Diagnostic accuracy of droplet digital PCR for BRAF mutation detection in thyroid specimens: a systematic review and meta-analysis

BRAF mutation testing, particularly for BRAF V600E, may provide adjunctive diagnostic information in the evaluation of thyroid nodules when interpreted alongside cytology, ultrasonography, and clinical risk. Droplet digital polymerase chain reaction (ddPCR) enables sensitive detection and quantification of low-variant-allele-frequency mutations in limited or heterogeneous specimens. However, its clinical diagnostic performance across thyroid specimen types has not been systematically quantified, and the influence of heterogeneous reference standards and clinical settings remains uncertain. We searched MEDLINE (PubMed), Scopus, Embase, Web of Science, and Google Scholar from inception through 26 July 2026. Eligible studies evaluated ddPCR for BRAF mutation detection in thyroid specimens. Random-effects generalized linear mixed models (GLMMs) were used to pool sensitivity and specificity. Additional analyses included a hierarchical summary receiver operating characteristic (HSROC) model, Deeks’ regression, prevalence-based predictive-value analyses, sensitivity analyses, and specimen-specific subgroup analyses. Twenty-two studies were included in the qualitative synthesis, and 11 studies comprising 1,403 units of analysis contributed to the meta-analysis. Pooled sensitivity was 72.2% (95% confidence interval [CI], 54.8%-84.8%), and pooled specificity was 99.9% (95%CI, 87.2%-100%). HSROC analysis yielded an area under the curve of 0.947. Histopathology-restricted and specimen-specific analyses yielded similarly high specificity estimates, whereas sensitivity remained heterogeneous. ddPCR yielded a high pooled specificity estimate but moderate and heterogeneous sensitivity for BRAF detection across thyroid specimens. These findings suggest a potential adjunctive rule-in role in selected clinical or cytologically indeterminate settings. However, negative results cannot exclude malignancy and should be interpreted alongside cytological, ultrasonographic, and clinical findings. • ddPCR yielded a high pooled specificity estimate for BRAF detection, whereas sensitivity was moderate and substantially heterogeneous across studies. • Wide prediction intervals, heterogeneous reference standards, and sparse false-positive events limit the generalizability of the pooled estimates. • ddPCR may provide adjunctive diagnostic information in selected clinical contexts, but it should not be used as a standalone test, and a negative result cannot exclude malignancy.