18F]FDG-positive. In our meta-analysis, 73 Hürthle cell nodules were included from eight studies (16.8%, 73/434). Histopathology was available in 71 nodules. A potentially clinically useful negative [18F]FDG-PET result was mentioned in merely 11 of the 73 Hürthle cell nodules (15% benign call rate). In comparison, 46.6% (109/234,
RkJQdWJsaXNoZXIy MTk4NDMw