PURPOSE: To investigate the capabilities of a neurovascular navigation prototype in phantom experiments. MATERIALS AND METHODS: The proposed navigation system integrates three-dimensional (3D) visualization of the anatomy and real-time electromagnetic localization of the endovascular tools. A 3D model of an endovascular phantom was reconstructed from thresholded preprocedural computed tomographic (CT) data. The vascular model was aligned with the reference frame of an electromagnetic tracker by using paired-point matching based on eight external fiducials. The robustness and accuracy of the registration were evaluated in 29 experiments. A magnetically tracked catheter was inserted into the carotid artery of the phantom, and the navigation system was used to reach five predefined vascular landmarks. The spatial accuracy of the prototype was evaluated during 50 endovascular targeting attempts. RESULTS: The navigation system achieved accurate co-registration of the location of a catheter inside a 3D reconstruction of a phantom vasculature. The experiments demonstrated the robustness of the registration, with a standard deviation for the translation and rotation components of 0.7 mm and 0.3 degrees , respectively. The maximal average error on the fiducials was 3.2 mm. Endovascular navigation by using the 3D real-time display was successfully performed with a mean overall accuracy of 2.7 mm +/- 0.7 and no projection limitation. CONCLUSION: The authors developed a navigation system that provides real-time 3D visualization of the position of endovascular components in a neurovascular phantom. The preliminary in vitro experiments showed clinically acceptable accuracy.
PURPOSE: To investigate the capabilities of a neurovascular navigation prototype in phantom experiments. MATERIALS AND METHODS: The proposed navigation system integrates three-dimensional (3D) visualization of the anatomy and real-time electromagnetic localization of the endovascular tools. A 3D model of an endovascular phantom was reconstructed from thresholded preprocedural computed tomographic (CT) data. The vascular model was aligned with the reference frame of an electromagnetic tracker by using paired-point matching based on eight external fiducials. The robustness and accuracy of the registration were evaluated in 29 experiments. A magnetically tracked catheter was inserted into the carotid artery of the phantom, and the navigation system was used to reach five predefined vascular landmarks. The spatial accuracy of the prototype was evaluated during 50 endovascular targeting attempts. RESULTS: The navigation system achieved accurate co-registration of the location of a catheter inside a 3D reconstruction of a phantom vasculature. The experiments demonstrated the robustness of the registration, with a standard deviation for the translation and rotation components of 0.7 mm and 0.3 degrees , respectively. The maximal average error on the fiducials was 3.2 mm. Endovascular navigation by using the 3D real-time display was successfully performed with a mean overall accuracy of 2.7 mm +/- 0.7 and no projection limitation. CONCLUSION: The authors developed a navigation system that provides real-time 3D visualization of the position of endovascular components in a neurovascular phantom. The preliminary in vitro experiments showed clinically acceptable accuracy.
Authors: L Zhang; S Parrini; C Freschi; V Ferrari; S Condino; M Ferrari; D Caramella Journal: Int J Comput Assist Radiol Surg Date: 2013-07-05 Impact factor: 2.924