Anatomy-Aware Prediction of Bronchoscopic Accessibility from 3D CT

2026-09-29

AI Summary

This study addresses the reliance on manual, time-consuming, and subjective assessments of bronchial accessibility prior to bronchoscopy by proposing the first end-to-end automated prediction framework. Methodologically, it pioneers a supervised learning task for bronchial accessibility and introduces an anatomy-aware Mixture-of-Experts (MoE) architecture that deeply integrates CT morphological features, lobar anatomical priors, and geometric path constraints of the bronchial tree. In terms of contributions, this work releases the first clinical benchmark dataset for this task. Evaluated on 438 cases, the proposed framework achieves an AUROC of 0.8052, significantly outperforming existing state-of-the-art baselines and senior clinicians, thereby validating its substantial value for clinical procedural planning.

Abstract

Pre-operative planning for bronchoscopy is critical for the diagnosis of lung lesions. Current accessibility assessment relies on subjective manual inspection of CT scans, which is time-consuming and prone to inter-observer variability. In this paper, we formalize bronchoscopy accessibility prediction as a novel supervised learning task and present the first end-to-end framework to address it. We propose an Anatomy-Aware Mixture-of-Experts (MoE) model that integrates specialized modules: a CT Expert for local morphological features, a Lobe Expert for anatomical priors, and a Path Geometry Expert that encodes the sequential constraints of the bronchial tree. To support this task, we curated the first clinical dataset of 438 cases with pre-operative CT scans and documented procedural outcomes. Experimental results demonstrate that our method achieves an AUROC of 0.8052, significantly outperforming both state-of-the-art baselines and experienced human experts. This work establishes a new benchmark for computer-aided interventional planning in pulmonary medicine. Our data and code will be publicly available at https://nubagcilab.github.io/BronchoAccess/.

Problem

• bronchoscopy accessibility
• pre-operative planning
• 3D CT
• lung lesions
• computer-aided intervention

Innovation

• Bronchoscopy Accessibility
• Anatomy-Aware Mixture-of-Experts
• End-to-End Framework
• 3D CT
• Interventional Planning