肺結節リスク層別化のための血液バイオマーカー：系統的レビューおよびメタアナリシス

To evaluate the diagnostic performance of blood-based assays for distinguishing malignant from benign indeterminate pulmonary nodules.

We performed a systematic review and meta-analysis of prospective and retrospective studies assessing blood-based biomarker tests in patients with indeterminate pulmonary nodules (6 to 30 mm). Thirteen studies (n=2771 patients) met inclusion criteria, spanning proteomic assays, circulating cell-free DNA fragmentomics, and integrated clinical-biomarker models. Pooled sensitivity, specificity, and area under the receiver operating characteristic curve (AUC) were estimated using random-effects models. Subgroup analyses, QUADAS-2 risk of bias assessment, and publication bias testing were conducted.

Across all studies, the pooled sensitivity was 84.6% and specificity was 65.3%, with a summary AUC of 0.78. At a modeled 15% prevalence, the negative predictive value was 96%, supporting clinical utility for ruling out malignancy. cfDNA fragmentomics assays demonstrated the highest accuracy (AUC: 0.87) with consistent performance even in small nodules and ground-glass opacities, and across smokers and never-smokers. Proteomic assays showed greater variability, while integrated models performed with high sensitivity but modest specificity. Risk of bias was generally low, though heterogeneity was moderate. No evidence of publication bias was detected.

Blood-based biomarker assays show promise as adjunctive tools for pulmonary nodule management, particularly for excluding malignancy and potentially reducing unnecessary invasive procedures. Fragmentomics platforms demonstrated the most consistent performance, but further large-scale prospective studies are needed to address outstanding questions as clinical applications expand.