Nan Zhang1, Guang Yang2, Weiwei Zhang3, Wenjing Wang1, Zhen Zhou1, Heye Zhang3, Lei Xu4, Yundai Chen5. 1. Department of Radiology, Beijing Anzhen Hospital, Capital Medical University, 2(nd) Anzhen Road, Chaoyang District, Beijing, China. 2. Cardiovascular Research Centre, Royal Brompton Hospital, SW3 6NP, London, UK; National Heart and Lung Institute, Imperial College London, London, SW7 2AZ, UK. 3. School of Biomedical Engineering, Sun Yat-Sen University, China. 4. Department of Radiology, Beijing Anzhen Hospital, Capital Medical University, 2(nd) Anzhen Road, Chaoyang District, Beijing, China. Electronic address: leixu2001@hotmail.com. 5. Department of Cardiology, Chinese PLA General Hospital, Beijing, China.
Abstract
OBJECTIVES: To develop a fully automatic multiview shape constraint framework for comprehensive coronary artery calcium scores (CACS) quantification via deep learning on nonenhanced cardiac CT images. METHODS: In this retrospective single-centre study, a multi-task deep learning framework was proposed to detect and quantify coronary artery calcification from CT images collected between October 2018 and March 2019. A total of 232 non-contrast cardiac-gated CT scans were retrieved and studied (80 % for model training and 20 % for testing). CACS results of testing datasets (n = 46), including Agatston score, calcium volume score, calcium mass score, were calculated fully automatically and manually at total and vessel-specific levels, respectively. RESULTS: No significant differences were found in CACS quantification obtained using automatic or manual methods at total and vessel-specific levels (Agatston score: automatic 535.3 vs. manual 542.0, P = 0.993; calcium volume score: automatic 454.2 vs. manual 460.6, P = 0.990; calcium mass score: automatic 128.9 vs. manual 129.5, P = 0.992). Compared to the ground truth, the number of calcified vessels can be accurate recognized automatically (total: automatic 107 vs. manual 102, P = 0.125; left main artery: automatic 15 vs. manual 14, P = 1.000 ; left ascending artery: automatic 37 vs. manual 37, P = 1.000; left circumflex artery: automatic 22 vs. manual 20, P = 0.625; right coronary artery: automatic 33 vs. manual 31, P = 0.500). At the patient's level, there was no statistic difference existed in the classification of Agatston scoring (P = 0.317) and the number of calcified vessels (P = 0.102) between the automatic and manual results. CONCLUSIONS: The proposed framework can achieve reliable and comprehensive quantification for the CACS, including the calcified extent and distribution indicators at both total and vessel-specific levels.
OBJECTIVES: To develop a fully automatic multiview shape constraint framework for comprehensive coronary artery calcium scores (CACS) quantification via deep learning on nonenhanced cardiac CT images. METHODS: In this retrospective single-centre study, a multi-task deep learning framework was proposed to detect and quantify coronary artery calcification from CT images collected between October 2018 and March 2019. A total of 232 non-contrast cardiac-gated CT scans were retrieved and studied (80 % for model training and 20 % for testing). CACS results of testing datasets (n = 46), including Agatston score, calcium volume score, calcium mass score, were calculated fully automatically and manually at total and vessel-specific levels, respectively. RESULTS: No significant differences were found in CACS quantification obtained using automatic or manual methods at total and vessel-specific levels (Agatston score: automatic 535.3 vs. manual 542.0, P = 0.993; calcium volume score: automatic 454.2 vs. manual 460.6, P = 0.990; calcium mass score: automatic 128.9 vs. manual 129.5, P = 0.992). Compared to the ground truth, the number of calcified vessels can be accurate recognized automatically (total: automatic 107 vs. manual 102, P = 0.125; left main artery: automatic 15 vs. manual 14, P = 1.000 ; left ascending artery: automatic 37 vs. manual 37, P = 1.000; left circumflex artery: automatic 22 vs. manual 20, P = 0.625; right coronary artery: automatic 33 vs. manual 31, P = 0.500). At the patient's level, there was no statistic difference existed in the classification of Agatston scoring (P = 0.317) and the number of calcified vessels (P = 0.102) between the automatic and manual results. CONCLUSIONS: The proposed framework can achieve reliable and comprehensive quantification for the CACS, including the calcified extent and distribution indicators at both total and vessel-specific levels.
Authors: Philip Greenland; Joseph S Alpert; George A Beller; Emelia J Benjamin; Matthew J Budoff; Zahi A Fayad; Elyse Foster; Mark A Hlatky; John McB Hodgson; Frederick G Kushner; Michael S Lauer; Leslee J Shaw; Sidney C Smith; Allen J Taylor; William S Weintraub; Nanette K Wenger; Alice K Jacobs; Sidney C Smith; Jeffrey L Anderson; Nancy Albert; Christopher E Buller; Mark A Creager; Steven M Ettinger; Robert A Guyton; Jonathan L Halperin; Judith S Hochman; Frederick G Kushner; Rick Nishimura; E Magnus Ohman; Richard L Page; William G Stevenson; Lynn G Tarkington; Clyde W Yancy Journal: J Am Coll Cardiol Date: 2010-12-14 Impact factor: 24.094
Authors: Matthew J Budoff; Stephan Achenbach; Roger S Blumenthal; J Jeffrey Carr; Jonathan G Goldin; Philip Greenland; Alan D Guerci; Joao A C Lima; Daniel J Rader; Geoffrey D Rubin; Leslee J Shaw; Susan E Wiegers Journal: Circulation Date: 2006-10-02 Impact factor: 29.690
Authors: Maros Ferencik; Karol M Pencina; Ting Liu; Khristine Ghemigian; Kristin Baltrusaitis; Joseph M Massaro; Ralph B D'Agostino; Christopher J O'Donnell; Udo Hoffmann Journal: Circ Cardiovasc Imaging Date: 2017-10 Impact factor: 7.792
Authors: Mahmoud Al Rifai; Miguel Cainzos-Achirica; Sina Kianoush; Mohammadhassan Mirbolouk; Allison Peng; Josep Comin-Colet; Michael J Blaha Journal: Curr Treat Options Cardiovasc Med Date: 2018-09-26
Authors: Borek Foldyna; Parastou Eslami; Jan-Erik Scholtz; Kristin Baltrusaitis; Michael T Lu; Joseph M Massaro; Ralph B D'Agostino; Maros Ferencik; Hugo J W L Aerts; Christopher J O'Donnell; Udo Hoffmann Journal: Eur Radiol Date: 2019-05-02 Impact factor: 5.315
Authors: Omar Dzaye; Ramzi Dudum; Mohammadhassan Mirbolouk; Olusola A Orimoloye; Albert D Osei; Zeina A Dardari; Daniel S Berman; Michael D Miedema; Leslee Shaw; Alan Rozanski; Matthias Holdhoff; Khurram Nasir; John A Rumberger; Matthew J Budoff; Mouaz H Al-Mallah; Ron Blankstein; Michael J Blaha Journal: J Cardiovasc Comput Tomogr Date: 2019-03-28
Authors: Szilard Voros; Juan J Rivera; Daniel S Berman; Ron Blankstein; Matthew J Budoff; Ricardo C Cury; Milind Y Desai; Damini Dey; Sandra S Halliburton; Harvey S Hecht; Khurram Nasir; Raul D Santos; Michael D Shapiro; Allen J Taylor; Uma S Valeti; Phillip M Young; Gaby Weissman Journal: J Cardiovasc Comput Tomogr Date: 2011-01-20
Authors: Jelmer M Wolterink; Tim Leiner; Bob D de Vos; Jean-Louis Coatrieux; B Michael Kelm; Satoshi Kondo; Rodrigo A Salgado; Rahil Shahzad; Huazhong Shu; Miranda Snoeren; Richard A P Takx; Lucas J van Vliet; Theo van Walsum; Tineke P Willems; Guanyu Yang; Yefeng Zheng; Max A Viergever; Ivana Išgum Journal: Med Phys Date: 2016-05 Impact factor: 4.071
Authors: Mallory S Lo-Kioeng-Shioe; Andrea L Vavere; Armin Arbab-Zadeh; Joanne D Schuijf; Carlos E Rochitte; Marcus Y Chen; Matthias Rief; Klaus F Kofoed; Melvin E Clouse; Arthur J Scholte; Julie M Miller; Aisha Betoko; Michael J Blaha; Christopher Cox; Jaap W Deckers; Joao A C Lima Journal: J Am Heart Assoc Date: 2019-03-19 Impact factor: 5.501
Authors: Joo Hyeok Choi; Min Jae Cha; Iksung Cho; William D Kim; Yera Ha; Hyewon Choi; Sun Hwa Lee; Seng Chan You; Jee Suk Chang Journal: Front Oncol Date: 2022-09-20 Impact factor: 5.738