Diagnostic accuracy and safety of nonsurgical biopsy for diagnosing pulmonary ground-glass opacities: a systematic review and meta-analysis

Introduction: Previous meta-analyses have explored the diagnostic accuracy and safety of computed tomography-guided percutaneous lung biopsy of ground-glass opacities (GGOs). However, no research investigated the role of nonsurgical biopsies (including transbronchial approaches). Additionally, studies reporting the diagnostic accuracy of GGOs with different characteristics are scarce, with no quantitative assessment published to date. We performed a systematic review to explore the diagnostic accuracy and safety of nonsurgical biopsy for diagnosing GGOs, especially those with higher ground-glass components and smaller nodule sizes. Methods: A thorough literature search of four databases was performed to compile studies evaluating both or either of the diagnostic accuracy and complications of nonsurgical biopsy for GGOs. A bivariate random-effects model and random-effect model were utilized for data synthesis. The methodological quality of the studies was assessed according to the Quality Assessment of Diagnostic Accuracy Studies-2 tool. Results: Nineteen eligible studies with a total of 1,379 biopsy-sampled lesions were analyzed, of which 1,124 were confirmed to be malignant. Nonsurgical biopsy reported a pooled sensitivity of 0.89, a specificity of 0.99, and a negative predictive value (NPV) of 60.3%. The overall sensitivity, specificity, and NPV of nonsurgical biopsy for diagnosing GGOs according to GGO component were 0.90, 0.99, and 77.2% in pure GGOs; 0.87, 0.99, and 67.2% in GG-predominant lesions; and 0.89, 1.00, and 44.1% in solid-predominant lesions, respectively. Additionally, the diagnostic sensitivity was better in lesions ≥20 mm than in small lesions (0.95 vs. 0.88). Factors that contributed to higher sensitivity were the use of a coaxial needle system and CT fluoroscopy but not the needle gauge. The summary sensitivity of core needle biopsy (CNB) was not significantly higher than fine needle aspiration (FNA) (0.92 vs. 0.84; p = 0.42); however, we found an increased incidence of hemorrhage in CNB compared with FNA (60.9 vs. 14.2%; p = 0.012). Conclusion: Nonsurgical biopsy for diagnosing GGOs shows high sensitivity and specificity with an acceptably low risk of complications. However, negative biopsy results are unreliable in excluding malignancy, necessitating resampling or subsequent follow-up. The applicability of our study is limited due to significant heterogeneity, indirect comparisons, and the paucity of data on bronchoscopic approaches, restricting the generalizability of our findings to patients requiring transbronchial biopsies.