Virtual bronchoscopic navigation improves the diagnostic yield of radial-endobronchial ultrasound for peripheral pulmonary lesions with involved bronchi on CT.

Abstract

Bronchoscopy using radial-endobronchial ultrasound (R-EBUS) and virtual bronchoscopic navigation (VBN) is a promising method for diagnosing peripheral pulmonary lesions. We previously performed a randomized comparative trial (RCT) (i.e., VBN combined with EBUS RCT) involving patients with 30-mm or smaller peripheral pulmonary lesions and found that the addition of VBN to R-EBUS improved the diagnostic yield. In the present study, we performed a retrospective subanalysis in order to identify patients for whom VBN is useful. The per-protocol population (194 cases) of the VBN combined with EBUS RCT was divided into subgroups based on the lesion size, lung lobe containing the lesion, lesion location, presence or absence of involved bronchi (bronchus sign) on thin-section CT and whether the lesion was detected on posterior-anterior (P-A) radiographs. The difference in the diagnostic yield between the VBN-assisted (VBNA) and non-VBN-assisted (NVBNA) groups was investigated. Within the bronchus sign-positive subgroup, the diagnostic yield in the VBNA and NVBNA groups was 94.4% (68/72) and 77.8% (56/72), respectively, showing a significantly higher yield in the VBNA group (p=0.004; odds ratio: 4.9). The yield was particularly high for lesions smaller than 20 mm (94.6% vs. 70.7%; p=0.006), lesions located in the peripheral third of the lung field (95.1% vs. 71.4%; p=0.005) and lesions invisible on P-A radiographs (90.0% vs. 41.7%; p=0.026). VBN improves the diagnostic yield when combined with R-EBUS to assess lesions exhibiting involved bronchi on CT images.