Computer-aided endovascular aortic repair using fully automated two- and three-dimensional fusion imaging

Area of Science:

• Medical Imaging
• Surgical Technology
• Cardiovascular Interventions

Background:

• Endovascular aortic repair (EVAR) requires precise alignment of preinterventional imaging with intraoperative guidance.
• Current methods may involve complex manual registration processes.
• Automated solutions could enhance efficiency and accuracy in EVAR procedures.

Purpose of the Study:

• To evaluate the usability and performance of a novel, fully automated fusion imaging engine prototype.
• To assess the engine's ability to match preinterventional computed tomography angiography (CTA) with intraoperative fluoroscopic angiography during EVAR.
• To determine the applicability, segmentation success, registration accuracy, and mesh model fidelity of the automated system.

Main Methods:

• Prospective enrollment of 25 patients undergoing elective EVAR for abdominal or thoracoabdominal aneurysms (June 2014 - February 2015).
• Preoperative planning using a fully automated fusion engine to create a 3D aortic mesh model from CTA.
• Intraoperative registration of the 3D model with 2D fluoroscopy.
• Outcome measures included engine applicability, freedom from failure in automatic segmentation and registration, and mesh model accuracy.

Main Results:

• The fusion imaging engine was successfully applied in 92% (23/25) of cases.
• Freedom from failure of automatic segmentation was 44% (11/25).
• Freedom from failure of automatic registration was 76% (19/25), with a median registration error of 0 mm (IQR, 0-5 mm).

Conclusions:

• The fully automated fusion imaging engine demonstrated high applicability in EVAR procedures.
• While some cases required manual intervention for data processing, the automatic registration accuracy was excellent.
• The technology shows promise as a useful and user-friendly tool for enhancing EVAR navigation.