Can computed tomography characteristics predict outcomes in patients undergoing radial endobronchial ultrasound-guided biopsy of peripheral lung lesions?

INTRODUCTION: Percutaneous computed tomography (CT)-guided lung biopsy is a standard minimally invasive technique for sampling peripheral lung lesions. Radial endobronchial ultrasound (EBUS) offers an alternative approach but it has yet to be defined which patients are most suited to this procedure. The primary aim of this study was to investigate whether CT characteristics could predict the success of radial EBUS-guided sampling.
METHODS: The University Hospital South Manchester provides a radial EBUS service, under conscious sedation without fluoroscopy, double-hinged curettes, or guide sheaths, to a large cancer Network in the United Kingdom. This retrospective analysis of a prospectively maintained database included all patients undergoing radial EBUS from January 2011 to June 2013. Lesion size, structure, location, and presence of a bronchus sign on thoracic CT were analyzed against predefined outcomes using multivariate analysis.
RESULTS: One-hundred and seventeen patients underwent radial EBUS in the study period (mean age 69.5, mean lesion size 36.6 mm). The presence of a bronchus sign on CT was the only independent predictor of all predefined outcomes: (1) lesion identification with radial EBUS, (2) positioning of probe within the center of the lesion, and (3) accurate pathological diagnosis; odds ratio (OR) 31.1 (7.8-123.9, p < 0.0001), OR 44.8 (5.6-354.9, p < 0.0001) and OR 46.6 (11.1-195.3, p < 0.0001) respectively. The sensitivity and diagnostic accuracy for those patients with a bronchus sign on CT was 87.3% and 86.7% compared with 12.5% and 11.1% for those lacking the bronchus sign. DISCUSSION: The patients most likely to benefit from radial EBUS, without the use of adjuncts, are those with a bronchus sign on CT.