# Vessel sign analysis paves the way to optimized CBCT application in interventional pulmonology: COMBINED algorithm as a one-stop-shop

## TL;DR

This study shows that using a combined approach with CBCT and specific tools in interventional pulmonology can achieve high diagnostic accuracy for lung nodules, similar to robotic systems.

## Contribution

The study introduces a combined CBCT algorithm using standard instruments as an effective alternative to robotic bronchoscopy for diagnosing peripheral lung nodules.

## Key findings

- The combined CBCT approach achieved an 89.4% diagnostic yield for malignant cases.
- Vessel sign analysis, especially of pulmonary artery branches, improved outcome prediction by 19% in nodules larger than 11mm.
- The diagnostic performance of the combined approach was comparable to robotic-assisted bronchoscopy.

## Abstract

Introduction: We used CBCT application as one-stop-shop nodule orientated approach in regards to increase DY, reduce complication rate, reduce time on-table and economical costs with classical peripheral instruments including mini-cryoprobe (ERBE 1,1mm), rEBUS (Olympus) and standard RUFBs (Olympus Company) with at least 2mm working channel and 4,2mm outer diameter for the diagnosis of peripheral targets (iSPNs) in a prospective all-comers registry after detailed analysis of pre-interventional CT for vessel- and bronchus sign classes.

Materials and Methods: From Jun 2017 until Nov 2019 in 90 all-comers patients between 16 and 95 years fit for bronchoscopy with 101 peripheral lesions in a daily routine scheme after informed consent about this prospective registry were included. For histological proven benign disease in any lesion patients had to adhere FU according radiological guidelines and further on by re-visits for at least 2 years after biopsy resulting into last visit in Feb 2022 without any drop-out. Present HRCT was mandatory to achieve one day before intervention. It had to be decided by the examiner mainly after analysis of the preset HRCT which of the 3 CBCT driven modalities were used for diagnostical approach: A) Pure endobronchial approach (CBCT, rEBUS, TBB), B) Pure transthoracical approach with a 21G core-biopsy needle (BIOPINCE needle) with CBCT only, or C) Combined approach as described below (CBCT, rEBUS, TTNA). As instruments were available common forceps and needles, EWC, curette and various RUFB (Olympus Company) mentioned in the materials section. A second CBCT was only allowed in the combined approach group to plan the 3D transthoracic approach in expiration whereas even a CBCT for tool-in-lesion control (TIL CBCT) was never allowed in all 3 groups.

Results: In 100 lesions predefined modalities pure endobiopsy, pure TTNA and combined approaches were performed in 77, 9 and 14 lesions respectively without any pneumothorax or bleeding. In these 3 modalities we found confirmed (mostly specific) benign and malignant cases 47 and 30, 4 and 5, 2 and 12 respectively. Lesion sizes in the 3 different groups were (median, mean) 14 and 17,7mm (of those 41 invisible of 77 under XR (53%) in the pure endobiopsy group), 27 and 31mm (11% invisible under XR in the pure TTNA group), 18,5 and 23mm (35% invisible under XR in the combined group) respectively. In the 3 groups for the malignant cases 25 of 30, 5 of 5 and 12 of 12 were diagnosed correctly rendering a diagnostical yield of 42 in 47 malignant cases for the whole algorithm (89,4%) with sizes (mean, median) for the whole algorithm of 16 and 19,7mm respectively which is comparable to published data for robotic-assisted bronchoscopy yield. In regards to vessel sign analysis it has to be clearly stated that the significance level for outcome prediction is inferior to bronchus sign analysis. In multivariate analysis there was a clear tendency towards higher outcome prediction especially if a pulmonary artery branch leads into such target even when a bronchus sign is missing. For NY when comparing univariate analysis and partition model analysis at a set diameter of >11mm with significance (p=0,0052) the additional advantage of analysing a given vessel sign (especially pulmonary artery branches) seems to add on 19% of valuable outcome prediction.

Conclusion: A nodule orientated approach in a manual CBCT-AF environment including typical instruments renders in experienced hands comparable results to robotic assisted bronchoscopy even without UTN bronchoscopes or other specialized, therefore expensive tools. In multivariate analysis only bronchus sign analysis revealed significant (p = 0,05) prediction of navigational yield outcome prediction whereas vessel sign analysis increases highly the odds ratio in favor of positive outcome prediction but without significance at the given level. In a partition model to erase outliers at a set iSPN diameter >11mm vessel sign analysis (especially pulmonary artery branches) renders a significant and ameliorated prediction of NY.