Nicholas Hardcastle1, Michael S Hofman2, Rodney J Hicks3, Jason Callahan2, Tomas Kron4, Michael P MacManus5, David L Ball5, Price Jackson6, Shankar Siva7. 1. Department of Physical Sciences, Peter MacCallum Cancer Centre, East Melbourne, Australia; Centre for Medical Radiation Physics, University of Wollongong, Wollongong, Australia. Electronic address: nick.hardcastle@gmail.com. 2. Molecular Imaging, Centre for Cancer Imaging, Peter MacCallum Cancer Centre, Melbourne, Australia. 3. Molecular Imaging, Centre for Cancer Imaging, Peter MacCallum Cancer Centre, Melbourne, Australia; Department of Medicine, University of Melbourne, Melbourne, Australia. 4. Department of Medical Imaging and Radiation Sciences, Monash University, Clayton, Australia; The Sir Peter MacCallum Department of Oncology, Melbourne University, Victoria, Australia. 5. Division of Radiation Oncology and Cancer Imaging, Peter MacCallum Cancer Centre, East Melbourne, Australia; The Sir Peter MacCallum Department of Oncology, University of Melbourne, Melbourne, Australia. 6. Department of Physical Sciences, Peter MacCallum Cancer Centre, East Melbourne, Australia. 7. Division of Radiation Oncology and Cancer Imaging, Peter MacCallum Cancer Centre, East Melbourne, Australia.
Abstract
PURPOSE: Measuring changes in lung perfusion resulting from radiation therapy dose requires registration of the functional imaging to the radiation therapy treatment planning scan. This study investigates registration accuracy and utility for positron emission tomography (PET)/computed tomography (CT) perfusion imaging in radiation therapy for non-small cell lung cancer. METHODS: (68)Ga 4-dimensional PET/CT ventilation-perfusion imaging was performed before, during, and after radiation therapy for 5 patients. Rigid registration and deformable image registration (DIR) using B-splines and Demons algorithms was performed with the CT data to obtain a deformation map between the functional images and planning CT. Contour propagation accuracy and correspondence of anatomic features were used to assess registration accuracy. Wilcoxon signed-rank test was used to determine statistical significance. Changes in lung perfusion resulting from radiation therapy dose were calculated for each registration method for each patient and averaged over all patients. RESULTS: With B-splines/Demons DIR, median distance to agreement between lung contours reduced modestly by 0.9/1.1 mm, 1.3/1.6 mm, and 1.3/1.6 mm for pretreatment, midtreatment, and posttreatment (P < .01 for all), and median Dice score between lung contours improved by 0.04/0.04, 0.05/0.05, and 0.05/0.05 for pretreatment, midtreatment, and posttreatment (P < .001 for all). Distance between anatomic features reduced with DIR by median 2.5 mm and 2.8 for pretreatment and midtreatment time points, respectively (P = .001) and 1.4 mm for posttreatment (P > .2). Poorer posttreatment results were likely caused by posttreatment pneumonitis and tumor regression. Up to 80% standardized uptake value loss in perfusion scans was observed. There was limited change in the loss in lung perfusion between registration methods; however, Demons resulted in larger interpatient variation compared with rigid and B-splines registration. CONCLUSIONS: DIR accuracy in the data sets studied was variable depending on anatomic changes resulting from radiation therapy; caution must be exercised when using DIR in regions of low contrast or radiation pneumonitis. Lung perfusion reduces with increasing radiation therapy dose; however, DIR did not translate into significant changes in dose-response assessment.
PURPOSE: Measuring changes in lung perfusion resulting from radiation therapy dose requires registration of the functional imaging to the radiation therapy treatment planning scan. This study investigates registration accuracy and utility for positron emission tomography (PET)/computed tomography (CT) perfusion imaging in radiation therapy for non-small cell lung cancer. METHODS: (68)Ga 4-dimensional PET/CT ventilation-perfusion imaging was performed before, during, and after radiation therapy for 5 patients. Rigid registration and deformable image registration (DIR) using B-splines and Demons algorithms was performed with the CT data to obtain a deformation map between the functional images and planning CT. Contour propagation accuracy and correspondence of anatomic features were used to assess registration accuracy. Wilcoxon signed-rank test was used to determine statistical significance. Changes in lung perfusion resulting from radiation therapy dose were calculated for each registration method for each patient and averaged over all patients. RESULTS: With B-splines/Demons DIR, median distance to agreement between lung contours reduced modestly by 0.9/1.1 mm, 1.3/1.6 mm, and 1.3/1.6 mm for pretreatment, midtreatment, and posttreatment (P < .01 for all), and median Dice score between lung contours improved by 0.04/0.04, 0.05/0.05, and 0.05/0.05 for pretreatment, midtreatment, and posttreatment (P < .001 for all). Distance between anatomic features reduced with DIR by median 2.5 mm and 2.8 for pretreatment and midtreatment time points, respectively (P = .001) and 1.4 mm for posttreatment (P > .2). Poorer posttreatment results were likely caused by posttreatment pneumonitis and tumor regression. Up to 80% standardized uptake value loss in perfusion scans was observed. There was limited change in the loss in lung perfusion between registration methods; however, Demons resulted in larger interpatient variation compared with rigid and B-splines registration. CONCLUSIONS: DIR accuracy in the data sets studied was variable depending on anatomic changes resulting from radiation therapy; caution must be exercised when using DIR in regions of low contrast or radiation pneumonitis. Lung perfusion reduces with increasing radiation therapy dose; however, DIR did not translate into significant changes in dose-response assessment.
Authors: Pierre-Yves Le Roux; Amir Iravani; Jason Callahan; Kate Burbury; Peter Eu; Daniel P Steinfort; Eddie Lau; Beverly Woon; Pierre-Yves Salaun; Rodney J Hicks; Michael S Hofman Journal: Eur J Nucl Med Mol Imaging Date: 2019-05-01 Impact factor: 9.236
Authors: Pegah Jahangiri; Kamyar Pournazari; Drew A Torigian; Thomas J Werner; Samuel Swisher-McClure; Charles B Simone; Abass Alavi Journal: Eur J Nucl Med Mol Imaging Date: 2018-09-18 Impact factor: 9.236
Authors: Pierre-Yves Le Roux; Tracy L Leong; Stephen A Barnett; Rodney J Hicks; Jason Callahan; Peter Eu; Renee Manser; Michael S Hofman Journal: Cancer Imaging Date: 2016-08-20 Impact factor: 3.909
Authors: Pierre-Yves Le Roux; Shankar Siva; Jason Callahan; Yannis Claudic; David Bourhis; Daniel P Steinfort; Rodney J Hicks; Michael S Hofman Journal: EJNMMI Res Date: 2017-10-10 Impact factor: 3.138
Authors: Jeffrey Barber; Johnson Yuen; Michael Jameson; Laurel Schmidt; Jonathan Sykes; Alison Gray; Nicholas Hardcastle; Callie Choong; Joel Poder; Amy Walker; Adam Yeo; Ben Archibald-Heeren; Kristie Harrison; Annette Haworth; David Thwaites Journal: J Med Radiat Sci Date: 2020-08-02