Eric D Morris1,2, Ahmed I Ghanem1,3, Ming Dong4, Milan V Pantelic5, Eleanor M Walker1, Carri K Glide-Hurst1,2. 1. Department of Radiation Oncology, Henry Ford Cancer Institute, Detroit, MI, USA. 2. Department of Radiation Oncology, Wayne State University School of Medicine, Detroit, MI, USA. 3. Department of Clinical Oncology, Alexandria University, Alexandria, Egypt. 4. Department of Computer Science, Wayne State University, Detroit, MI, USA. 5. Department of Radiology, Henry Ford Cancer Institute, Detroit, MI, USA.
Abstract
PURPOSE: Radiation dose to cardiac substructures is related to radiation-induced heart disease. However, substructures are not considered in radiation therapy planning (RTP) due to poor visualization on CT. Therefore, we developed a novel deep learning (DL) pipeline leveraging MRI's soft tissue contrast coupled with CT for state-of-the-art cardiac substructure segmentation requiring a single, non-contrast CT input. MATERIALS/ METHODS: Thirty-two left-sided whole-breast cancer patients underwent cardiac T2 MRI and CT-simulation. A rigid cardiac-confined MR/CT registration enabled ground truth delineations of 12 substructures (chambers, great vessels (GVs), coronary arteries (CAs), etc.). Paired MRI/CT data (25 patients) were placed into separate image channels to train a three-dimensional (3D) neural network using the entire 3D image. Deep supervision and a Dice-weighted multi-class loss function were applied. Results were assessed pre/post augmentation and post-processing (3D conditional random field (CRF)). Results for 11 test CTs (seven unique patients) were compared to ground truth and a multi-atlas method (MA) via Dice similarity coefficient (DSC), mean distance to agreement (MDA), and Wilcoxon signed-ranks tests. Three physicians evaluated clinical acceptance via consensus scoring (5-point scale). RESULTS: The model stabilized in ~19 h (200 epochs, training error <0.001). Augmentation and CRF increased DSC 5.0 ± 7.9% and 1.2 ± 2.5%, across substructures, respectively. DL provided accurate segmentations for chambers (DSC = 0.88 ± 0.03), GVs (DSC = 0.85 ± 0.03), and pulmonary veins (DSC = 0.77 ± 0.04). Combined DSC for CAs was 0.50 ± 0.14. MDA across substructures was <2.0 mm (GV MDA = 1.24 ± 0.31 mm). No substructures had statistical volume differences (P > 0.05) to ground truth. In four cases, DL yielded left main CA contours, whereas MA segmentation failed, and provided improved consensus scores in 44/60 comparisons to MA. DL provided clinically acceptable segmentations for all graded patients for 3/4 chambers. DL contour generation took ~14 s per patient. CONCLUSIONS: These promising results suggest DL poses major efficiency and accuracy gains for cardiac substructure segmentation offering high potential for rapid implementation into RTP for improved cardiac sparing.
PURPOSE: Radiation dose to cardiac substructures is related to radiation-induced heart disease. However, substructures are not considered in radiation therapy planning (RTP) due to poor visualization on CT. Therefore, we developed a novel deep learning (DL) pipeline leveraging MRI's soft tissue contrast coupled with CT for state-of-the-art cardiac substructure segmentation requiring a single, non-contrast CT input. MATERIALS/ METHODS: Thirty-two left-sided whole-breast cancerpatients underwent cardiac T2 MRI and CT-simulation. A rigid cardiac-confined MR/CT registration enabled ground truth delineations of 12 substructures (chambers, great vessels (GVs), coronary arteries (CAs), etc.). Paired MRI/CT data (25 patients) were placed into separate image channels to train a three-dimensional (3D) neural network using the entire 3D image. Deep supervision and a Dice-weighted multi-class loss function were applied. Results were assessed pre/post augmentation and post-processing (3D conditional random field (CRF)). Results for 11 test CTs (seven unique patients) were compared to ground truth and a multi-atlas method (MA) via Dice similarity coefficient (DSC), mean distance to agreement (MDA), and Wilcoxon signed-ranks tests. Three physicians evaluated clinical acceptance via consensus scoring (5-point scale). RESULTS: The model stabilized in ~19 h (200 epochs, training error <0.001). Augmentation and CRF increased DSC 5.0 ± 7.9% and 1.2 ± 2.5%, across substructures, respectively. DL provided accurate segmentations for chambers (DSC = 0.88 ± 0.03), GVs (DSC = 0.85 ± 0.03), and pulmonary veins (DSC = 0.77 ± 0.04). Combined DSC for CAs was 0.50 ± 0.14. MDA across substructures was <2.0 mm (GV MDA = 1.24 ± 0.31 mm). No substructures had statistical volume differences (P > 0.05) to ground truth. In four cases, DL yielded left main CA contours, whereas MA segmentation failed, and provided improved consensus scores in 44/60 comparisons to MA. DL provided clinically acceptable segmentations for all graded patients for 3/4 chambers. DL contour generation took ~14 s per patient. CONCLUSIONS: These promising results suggest DL poses major efficiency and accuracy gains for cardiac substructure segmentation offering high potential for rapid implementation into RTP for improved cardiac sparing.
Authors: Geert Litjens; Thijs Kooi; Babak Ehteshami Bejnordi; Arnaud Arindra Adiyoso Setio; Francesco Ciompi; Mohsen Ghafoorian; Jeroen A W M van der Laak; Bram van Ginneken; Clara I Sánchez Journal: Med Image Anal Date: 2017-07-26 Impact factor: 8.545
Authors: Robert Kaderka; Erin F Gillespie; Robert C Mundt; Alex K Bryant; Camila B Sanudo-Thomas; Anna L Harrison; Emilie L Wouters; Vitali Moiseenko; Kevin L Moore; Todd F Atwood; James D Murphy Journal: Radiother Oncol Date: 2018-08-11 Impact factor: 6.280
Authors: Eric D Morris; Ahmed I Ghanem; Milan V Pantelic; Eleanor M Walker; Xiaoxia Han; Carri K Glide-Hurst Journal: Int J Radiat Oncol Biol Phys Date: 2018-11-22 Impact factor: 7.038
Authors: Martin Rajchl; Matthew C H Lee; Ozan Oktay; Konstantinos Kamnitsas; Jonathan Passerat-Palmbach; Wenjia Bai; Mellisa Damodaram; Mary A Rutherford; Joseph V Hajnal; Bernhard Kainz; Daniel Rueckert Journal: IEEE Trans Med Imaging Date: 2016-11-09 Impact factor: 10.048
Authors: S Vivekanandan; D B Landau; N Counsell; D R Warren; A Khwanda; S D Rosen; E Parsons; Y Ngai; L Farrelly; L Hughes; M A Hawkins; J D Fenwick Journal: Int J Radiat Oncol Biol Phys Date: 2017-04-27 Impact factor: 7.038
Authors: Inas A Yassine; Ahmed M Ghanem; Nader S Metwalli; Ahmed Hamimi; Ronald Ouwerkerk; Jatin R Matta; Michael A Solomon; Jason M Elinoff; Ahmed M Gharib; Khaled Z Abd-Elmoniem Journal: Comput Biol Med Date: 2021-11-18 Impact factor: 4.589
Authors: Yabo Fu; Hao Zhang; Eric D Morris; Carri K Glide-Hurst; Suraj Pai; Alberto Traverso; Leonard Wee; Ibrahim Hadzic; Per-Ivar Lønne; Chenyang Shen; Tian Liu; Xiaofeng Yang Journal: IEEE Trans Radiat Plasma Med Sci Date: 2021-08-24
Authors: Shadab Momin; Yang Lei; Neal S McCall; Jiahan Zhang; Justin Roper; Joseph Harms; Sibo Tian; Michael S Lloyd; Tian Liu; Jeffrey D Bradley; Kristin Higgins; Xiaofeng Yang Journal: Phys Med Biol Date: 2022-05-11 Impact factor: 4.174
Authors: Robert Neil Finnegan; Lucia Orlandini; Xiongfei Liao; Jun Yin; Jinyi Lang; Jason Dowling; Davide Fontanarosa Journal: PLoS One Date: 2021-01-14 Impact factor: 3.240
Authors: Eric D Morris; Ahmed I Ghanem; Simeng Zhu; Ming Dong; Milan V Pantelic; Carri K Glide-Hurst Journal: Phys Imaging Radiat Oncol Date: 2021-04-16