Cone Beam CT-Guided Bronchoscopy

Abstract

Despite recent advances in peripheral bronchoscopy, its diagnostic yield for lung nodules remains suboptimal and varies considerably among different publications.1–5 The explanation for this wide range in the reported diagnostic yield is multifactorial. Different populations, operators, techniques, and study designs are some of the contributing factors. Unlike our gratifying experience with the advent of endobronchial ulrasound-guided transbronchial needle aspiration and its steep learning curve, the wide range in diagnostic yield of peripheral bronchoscopy highlights the lack of a reliable and reproducible technique. From the technical standpoint, the diagnostic yield of peripheral bronchoscopy can be influenced at 3 major levels: “navigation,” “confirmation,” and “acquisition.” “Navigation” is the ability to reach a target. This is dependent, among others, on target-related factors (ie, size, location, “bronchus sign”), the operator’s ability to interpret chest computed tomography (CT) findings/anatomy and apply them during bronchoscopy, and the ability to follow the desired pathway. Navigation software, electromagnetic-guided navigation, ultrathin scopes, steerable catheters, and robotic bronchoscopy may have an impact at this level. “Confirmation” is the ability to demonstrate that our navigation or sampling tools are in contact with the target. Leaving chest CT aside, the only technique that can potentially confirm contact with the target in real time is radial-probe endobronchial ultrasound. The pseudo-confirmation provided by navigational software is based on prebronchoscopy CT scans, it is not real time, and it is bound to a margin of error. Cone beam CT (CBCT), as we will discuss later on, may play a key role at this level. “Acquisition” is our ability to obtain diagnostic samples. This can be influenced by factors associated with the target (ie, malignant vs. benign histology), with our sampling tools, and also with the type of contact that we achieve with the target (ie, center, periphery, or adjacent to it). This last step—acquisition—is responsible for the gap between navigational yield and diagnostic yield. We may successfully reach a lesion, and yet, not be able to obtain diagnosis. CBCT is a newer CT modality that has been adopted widely by interventional radiologists. Unlike standard (fan beam) CT, the system is compact enough to allow mounting it on a moving C-arm. CBCT performs volumetric data acquisition in a single rotation of the source and detector, with the patient remaining stationary during the examination.5 This latter characteristic and its ability to provide standard fluoroscopy images make CBCT more suitable than standard CT as an aid in peripheral bronchoscopy. In the current issue of this journal Pritchett and coworkers retrospectively describe their experience on the use of CBCT during peripheral bronchoscopy.6 The authors performed a CBCT scan once patients were intubated for bronchoscopy before navigation. They utilized this scan to identify the target and, with a dedicated software, to create an overlay of the target on live fluoroscopy images (a modality that they termed “augmented fluoroscopy”). They then navigated to the lesions utilizing electromagnetic navigation equipment and tools—without using RP-EBUS for confirmation—and they utilized on-site cytology. They report that a second CBCT scan was only performed when deemed necessary. This technique was utilized for 93 lesions in 75 consecutive patients. The median size of these lesions was 16 mm (range, 7 to 55 mm). The overall diagnostic yield was 83.7% (95% confidence interval, 74.8%-89.9%) and pneumothorax occurred in 3 patients (4%). Radiation exposure data originated in a small subset of 9 patients and the mean effective dose (E) was 2 mSv per CBCT scan. The average number of CBCT scans per patient was 1.5. The first scan was performed before bronchoscopic navigation to generate augmented fluoroscopy images. Hence, the impact of CBC