Radial-probe endobronchial ultrasound outcomes in the investigation of peripheral pulmonary lesions: a New Zealand perspective.

Abstract

Radial-probe endobronchial ultrasound (radial-EBUS) is becoming an important investigation for peripheral pulmonary lesions (PPL). A key advantage of radial-EBUS is the favourable risk profile compared with current gold-standard computerised tomography-guided biopsy. To investigate the diagnostic yield, predictors of positive yield and radial-EBUS safety in a New Zealand institution. We also determined whether molecular analysis was possible on the same tissue samples. We performed a retrospective analysis of all patients (n = 68) from Middlemore Hospital, Auckland, undergoing radial-EBUS with guide-sheath for PPL from March 2015 to August 2016. Clinical, radiological and procedural data were collected. Radial-EBUS diagnostic yield was determined for malignant and benign diagnoses, and molecular analysis yield was determined on appropriate malignant samples. Logistic regression was used to determine factors predicting successful radial-EBUS. Overall diagnostic yield of radial-EBUS was 55.9% (95% confidence interval (CI): 44.3-67.9). Malignant diagnostic sensitivity was 60.8% (95% CI: 46.1-74.2) and benign diagnostic sensitivity was 50% (95% CI: 23-77). Lesions close to the hilum (P = 0.039), concentric radial-probe positioning (P = 0.008) and the use of forceps as first instrument (P = 0.0049) significantly predicted successful diagnostic yield. Of the malignant cases 81.0% (95% CI: 58.1-94.6) were sufficient for molecular analysis. Pneumothorax occurred in 4.4% (95% CI: 0.9-12.4), none required chest drain intervention. There were no cases of significant pulmonary haemorrhage. Radial-EBUS was shown to be safe with diagnostic yield similar to international reports. Important predictors of success include distance from hilum, probe position and forceps as first instrument. We also demonstrated that molecular analysis is possible in radial-EBUS obtained samples.