Anna Belkova1, David R Pichora2, Randy E Ellis3,4. 1. School of Computing, Queen's University, Kingston, K7L 3N6, Canada. 2. Department of Surgery, Queen's University, Kingston, K7L 3N6, Canada. 3. School of Computing, Queen's University, Kingston, K7L 3N6, Canada. ellis@queensu.ca. 4. Department of Surgery, Queen's University, Kingston, K7L 3N6, Canada. ellis@queensu.ca.
Abstract
PURPOSE: Image-guided interventions that use preoperative 3D computed tomography (CT) models are limited by the preoperative segmentation time 3D image and collection of intraoperative registration data. Intraoperative CT imaging can be ergonomically efficient in a direct navigation system if the imaging device is accurately calibrated. A mobile-gantry CT scanner offers improved patient safety but presents technical challenges beyond those of a conventional scanner. The goal was to calibrate an optoelectronic navigation system to mobile-gantry CT with millimeter-level accuracy. METHODS: A custom calibration device was designed and manufactured. The calibrator contained optoelectronic markers for navigation reference and radio-opaque markers for CT reference. Calibrations were performed with a ceiling-mounted optoelectronic camera and with a portable camera, and then verified for accuracy. RESULTS: The component fiducial registration errors were extremely small, being 0.36 mm, with standard deviation of 0.16 mm, for the ceiling-mounted camera, and 0.05 mm, with standard deviation of 0.01 mm, for the portable camera. The net target registration error, measured as RMS deviation, was 1.58 mm for the ceiling-mounted camera and 0.73 mm for the portable camera. CONCLUSIONS: High-accuracy calibration of the mobile-gantry CT scanner was possible from a single preoperative CT image. A ceiling-mounted optoelectronic camera, which is ergonomically preferable, marginally met the accuracy criteria. The portable camera, which is in widespread use for conventional navigated surgery, had deep sub-millimeter error. This study demonstrates that high accuracy is achievable and offers a system developer options to trade off accuracy and user convenience in direct surgical navigation.
PURPOSE: Image-guided interventions that use preoperative 3D computed tomography (CT) models are limited by the preoperative segmentation time 3D image and collection of intraoperative registration data. Intraoperative CT imaging can be ergonomically efficient in a direct navigation system if the imaging device is accurately calibrated. A mobile-gantry CT scanner offers improved patient safety but presents technical challenges beyond those of a conventional scanner. The goal was to calibrate an optoelectronic navigation system to mobile-gantry CT with millimeter-level accuracy. METHODS: A custom calibration device was designed and manufactured. The calibrator contained optoelectronic markers for navigation reference and radio-opaque markers for CT reference. Calibrations were performed with a ceiling-mounted optoelectronic camera and with a portable camera, and then verified for accuracy. RESULTS: The component fiducial registration errors were extremely small, being 0.36 mm, with standard deviation of 0.16 mm, for the ceiling-mounted camera, and 0.05 mm, with standard deviation of 0.01 mm, for the portable camera. The net target registration error, measured as RMS deviation, was 1.58 mm for the ceiling-mounted camera and 0.73 mm for the portable camera. CONCLUSIONS: High-accuracy calibration of the mobile-gantry CT scanner was possible from a single preoperative CT image. A ceiling-mounted optoelectronic camera, which is ergonomically preferable, marginally met the accuracy criteria. The portable camera, which is in widespread use for conventional navigated surgery, had deep sub-millimeter error. This study demonstrates that high accuracy is achievable and offers a system developer options to trade off accuracy and user convenience in direct surgical navigation.
Authors: Jonathan S Hott; Vivek R Deshmukh; Jeffrey D Klopfenstein; Volker K H Sonntag; Curtis A Dickman; Robert F Spetzler; Stephen M Papadopoulos Journal: Neurosurgery Date: 2004-05 Impact factor: 4.654
Authors: Erin Janine Smith; Anton Oentoro; Hisham Al-Sanawi; Braden Gammon; Paul St John; David R Pichora; Randy E Ellis Journal: Med Image Comput Comput Assist Interv Date: 2010
Authors: Erin J Smith; Hisham A Al-Sanawi; Braden Gammon; Paul J St John; David R Pichora; Randy E Ellis Journal: Int J Comput Assist Radiol Surg Date: 2011-06-25 Impact factor: 2.924
Authors: Klaus Stelter; Georg Ledderose; John Martin Hempel; Dominik F B Morhard; Wilhelm Flatz; Eike Krause; Joachim Mueller Journal: Comput Aided Surg Date: 2012-03-27