Diagnostic Accuracy of Ultrasound Guidance in Transthoracic Needle Biopsy: A Systematic Review and Meta-Analysis

Simon Lemieux,Lorence Pinard,Raphaël Marchand,Sonia Kali,Stephan Altmayer,Vicky Mai,Steeve Provencher
DOI: https://doi.org/10.1097/RTI.0000000000000811
2024-09-17
Abstract:Purpose: To perform a systematic review and meta-analysis of relevant studies to assess the diagnostic accuracy and safety outcomes of ultrasound (US)-guided transthoracic needle biopsy (TTNB) for peripheral lung and pleural lesions. Materials and methods: A search was performed through Medline, Embase, Web of Science, and Cochrane Central from inception up to September 23, 2022 for diagnostic accuracy studies reporting US-guided TTNB (Prospero registration: CRD42021225168). The primary outcome was diagnostic accuracy, which was assessed by sensitivity, specificity, likelihood ratios (LR), and diagnostic odds ratio. Sensitivity and subgroup analyses were performed to evaluate inter-study heterogeneity. The secondary outcome was the frequency of complications. Random-effects models were used for the analyses. The risk of bias and the applicability of the included studies were assessed using the QUADAS-2 tool. Publication bias was assessed by testing the association between the natural logarithm of the diagnostic odds ratio and the effective sample size. Results: Of the 7841 citations identified, 83 independent cohorts (11,767 patients) were included in the analysis. The pooled sensitivity of US-TTNB was 88% (95% CI: 86%-91%, 80 studies). Pooled specificity was 100% (95% CI: 99%-100%, 72 studies), resulting in positive LR, negative LR, and diagnostic odds ratio of 946 (-743 to 2635), 0.12 (0.09 to 0.14), and 8141 (1344 to 49,321), respectively. Complications occurred in 4% (95% CI: 3%-5%) of the procedures, with pneumothorax being the most frequent (3%; 95% CI: 2%-3%, 72 studies) and resulting in chest tube placement in 0.4% (95% CI: 0.2%-0.7%, 64 studies) of the procedures. Conclusions: US-TTNB is an effective and safe procedure for pleural lesions and peripheral lung lesions.
What problem does this paper attempt to address?