Performance of endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) for staging the radiologically normal mediastinum has been reported with inconsistent findings. We assessed the sensitivity of systematic staging using EBUS-TBNA for detection of radiologically occult mediastinal metastases in cN0/N1 lung cancer.
Studies evaluating EBUS-TBNA for systematic mediastinal staging in cN0/N1 lung cancer were identified by systematic review. Data extracted included: participant age and sex; EBUS-TBNA protocol; stage determined by radiology, EBUS-TBNA and surgery; 2×2 tables. Primary outcome was diagnostic accuracy of EBUS-TBNA for detection of unsuspected N2/N3 disease.
We identified 1173 articles. In total, 13 were included in a qualitative review and 9 (1146 patients) in a quantitative meta-analysis. Mean prevalence of N2/N3 disease was 15% (6% to 24%). EBUS-TBNA had pooled sensitivity of 49% [95% confidence interval (CI), 41%-57%], pooled specificity of 100% (95% CI, 99%-100%), mean negative predictive value 91% (82% to 100%) for detection of unsuspected N2/N3 metastases. Number needed to test to detect occult N2/N3 disease was 14 (95% CI, 10.8-16.3), which halved with addition of per-esophageal endoscopic ultrasound.
Preoperative systematic staging by EBUS-TBNA of early lung cancer can reduce postoperative upstaging. Sensitivity for detection of radiologically occult mediastinal metastases seems lower than selective sampling of pathologic lymph nodes. Verification of negative results by mediastinoscopy in selected cases remains of value.
*The Walter and Eliza Hall Institute of Medical Research
†Department of Medicine, University of Melbourne
Departments of ‡General Medicine
∥Respiratory Medicine, Royal Melbourne Hospital, Parkville
§Institute for Health and Aging, Australian Catholic University, Melbourne, Vic., Australia
T.L.L., D.P.S.: plan and design of study, performance of literature search and article review, contribution to writing manuscript. T.L.L., P.M.L.: data extraction. T.L.L., D.P.S., and A.G.: data analysis and interpretation. P.M.L., L.I., and A.G.: critical review of manuscript.
Disclosure: There is no conflict of interest or other disclosures.
Reprints: Tracy L. Leong, FRACP, The Walter and Eliza Hall Institute of Medical Research, 1G Royal Parade, Parkville, Vic. 3052, Australia (e-mail: firstname.lastname@example.org).
Received May 9, 2018
Accepted July 19, 2018