Statistical modeling of pneumothorax deformation by mapping CT and cone-beam CT images

TL;DR

This paper presents a statistical modeling framework for pneumothorax deformation using paired CT images, enabling surgical guidance and intraoperative lung state reconstruction through deformable mesh registration and kernel-based learning.

Contribution

It introduces a novel deformable mesh registration and kernel-based deformation learning framework for modeling pneumothorax deformation from paired CT images.

Findings

Detailed pneumothorax deformation models

Evaluation of kernel-based deformation reconstruction

Mesh registration framework for surgical guidance

Abstract

In this study, we introduce statistical modeling methods for pneumothorax deformation using paired cone-beam computed tomography (CT) images. We designed a deformable mesh registration framework for shape changes involving non-linear deformation and rotation of the lungs. The registered meshes with local correspondences are available for both surgical guidance in thoracoscopic surgery and building statistical deformation models with inter-patient variations. In addition, a kernel-based deformation learning framework is proposed to reconstruct intraoperative deflated states of the lung from the preoperative CT models. This paper reports the findings of pneumothorax deformation and evaluation results of the kernel-based deformation framework.