Automatic segmentation of multiple cardiovascular structures from cardiac computed tomography angiography images using deep learning

被引：24

作者：

Baskaran, Lohendran ^{[1
,2
,3
,4
]}

Al'Aref, Subhi J. ^{[1
,2
,3
]}

Maliakal, Gabriel ^{[5
]}

Lee, Benjamin C. ^{[1
]}

Xu, Zhuoran ^{[1
]}

Choi, Jeong W. ^{[1
]}

Lee, Sang-Eun ^{[6
,7
]}

Sung, Ji Min ^{[6
]}

Lin, Fay Y. ^{[1
,2
,3
]}

Dunham, Simon ^{[1
]}

Mosadegh, Bobak ^{[1
]}

Kim, Yong-Jin ^{[8
]}

Gottlieb, Ilan ^{[9
]}

Lee, Byoung Kwon ^{[10
]}

Chun, Eun Ju ^{[11
]}

Cademartiri, Filippo ^{[12
]}

Maffei, Erica ^{[13
]}

Marques, Hugo ^{[14
]}

Shin, Sanghoon ^{[7
]}

Choi, Jung Hyun ^{[15
]}

Chinnaiyan, Kavitha ^{[16
]}

Hadamitzky, Martin ^{[17
]}

Conte, Edoardo ^{[18
]}

Andreini, Daniele ^{[18
]}

Pontone, Gianluca ^{[18
]}

Budoff, Matthew J. ^{[19
]}

Leipsic, Jonathon A. ^{[20
]}

Raff, Gilbert L. ^{[16
]}

Virmani, Renu ^{[16
,21
]}

Samady, Habib ^{[22
]}

Stone, Peter H. ^{[23
]}

Berman, Daniel S. ^{[24
]}

Narula, Jagat ^{[25
,26
]}

Bax, Jeroen J. ^{[27
]}

Chang, Hyuk-Jae ^{[6
]}

Min, James K. ^{[1
,2
,3
]}

Shaw, Leslee J. ^{[1
,2
,3
]}

机构：

[1] Weill Cornell Med, Dalio Inst Cardiovasc Imaging, New York, NY 10065 USA

[2] New York Presbyterian Hosp, Dept Radiol, New York, NY 10038 USA

[3] Weill Cornell Med, New York, NY 10065 USA

[4] Natl Heart Ctr, Dept Cardiovasc Med, Singapore, Singapore

[5] Cleerly Inc, New York, NY USA

[6] Yonsei Univ, Integrat Cardiovasc Imaging Ctr, Severance Cardiovasc Hosp, Div Cardiol,Coll Med, Seoul, South Korea

[7] Ewha Womans Univ, Dept Internal Med, Div Cardiol, Seoul Hosp, Seoul, South Korea

[8] Seoul Natl Univ, Seoul Natl Univ Hosp, Cardiovasc Ctr, Dept Internal Med,Coll Med, Seoul, South Korea

[9] Casa Saude Sao Jose, Dept Radiol, Rio De Janeiro, Brazil

[10] Yonsei Univ, Gangnam Severance Hosp, Div Cardiol, Coll Med, Seoul, South Korea

[11] Seoul Natl Univ, Dept Radiol, Bundang Hosp, Sungnam, South Korea

[12] SDN IRCCS, Cardiovasc Imaging Ctr, Naples, Italy

[13] ASUR Marche, Dept Radiol, Area Vasta 1, Urbino, Italy

[14] Hosp Luz, Unit Cardiovasc Imaging, UNICA, Lisbon, Portugal

[15] Pusan Univ Hosp, Busan, South Korea

[16] William Beaumont Hosp, Dept Cardiol, Royal Oak, MI 48072 USA

[17] German Heart Ctr Munich, Dept Radiol & Nucl Med, Munich, Germany

[18] IRCCS, Ctr Cardiol Monzino, Milan, Italy

[19] Los Angeles Biomed Res Inst, Dept Med, Torrance, CA USA

[20] Univ British Columbia, Dept Med & Radiol, Vancouver, BC, Canada

[21] CVPath Inst, Dept Pathol, Gaithersburg, MD USA

[22] Emory Univ, Sch Med, Div Cardiol, Atlanta, GA 30322 USA

[23] Harvard Med Sch, Brigham & Womens Hosp, Cardiovasc Div, Boston, MA 02115 USA

[24] Cedars Sinai Med Ctr, Dept Imaging & Med, Los Angeles, CA 90048 USA

[25] Icahn Sch Med Mt Sinai, Mt Sinai Heart Zena & Michael Wiener Cardiovasc I, New York, NY 10029 USA

[26] Icahn Sch Med Mt Sinai, Henry R Kravis Ctr Cardiovasc Hlth, New York, NY 10029 USA

[27] Leiden Univ, Dept Cardiol, Med Ctr, Leiden, Netherlands

来源：

PLOS ONE | 2020年 / 15卷 / 05期

基金：

美国国家卫生研究院;

关键词：

HEART;

D O I：

10.1371/journal.pone.0232573

中图分类号：

O [数理科学和化学]; P [天文学、地球科学]; Q [生物科学]; N [自然科学总论];

学科分类号：

07 ; 0710 ; 09 ;

摘要：

Background Segmentation of cardiovascular images is resource-intensive. We design an automated deep learning method for the segmentation of multiple structures from Coronary Computed Tomography Angiography (CCTA) images. Methods Images from a multicenter registry of patients that underwent clinically-indicated CCTA were used. The proximal ascending and descending aorta (PAA, DA), superior and inferior vena cavae (SVC, IVC), pulmonary artery (PA), coronary sinus (CS), right ventricular wall (RVW) and left atrial wall (LAW) were annotated as ground truth. The U-net-derived deep learning model was trained, validated and tested in a 70:20:10 split. Results The dataset comprised 206 patients, with 5.130 billion pixels. Mean age was 59.9 +/- 9.4 yrs., and was 42.7% female. An overall median Dice score of 0.820 (0.782, 0.843) was achieved. Median Dice scores for PAA, DA, SVC, IVC, PA, CS, RVW and LAW were 0.969 (0.979, 0.988), 0.953 (0.955, 0.983), 0.937 (0.934, 0.965), 0.903 (0.897, 0.948), 0.775 (0.724, 0.925), 0.720 (0.642, 0.809), 0.685 (0.631, 0.761) and 0.625 (0.596, 0.749) respectively. Apart from the CS, there were no significant differences in performance between sexes or age groups. Conclusions An automated deep learning model demonstrated segmentation of multiple cardiovascular structures from CCTA images with reasonable overall accuracy when evaluated on a pixel level.

引用

页数：13

共 26 条

[1] SCCT guidelines for the performance and acquisition of coronary computed tomographic angiography: A report of the society of Cardiovascular Computed Tomography Guidelines Committee Endorsed by the North American Society for Cardiovascular Imaging (NASCI) [J].

Abbara, Suhny ;

Blanke, Philipp ;

Maroules, Christopher D. ;

Cheezum, Michael ;

Choi, Andrew D. ;

Han, B. Kelly ;

Marwan, Mohamed ;

Naoum, Chris ;

Norgaard, Bjarne L. ;

Rubinshtein, Ronen ;

Schoenhagen, Paul ;

Villines, Todd ;

Leipsic, Jonathon .

JOURNAL OF CARDIOVASCULAR COMPUTED TOMOGRAPHY, 2016, 10 (06) :435-449

[2]

[Anonymous], 2019, PYTHON PACKAGE READI

[3]

Baskaran L, 2019, JACC CARDIOVASCULAR

[4]

Cho J., 2015, ARXIV151106348CS

[5] Segmentation of the heart and great vessels in CT images using a model-based adaptation framework [J].

Ecabert, Olivier ;

Peters, Jochen ;

Walker, Matthew J. ;

Ivanc, Thomas ;

Lorenz, Cristian ;

von Berg, Jens ;

Lessick, Jonathan ;

Vembar, Mani ;

Weese, Juergen .

MEDICAL IMAGE ANALYSIS, 2011, 15 (06) :863-876

[6] Normal values of left ventricular mass and cardiac chamber volumes assessed by 320-detector computed tomography angiography in the Copenhagen General Population Study [J].

Fuchs, Andreas ;

Mejdahl, Mads R. ;

Kuhl, J. Tobias ;

Stisen, Zara R. ;

Nilsson, Emma Julia P. ;

Kober, Lars V. ;

Nordestgaard, Borge G. ;

Kofoed, Klaus F. .

EUROPEAN HEART JOURNAL-CARDIOVASCULAR IMAGING, 2016, 17 (09) :1009-1017

[7] Automated 3D segmentation and diameter measurement of the thoracic aorta on non-contrast enhanced CT [J].

Gamechi, Zahra Sedghi ;

Bons, Lidia R. ;

Giordano, Marco ;

Bos, Daniel ;

Budde, Ricardo P. J. ;

Kofoed, Klaus F. ;

Pedersen, Jesper Holst ;

Roos-Hesselink, Jolien W. ;

de Bruijne, Marleen .

EUROPEAN RADIOLOGY, 2019, 29 (09) :4613-4623

[8] Deep Learning-A Technology With the Potential to Transform Health Care [J].

Hinton, Geoffrey .

JAMA-JOURNAL OF THE AMERICAN MEDICAL ASSOCIATION, 2018, 320 (11) :1101-1102

[9] Multi-organ Segmentation of Chest CT Images in Radiation Oncology: Comparison of Standard and Dilated UNet [J].

Javaid, Umair ;

Dasnoy, Damien ;

Lee, John A. .

ADVANCED CONCEPTS FOR INTELLIGENT VISION SYSTEMS, ACIVS 2018, 2018, 11182 :188-199

[10]

Kivinen JJ, VISUAL BOUNDARY PRED, V10

← 1 2 3 →