Confocal laser endomicroscopy-assisted endobronchial ultrasound-guided needle aspiration in diagnosing mediastinal and/or hilar lymphadenopathy (COLLABORATION-III): protocol for a randomised controlled trial
Ziwen Zheng, Li Xu, Chao Wu, Jiangrong Liao, Mingming Deng, Run Tong, Gang HouIntroduction
Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) is an established first-line technique for sampling mediastinal and/or hilar lymphadenopathy (MHL). However, its diagnostic accuracy is constrained by the procurement of small-volume specimens, which may not reliably capture the diagnostically relevant tissue. Needle-based confocal laser endomicroscopy (nCLE) provides real-time in vivo microscopic imaging and has the potential to facilitate targeted tissue acquisition. EBUS-guided transbronchial mediastinal cryobiopsy (EBUS-TBMC) can yield larger and better-preserved specimens; however, its availability is limited. This study aims to evaluate whether nCLE-assisted EBUS-TBNA is non-inferior to EBUS-TBMC for diagnostic yield for MHL.
Methods and analysis
The confocal laser endomicroscopy-assisted endobronchial ultrasound-guided needle aspiration (COLLABORATION III) trial is a prospective, multicentre, stratified, non-inferiority randomised controlled trial conducted across five centres in China. The study aims to enroll 640 adult participants presenting with unexplained MHL with a short-axis measurement ≥10 mm necessitating histopathological confirmation. Participants will be stratified based on the preprocedural likelihood of malignancy and randomised at a 1:1 ratio to undergo nCLE-assisted EBUS-TBNA or EBUS-TBMC. The primary endpoint is the overall diagnostic yield, defined as the proportion of procedures that obtain specimens sufficient for a definitive histological or cytological diagnosis, with a non-inferiority margin set at −10%. Secondary endpoints include diagnostic sensitivity for both malignant and benign conditions, rate of granuloma detection, procedure duration and incidence of adverse events. The primary analysis will be conducted on an intention-to-treat basis, supplemented by a per-protocol sensitivity analysis. Non-inferiority will be established if the difference in the diagnostic yield exceeds −10%.
Ethics and dissemination
The study protocol was approved by the Institutional Review Board of the China–Japan Friendship Hospital (2025-KY-388). Written informed consent will be required from all participants. These findings will be disseminated through peer-reviewed journal publications and conference presentations.
Trial registration number
COLLABORATION-III was prospectively registered at ClinicalTrials.gov (ID: