Bronchoscopy using a head-mounted mixed reality device—a phantom study and a first in-patient user experience

Abstract

Background: Bronchoscopy for peripheral lung lesions may involve image sources such as computed tomography (CT), fluoroscopy, radial endobronchial ultrasound (R-EBUS), and virtual/electromagnetic navigation bronchoscopy. Our objective was to evaluate the feasibility of replacing these multiple monitors with a head-mounted display (HMD), always providing relevant image data in the line of sight of the bronchoscopist. Methods: A total of 17 pulmonologists wearing a HMD (Microsoft® HoloLens 2) performed bronchoscopy with electromagnetic navigation in a lung phantom. The bronchoscopists first conducted an endobronchial inspection and navigation to the target, followed by an endobronchial ultrasound bronchoscopy. The HMD experience was evaluated using a questionnaire. Finally, the HMD was used in bronchoscopy inspection and electromagnetic navigation of two patients presenting with hemoptysis. Results: In the phantom study, the perceived quality of video and ultrasound images was assessed using a visual analog scale, with 100% representing optimal image quality. The score for video quality was 58% (95% confidence interval [CI] 48%–68%) and for ultrasound image quality, the score was 43% (95% CI 30%–56%). Contrast, color rendering, and resolution were all considered suboptimal. Despite adjusting the brightness settings, video image rendering was considered too dark. Navigation to the target for biopsy sampling was accomplished by all participants, with no significant difference in procedure time between experienced and less experienced bronchoscopists. The overall system latency for the image stream was 0.33–0.35 s. Fifteen of the pulmonologists would consider using HoloLens for navigation in the periphery, and two would not consider using HoloLens in bronchoscopy at all. In the human study, bronchoscopy inspection was feasible for both patients. Conclusion: Bronchoscopy using an HMD was feasible in a lung phantom and in two patients. Video and ultrasound image quality was considered inferior to that of video monitors. HoloLens 2 was suboptimal for airway and mucosa inspection but may be adequate for virtual bronchoscopy navigation.