112 chapter 2 In the entire pooled population, eight nodules tested false-positive and proved histopathologically benign: two follicular adenomas [118, 121], one Hürthle cell adenoma (oncocytic variant of follicular adenoma, or FA-OV) [121], one nodular hyperplasia [72], one lymphocytic thyroiditis [119], one adenomatous hyperplasia [116], and two unspecified benign lesions [111, 120]. Most of these nodules were from South Korean studies that attributed their false-positive results to the technique they used for mutation analysis. All but two of the studies used similarly sensitive techniques such as real-time polymerase chain reaction (PCR), dual priming oligonucleotide PCR (DPO-PCR) or pyrosequencing [72, 77, 111, 116, 119, 121]. The high sensitivity of these methods likely comes at the expense of a small number of false-positive results and impairs the strong specificity [116, 119, 121]. Notable, Hyeon et al. used direct sequencing, a method known for its high specificity, but still found one false-positive BRAF mutation in a benign lesion [120]. BRAF mutation analysis in Bethesda III nodules Separate meta-analysis was performed for nodules with Bethesda III cytology. In total, 3,186 Bethesda III nodules were included from 24 studies [67-69, 72, 74, 77, 79, 89, 92, 95-98, 101, 104, 111-114, 116, 118-121]. A BRAF mutation was found in 12.3% (392/3,186). Surgery was performed in only 44.4% (1,410/3,176) of the Bethesda III nodules with a conclusive index test result, including 82.4% (323/392) of the BRAF mutation-positive ones. The malignancy rate was 19.7% (629/3,186), and at least 51.2% (322/629) of which were proven PTC (164 Bethesda III nodules harboured an unspecified malignancy). Estimated pooled sensitivity, specificity, positive LR, negative LR and AUC are 34.2% (95% CI: 21.6%-49.5%), 99.6% (95% CI: 97.2%-99.9%), 89.04 (95% CI: 13.31-595.79), 0.66 (95% CI: 0.53-0.82) and 0.92 (95% CI: 0.89-0.94), respectively (Table 6, Figure 8 and Figure 9). FN: false negative. FP: false positive. LR+: positive likelihood ratio. LR-: negative likelihood ratio. nH: number of indeterminate thyroid nodules with conclusive index test and histopathological follow-up. ni: number of indeterminate thyroid nodules. test+: number of positive index tests. test-: number of negative index tests. test?: number of unknown or nondiagnostic index tests. TN: true negative. TP: true positive. Our results capture the clear association between the BRAF mutation and PTC: 72.5% (308/425) of the histopathology confirmed BRAF mutation-positive nodules were confirmed as PTC and 60.4% (308/510) of PTC were BRAF mutation-positive. The overall poor sensitivity of BRAF mutation analysis is explained by the large fraction of other types of thyroid carcinoma in indeterminate nodules, which were infrequently BRAF mutation-positive: only 37 of 318 FVPTC (12%) were BRAF mutationpositive, 1 of 32 Hürthle cell carcinoma (oncocytic variant of follicular thyroid carcinoma, FTC-OV) (3.1%), 3 of 12 mPTC (25%), and none of the 143 FTC, seven medullary thyroid carcinoma (MTC) and 11 other thyroid carcinomas (e.g. poorly differentiated carcinoma) (Table 5). Eleven studies did not correlate presence of a BRAF mutation to the distinctive subtypes of thyroid carcinoma in (a large part of) their study population; 68 of these 262 unspecified malignancies were BRAF mutationpositive [60, 93, 98, 101, 107, 110, 112, 116, 118-120]. The prevalence of BRAF mutations in FVPTC was significantly lower than in (the classical variant of) PTC (Pearson χ2, p<0.0001). Of the 16 BRAFK601E mutations found in five studies, eight mutations were found in FVPTC, one in follicular adenoma and seven in thyroid carcinoma of unreported subtype [60, 75-77, 118].
RkJQdWJsaXNoZXIy MTk4NDMw