Giuseppe Panuccio1, Giovanni Federico Torsello2, Markus Pfister3, Theodosios Bisdas4, Michel J Bosiers4, Giovanni Torsello4, Martin Austermann4. 1. Department of Vascular Surgery, St. Franziskus Hospital and University Clinic of Muenster, Muenster, Germany. Electronic address: giuseppe.panuccio@gmail.com. 2. Department of Vascular Surgery, St. Franziskus Hospital and University Clinic of Muenster, Muenster, Germany; Westphalian Center for Radiology, Muenster, Germany. 3. Siemens AG, Healthcare Sector, Forchheim, Germany. 4. Department of Vascular Surgery, St. Franziskus Hospital and University Clinic of Muenster, Muenster, Germany.
Abstract
OBJECTIVE: To assess the usability of a fully automated fusion imaging engine prototype, matching preinterventional computed tomography with intraoperative fluoroscopic angiography during endovascular aortic repair. METHODS: From June 2014 to February 2015, all patients treated electively for abdominal and thoracoabdominal aneurysms were enrolled prospectively. Before each procedure, preoperative planning was performed with a fully automated fusion engine prototype based on computed tomography angiography, creating a mesh model of the aorta. In a second step, this three-dimensional dataset was registered with the two-dimensional intraoperative fluoroscopy. The main outcome measure was the applicability of the fully automated fusion engine. Secondary outcomes were freedom from failure of automatic segmentation or of the automatic registration as well as accuracy of the mesh model, measuring deviations from intraoperative angiography in millimeters, if applicable. RESULTS: Twenty-five patients were enrolled in this study. The fusion imaging engine could be used in successfully 92% of the cases (n = 23). Freedom from failure of automatic segmentation was 44% (n = 11). The freedom from failure of the automatic registration was 76% (n = 19), the median error of the automatic registration process was 0 mm (interquartile range, 0-5 mm). CONCLUSIONS: The fully automated fusion imaging engine was found to be applicable in most cases, albeit in several cases a fully automated data processing was not possible, requiring manual intervention. The accuracy of the automatic registration yielded excellent results and promises a useful and simple to use technology.
OBJECTIVE: To assess the usability of a fully automated fusion imaging engine prototype, matching preinterventional computed tomography with intraoperative fluoroscopic angiography during endovascular aortic repair. METHODS: From June 2014 to February 2015, all patients treated electively for abdominal and thoracoabdominal aneurysms were enrolled prospectively. Before each procedure, preoperative planning was performed with a fully automated fusion engine prototype based on computed tomography angiography, creating a mesh model of the aorta. In a second step, this three-dimensional dataset was registered with the two-dimensional intraoperative fluoroscopy. The main outcome measure was the applicability of the fully automated fusion engine. Secondary outcomes were freedom from failure of automatic segmentation or of the automatic registration as well as accuracy of the mesh model, measuring deviations from intraoperative angiography in millimeters, if applicable. RESULTS: Twenty-five patients were enrolled in this study. The fusion imaging engine could be used in successfully 92% of the cases (n = 23). Freedom from failure of automatic segmentation was 44% (n = 11). The freedom from failure of the automatic registration was 76% (n = 19), the median error of the automatic registration process was 0 mm (interquartile range, 0-5 mm). CONCLUSIONS: The fully automated fusion imaging engine was found to be applicable in most cases, albeit in several cases a fully automated data processing was not possible, requiring manual intervention. The accuracy of the automatic registration yielded excellent results and promises a useful and simple to use technology.
Authors: Katharina Breininger; Moritz Hanika; Mareike Weule; Markus Kowarschik; Marcus Pfister; Andreas Maier Journal: Int J Comput Assist Radiol Surg Date: 2019-08-22 Impact factor: 2.924