PURPOSE: In motion-compensated image-guided radiotherapy, accurate tracking of the target region is required. This tracking process includes building a correlation model between external surrogate motion and the motion of the target region. A novel correlation method is presented and compared with the commonly used polynomial model. METHODS AND MATERIALS: The CyberKnife system (Accuray, Inc., Sunnyvale/CA) uses a polynomial correlation model to relate externally measured surrogate data (optical fibres on the patient's chest emitting red light) to infrequently acquired internal measurements (X-ray data). A new correlation algorithm based on ɛ -Support Vector Regression (SVR) was developed. Validation and comparison testing were done with human volunteers using live 3D ultrasound and externally measured infrared light-emitting diodes (IR LEDs). Seven data sets (5:03-6:27 min long) were recorded from six volunteers. RESULTS: Polynomial correlation algorithms were compared to the SVR-based algorithm demonstrating an average increase in root mean square (RMS) accuracy of 21.3% (0.4 mm). For three signals, the increase was more than 29% and for one signal as much as 45.6% (corresponding to more than 1.5 mm RMS). Further analysis showed the improvement to be statistically significant. CONCLUSION: The new SVR-based correlation method outperforms traditional polynomial correlation methods for motion tracking. This method is suitable for clinical implementation and may improve the overall accuracy of targeted radiotherapy.
PURPOSE: In motion-compensated image-guided radiotherapy, accurate tracking of the target region is required. This tracking process includes building a correlation model between external surrogate motion and the motion of the target region. A novel correlation method is presented and compared with the commonly used polynomial model. METHODS AND MATERIALS: The CyberKnife system (Accuray, Inc., Sunnyvale/CA) uses a polynomial correlation model to relate externally measured surrogate data (optical fibres on the patient's chest emitting red light) to infrequently acquired internal measurements (X-ray data). A new correlation algorithm based on ɛ -Support Vector Regression (SVR) was developed. Validation and comparison testing were done with human volunteers using live 3D ultrasound and externally measured infrared light-emitting diodes (IR LEDs). Seven data sets (5:03-6:27 min long) were recorded from six volunteers. RESULTS: Polynomial correlation algorithms were compared to the SVR-based algorithm demonstrating an average increase in root mean square (RMS) accuracy of 21.3% (0.4 mm). For three signals, the increase was more than 29% and for one signal as much as 45.6% (corresponding to more than 1.5 mm RMS). Further analysis showed the improvement to be statistically significant. CONCLUSION: The new SVR-based correlation method outperforms traditional polynomial correlation methods for motion tracking. This method is suitable for clinical implementation and may improve the overall accuracy of targeted radiotherapy.
Authors: David P Gierga; Johanna Brewer; Gregory C Sharp; Margrit Betke; Christopher G Willett; George T Y Chen Journal: Int J Radiat Oncol Biol Phys Date: 2005-04-01 Impact factor: 7.038
Authors: Daniel A Low; Parag J Parikh; Wei Lu; James F Dempsey; Sasha H Wahab; James P Hubenschmidt; Michelle M Nystrom; Maureen Handoko; Jeffrey D Bradley Journal: Int J Radiat Oncol Biol Phys Date: 2005-11-01 Impact factor: 7.038
Authors: Ross I Berbeco; Steve B Jiang; Gregory C Sharp; George T Chen; Hassan Mostafavi; Hiroki Shirato Journal: Phys Med Biol Date: 2004-01-21 Impact factor: 3.609
Authors: Stephen R Bowen; Larry A Pierce; Adam M Alessio; Chi Liu; Scott D Wollenweber; Charles W Stearns; Paul E Kinahan Journal: J Med Imaging (Bellingham) Date: 2014-09-24
Authors: Guang Li; Jie Wei; Hailiang Huang; Qing Chen; Carl P Gaebler; Tiffany Lin; Amy Yuan; Andreas Rimner; James Mechalakos Journal: Med Phys Date: 2016-03 Impact factor: 4.071
Authors: Stephen R Bowen; Matthew J Nyflot; Michael Gensheimer; Kristi R G Hendrickson; Paul E Kinahan; George A Sandison; Shilpen A Patel Journal: Clin Transl Med Date: 2012-08-31