Ebus/ct Fusion for Image-guided Bronchoscopy

Open Access
Zang, Xiaonan
Graduate Program:
Electrical Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
December 17, 2015
Committee Members:
  • William Evan Higgins, Dissertation Advisor
  • William Evan Higgins, Committee Chair
  • Vishal Monga, Committee Member
  • Jeffrey Louis Schiano, Committee Member
  • Robert Collins, Committee Member
  • image segmentation
  • image registration
  • ultrasound
Many technical innovations in multimodal radiologic imaging and bronchoscopy have arisen to improve lung-cancer staging techniques. Among the innovations, endobronchial ultrasound (EBUS), which provides live guidance for localizing extraluminal sites, has become a standard procedure for cancerstaging bronchoscopy. Unfortunately, no direct linkage has been established between EBUS and the other modalities, leading to wide inter-physician performance variations and biopsy failures. In this thesis, we propose a multimodal image-guided system that synergistically combines EBUS with the other imaging sources and enables a true fusion of EBUS and computed tomography (CT) for biopsy-site localization of extraluminal sites during bronchoscopy. In particular, we develop EBUS segmentation methods to better interpret EBUS.We also develop EBUS reconstruction and registration methods to correlate EBUS findings with the information derived from the other modalities. Moreover, at the system-level, we integrate EBUS-related constraints to a pre-existing routeplanning method and introduce a procedure-planning approach to derive specific bronchoscope maneuvers for EBUS-guided region-of-interest (ROI) localization and biopsy. These bronchoscope maneuvers, presented in a pre-bronchoscopy report, ensure that the physician can naturally follow the multimodal guidance during live bronchoscopy. We also develop tools, establish a EBUS-based guidance protocol, and use our system to help the physician localize the diagnostic targets during live bronchoscopy.