Yuan Dong1,2, Chenxi Zhang3,4, Dafeng Ji1,2,5, Manning Wang1,2, Zhijian Song6,7. 1. Digital Medical Research Center, School of Basic Medical Sciences, Fudan University, Shanghai, China. 2. Shanghai Key Laboratory of Medical Imaging Computing and Computer Assisted Intervention, Shanghai, China. 3. Digital Medical Research Center, School of Basic Medical Sciences, Fudan University, Shanghai, China. chenxizhang@fudan.edu.cn. 4. Shanghai Key Laboratory of Medical Imaging Computing and Computer Assisted Intervention, Shanghai, China. chenxizhang@fudan.edu.cn. 5. Anatomy Department of Medical School, Nantong University, Nantong, China. 6. Digital Medical Research Center, School of Basic Medical Sciences, Fudan University, Shanghai, China. zjsong@fudan.edu.cn. 7. Shanghai Key Laboratory of Medical Imaging Computing and Computer Assisted Intervention, Shanghai, China. zjsong@fudan.edu.cn.
Abstract
PURPOSE: The conventional surface-based method only registers the facial zone with preoperative point cloud, resulting in low accuracy away from the facial area. Acquiring a point cloud of the entire head for registration can improve registration accuracy in all parts of the head. However, it takes a long time to collect a point cloud of the entire head. It may be more practical to selectively scan part of the head to ensure high registration accuracy in the surgical area of interest. In this study, we investigate the effects of different scan regions on registration errors in different target areas when using a surface-based registration method. METHODS: We first evaluated the correlation between the laser scan resolution and registration accuracy to determine an appropriate scan resolution. Then, with the appropriate resolution, we explored the effects of scan modes on registration error in computer simulation experiments, phantom experiments and two clinical cases. The scan modes were designed based on different combinations of five zones of the head surface, i.e., the sphenoid-frontal zone, parietal zone, left temporal zone, right temporal zone and occipital zone. In the phantom experiment, a handheld scanner was used to acquire a point cloud of the head. A head model containing several tumors was designed, enabling us to calculate the target registration errors deep in the brain to evaluate the effect of regional-surface-based registration. RESULT: The optimal scan modes for tumors located in the sphenoid-frontal, parietal and temporal areas are mode 4 (i.e., simultaneously scanning the sphenoid-frontal zone and the temporal zone), mode 4 and mode 6 (i.e., simultaneously scanning the sphenoid-frontal zone, the temporal zone and the parietal zone), respectively. For the tumor located in the occipital area, no modes were able to achieve reliable accuracy. CONCLUSION: The results show that selecting an appropriate scan resolution and scan mode can achieve reliable accuracy for use in sphenoid-frontal, parietal and temporal area surgeries while effectively reducing the operation time.
PURPOSE: The conventional surface-based method only registers the facial zone with preoperative point cloud, resulting in low accuracy away from the facial area. Acquiring a point cloud of the entire head for registration can improve registration accuracy in all parts of the head. However, it takes a long time to collect a point cloud of the entire head. It may be more practical to selectively scan part of the head to ensure high registration accuracy in the surgical area of interest. In this study, we investigate the effects of different scan regions on registration errors in different target areas when using a surface-based registration method. METHODS: We first evaluated the correlation between the laser scan resolution and registration accuracy to determine an appropriate scan resolution. Then, with the appropriate resolution, we explored the effects of scan modes on registration error in computer simulation experiments, phantom experiments and two clinical cases. The scan modes were designed based on different combinations of five zones of the head surface, i.e., the sphenoid-frontal zone, parietal zone, left temporal zone, right temporal zone and occipital zone. In the phantom experiment, a handheld scanner was used to acquire a point cloud of the head. A head model containing several tumors was designed, enabling us to calculate the target registration errors deep in the brain to evaluate the effect of regional-surface-based registration. RESULT: The optimal scan modes for tumors located in the sphenoid-frontal, parietal and temporal areas are mode 4 (i.e., simultaneously scanning the sphenoid-frontal zone and the temporal zone), mode 4 and mode 6 (i.e., simultaneously scanning the sphenoid-frontal zone, the temporal zone and the parietal zone), respectively. For the tumor located in the occipital area, no modes were able to achieve reliable accuracy. CONCLUSION: The results show that selecting an appropriate scan resolution and scan mode can achieve reliable accuracy for use in sphenoid-frontal, parietal and temporal area surgeries while effectively reducing the operation time.
Authors: Michael I Miga; Tuhin K Sinha; David M Cash; Robert L Galloway; Robert J Weil Journal: IEEE Trans Med Imaging Date: 2003-08 Impact factor: 10.048
Authors: Kurt Schicho; Michael Figl; Rudolf Seemann; Markus Donat; Michael L Pretterklieber; Wolfgang Birkfellner; Astrid Reichwein; Felix Wanschitz; Franz Kainberger; Helmar Bergmann; Arne Wagner; Rolf Ewers Journal: J Neurosurg Date: 2007-04 Impact factor: 5.115
Authors: Andreas Raabe; René Krishnan; Robert Wolff; Elvis Hermann; Michael Zimmermann; Volker Seifert Journal: Neurosurgery Date: 2002-04 Impact factor: 4.654
Authors: Peter A Woerdeman; Peter W A Willems; Herke J Noordmans; Cornelis A F Tulleken; Jan Willem Berkelbach van der Sprenkel Journal: J Neurosurg Date: 2007-06 Impact factor: 5.115