Wei Zhao1, Bin Han2, Yong Yang3, Mark Buyyounouski4, Steven L Hancock5, Hilary Bagshaw6, Lei Xing7. 1. Stanford University, Department of Radiation Oncology, Stanford, USA. Electronic address: zhaow85@stanford.edu. 2. Stanford University, Department of Radiation Oncology, Stanford, USA. Electronic address: hanbin@stanford.edu. 3. Stanford University, Department of Radiation Oncology, Stanford, USA. Electronic address: yongy66@stanford.edu. 4. Stanford University, Department of Radiation Oncology, Stanford, USA. Electronic address: mbuyyou@stanford.edu. 5. Stanford University, Department of Radiation Oncology, Stanford, USA. Electronic address: shancock@stanford.edu. 6. Stanford University, Department of Radiation Oncology, Stanford, USA. Electronic address: hbagshaw@stanford.edu. 7. Stanford University, Department of Radiation Oncology, Stanford, USA. Electronic address: lei@stanford.edu.
Abstract
BACKGROUND AND PURPOSE: To investigate a novel markerless prostate localization strategy using a pre-trained deep learning model to interpret routine projection kilovoltage (kV) X-ray images in image-guided radiation therapy (IGRT). MATERIALS AND METHODS: We developed a personalized region-based convolutional neural network to localize the prostate treatment target without implanted fiducials. To train the deep neural network (DNN), we used the patient's planning computed tomography (pCT) images with pre-delineated prostate target to generate a large amount of synthetic kV projection X-ray images in the geometry of onboard imager (OBI) system. The DNN model was evaluated by retrospectively studying 10 patients who underwent prostate IGRT. Three out of the ten patients who had implanted fiducials and the fiducials' positions in the OBI images acquired for treatment setup were examined to show the potential of the proposed method for prostate IGRT. Statistical analysis using Lin's concordance correlation coefficient was calculated to assess the results along with the difference between the digitally reconstructed radiographs (DRR) derived and DNN predicted locations of the prostate. RESULTS: Differences between the predicted target positions using DNN and their actual positions are (mean ± standard deviation) 1.58 ± 0.43 mm, 1.64 ± 0.43 mm, and 1.67 ± 0.36 mm in anterior-posterior, lateral, and oblique directions, respectively. Prostate position identified on the OBI kV images is also found to be consistent with that derived from the implanted fiducials. CONCLUSIONS: Highly accurate, markerless prostate localization based on deep learning is achievable. The proposed method is useful for daily patient positioning and real-time target tracking during prostate radiotherapy.
BACKGROUND AND PURPOSE: To investigate a novel markerless prostate localization strategy using a pre-trained deep learning model to interpret routine projection kilovoltage (kV) X-ray images in image-guided radiation therapy (IGRT). MATERIALS AND METHODS: We developed a personalized region-based convolutional neural network to localize the prostate treatment target without implanted fiducials. To train the deep neural network (DNN), we used the patient's planning computed tomography (pCT) images with pre-delineated prostate target to generate a large amount of synthetic kV projection X-ray images in the geometry of onboard imager (OBI) system. The DNN model was evaluated by retrospectively studying 10 patients who underwent prostate IGRT. Three out of the ten patients who had implanted fiducials and the fiducials' positions in the OBI images acquired for treatment setup were examined to show the potential of the proposed method for prostate IGRT. Statistical analysis using Lin's concordance correlation coefficient was calculated to assess the results along with the difference between the digitally reconstructed radiographs (DRR) derived and DNN predicted locations of the prostate. RESULTS: Differences between the predicted target positions using DNN and their actual positions are (mean ± standard deviation) 1.58 ± 0.43 mm, 1.64 ± 0.43 mm, and 1.67 ± 0.36 mm in anterior-posterior, lateral, and oblique directions, respectively. Prostate position identified on the OBI kV images is also found to be consistent with that derived from the implanted fiducials. CONCLUSIONS: Highly accurate, markerless prostate localization based on deep learning is achievable. The proposed method is useful for daily patient positioning and real-time target tracking during prostate radiotherapy.
Authors: Lei Xing; Brian Thorndyke; Eduard Schreibmann; Yong Yang; Tian-Fang Li; Gwe-Ya Kim; Gary Luxton; Albert Koong Journal: Med Dosim Date: 2006 Impact factor: 1.482
Authors: Jin Aun Ng; Jeremy T Booth; Per R Poulsen; Walther Fledelius; Esben Schjødt Worm; Thomas Eade; Fiona Hegi; Andrew Kneebone; Zdenka Kuncic; Paul J Keall Journal: Int J Radiat Oncol Biol Phys Date: 2012-09-11 Impact factor: 7.038
Authors: Colien Hazelaar; Lineke van der Weide; Hassan Mostafavi; Ben J Slotman; Wilko F A R Verbakel; Max Dahele Journal: Radiother Oncol Date: 2018-08-29 Impact factor: 6.280
Authors: Colien Hazelaar; Wilko F A R Verbakel; Hassan Mostafavi; Lineke van der Weide; Ben J Slotman; Max Dahele Journal: Int J Radiat Oncol Biol Phys Date: 2018-04-22 Impact factor: 7.038
Authors: Alexis N T J Kotte; Pieter Hofman; Jan J W Lagendijk; Marco van Vulpen; Uulke A van der Heide Journal: Int J Radiat Oncol Biol Phys Date: 2007-05-21 Impact factor: 7.038