Three-Dimensional Deformations of Pulmonary Collapse for Intraoperative Augmented Reality Guidance: A Proof-of-Concept Study

OBJECTIVE: During pulmonary surgery, the lung is deflated to facilitate the procedure. This study aimed to assess the deformation of the bronchial tree and pulmonary parenchyma during lung collapse, for eventual use in augmented reality (AR) guidance during pulmonary resections.

METHODS: The concept was first tested in 2 porcine models by analyzing paired computed tomography scans of collapsed and inflated lungs, then applied to 6 human patients. Bronchus and parenchyma were segmented, and a bronchus centerline was calculated. The diameter, length differences, angular deformations, and volume differences of the parenchyma were calculated. Finally, these deformations were applied on the inflated bronchus centerline to generate an artificially collapsed bronchus.

RESULTS: In both the porcine and human models, the pulmonary collapse resulted in substantial volumetric and anatomical changes. For the humans, the right lung showed a median displacement of 14.41 mm in the dorsomedial direction, while the left lung was displaced 11.99 mm in the dorsolateral direction (P = 0.79). Median volume reduction was 970 mL for the right lung and 878 mL for the left lung. Bronchial narrowing was observed, with a median diameter reduction of 0.14 mm for the right lung and 1.23 mm for the left lung. Moreover, the lengths of the bronchial segments were reduced, with a median length reduction of 0.20 mm for the right sided and 0.72 mm for the left sided.

CONCLUSIONS: Algorithmically driven calculations of the intraoperative pulmonary collapse of human and porcine lungs were performed and applied onto an inspirated bronchus. This resulted in an artificial collapsed bronchus. This method could be a foundation for a dynamical deformable deflation model, suitable for intraoperative AR-based pulmonary navigation.