Endovascular image-guided navigation: validation of two volume-volume registration algorithms

Area of Science:

• Medical Imaging
• Surgical Navigation
• Computational Anatomy

Background:

• Intraoperative CT scans have limited volume, hindering navigation system use.
• Preoperative CT scans offer larger volume coverage, essential for comprehensive navigation.
• Accurate registration between preoperative and intraoperative CT is vital for endovascular navigation.

Purpose of the Study:

• To evaluate the accuracy of two volume-volume registration algorithms (Siemens and CAMP) for endovascular navigation.
• To compare the performance of Siemens and CAMP algorithms against reference registrations.

Main Methods:

• Two volume-volume registration algorithms (Siemens, CAMP) were assessed.
• Fourteen intra- and postoperative CT datasets from abdominal aortic aneurysm patients were used.
• Registration accuracy was evaluated using five landmarks and calculating root mean square distances, with rigid body registration as reference.

Main Results:

• CAMP algorithm achieved a mean accuracy of 4.02 ± 1.52 mm.
• Siemens algorithm achieved a mean accuracy of 5.05 ± 4.74 mm.
• Reference registrations showed the highest accuracy at 2.72 ± 1.18 mm, with significant differences between algorithms (p < 0.001).

Conclusions:

• The CAMP algorithm offers more accurate registration than the Siemens algorithm for endovascular navigation.
• Both evaluated algorithms show potential for improving image guidance in endovascular procedures.
• Further refinement of registration algorithms is needed to match the accuracy of reference methods.