A deep learning pipeline for automatic analysis of multi-scan cardiovascular magnetic resonance

被引:24
|
作者
Fadil, Hakim [1 ]
Totman, John J. [1 ]
Hausenloy, Derek J. [2 ,3 ,4 ,5 ,6 ]
Ho, Hee-Hwa [7 ]
Joseph, Prabath [7 ]
Low, Adrian Fatt-Hoe [8 ]
Richards, A. Mark [9 ,10 ]
Chan, Mark Y. [4 ]
Marchesseau, Stephanie [1 ]
机构
[1] Natl Univ Singapore, Ctr Translat MR Res TMR, Singapore 117549, Singapore
[2] Duke Natl Univ Singapore Med Sch, Cardiovasc & Metab Disorders Program, Singapore 169857, Singapore
[3] Natl Heart Ctr, Natl Heart Res Inst Singapore, Singapore, Singapore
[4] Natl Univ Singapore, Dept Med, Yong Loo Lin SoM, Singapore 117597, Singapore
[5] UCL, Hatter Cardiovasc Inst, London, England
[6] Asia Univ, Coll Med & Hlth Sci, Cardiovasc Res Ctr, Taichung, Taiwan
[7] Tan Tock Seng Hosp, Singapore 308433, Singapore
[8] Natl Univ Heart Ctr, Singapore 119074, Singapore
[9] Natl Univ Singapore, Cardiovasc Res Inst, Singapore 119228, Singapore
[10] Univ Otago, Christchurch Heart Inst, Christchurch 8140, New Zealand
基金
英国医学研究理事会;
关键词
T1; mapping; T2; Cine short-axis; Aortic flow; Deep learning; Segmentation; Automatic analysis; IMAGE-ANALYSIS;
D O I
10.1186/s12968-020-00695-z
中图分类号
R5 [内科学];
学科分类号
1002 ; 100201 ;
摘要
Background Cardiovascular magnetic resonance (CMR) sequences are commonly used to obtain a complete description of the function and structure of the heart, provided that accurate measurements are extracted from images. New methods of extraction of information are being developed, among them, deep neural networks are powerful tools that showed the ability to perform fast and accurate segmentation. Iq1n order to reduce the time spent by reading physicians to process data and minimize intra- and inter-observer variability, we propose a fully automatic multi-scan CMR image analysis pipeline. Methods Sequence specific U-Net 2D models were trained to perform the segmentation of the left ventricle (LV), right ventricle (RV) and aorta in cine short-axis, late gadolinium enhancement (LGE), native T1 map, post-contrast T1, native T2 map and aortic flow sequences depending on the need. The models were trained and tested on a set of data manually segmented by experts using semi-automatic and manual tools. A set of parameters were computed from the resulting segmentations such as the left ventricular and right ventricular ejection fraction (EF), LGE scar percentage, the mean T1, T1 post, T2 values within the myocardium, and aortic flow. The Dice similarity coefficient, Hausdorff distance, mean surface distance, and Pearson correlation coefficient R were used to assess and compare the results of the U-Net based pipeline with intra-observer variability. Additionally, the pipeline was validated on two clinical studies. Results The sequence specific U-Net 2D models trained achieved fast (<= 0.2 s/image on GPU) and precise segmentation over all the targeted region of interest with high Dice scores (= 0.91 for LV, = 0.92 for RV, = 0.93 for Aorta in average) comparable to intra-observer Dice scores (= 0.86 for LV, = 0.87 for RV, = 0.95 for aorta flow in average). The automatically and manually computed parameters were highly correlated (R = 0.91 in average) showing results superior to the intra-observer variability (R = 0.85 in average) for every sequence presented here. Conclusion The proposed pipeline allows for fast and robust analysis of large CMR studies while guaranteeing reproducibility, hence potentially improving patient's diagnosis as well as clinical studies outcome.
引用
收藏
页数:13
相关论文
共 50 条
  • [21] Deep learning-based automated scan plane positioning for brain magnetic resonance imaging
    Zhu, Gaojie
    Shen, Xiongjie
    Sun, Zhiguo
    Xiao, Zhongjie
    Zhong, Junjie
    Yin, Zhe
    Li, Shengxiang
    Guo, Hua
    QUANTITATIVE IMAGING IN MEDICINE AND SURGERY, 2024, 14 (06) : 4015 - 4030
  • [22] Automatic classification of the vertebral endplate lesions in magnetic resonance imaging by deep learning model
    Bassani, Tito
    Cina, Andrea
    Galbusera, Fabio
    Sconfienza, Luca Maria
    Albano, Domenico
    Barcellona, Federica
    Colombini, Alessandra
    Luca, Andrea
    Brayda-Bruno, Marco
    FRONTIERS IN SURGERY, 2023, 10
  • [24] Automatic classification of brain magnetic resonance images with hypercolumn deep features and machine learning
    Kemal Akyol
    Physical and Engineering Sciences in Medicine, 2022, 45 : 935 - 947
  • [25] Automatic Detection of a Standard Line for Brain Magnetic Resonance Imaging Using Deep Learning
    Sugimori, Hiroyuki
    Kawakami, Masashi
    APPLIED SCIENCES-BASEL, 2019, 9 (18):
  • [26] A Deep Learning Model for Automatic Detection and Classification of Disc Herniation in Magnetic Resonance Images
    Sustersic, Tijana
    Rankovic, Vesna
    Milovanovic, Vladimir
    Kovacevic, Vojin
    Rasulic, Lukas
    Filipovic, Nenad
    IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS, 2022, 26 (12) : 6036 - 6046
  • [27] Lumbar Disc Herniation Automatic Detection in Magnetic Resonance Imaging Based on Deep Learning
    Tsai, Jen-Yung
    Hung, Isabella Yu-Ju
    Guo, Yue Leon
    Jan, Yih-Kuen
    Lin, Chih-Yang
    Shih, Tiffany Ting-Fang
    Chen, Bang-Bin
    Lung, Chi-Wen
    FRONTIERS IN BIOENGINEERING AND BIOTECHNOLOGY, 2021, 9
  • [28] A Deep Learning Pipeline for Automatic Skull Stripping and Brain Segmentation
    Yogananda, Chandan Ganesh Bangalore
    Wagner, Benjamin C.
    Murugesan, Gowtham K.
    Madhuranthakam, Ananth
    Maldjian, Joseph A.
    2019 IEEE 16TH INTERNATIONAL SYMPOSIUM ON BIOMEDICAL IMAGING (ISBI 2019), 2019, : 727 - 731
  • [29] Deep Learning-Based Automatic Segmentation of Brain Tumors in Heterogenous Multi-Center Magnetic Resonance Imaging Sets
    Liao, W.
    Luo, X.
    Zhang, S.
    INTERNATIONAL JOURNAL OF RADIATION ONCOLOGY BIOLOGY PHYSICS, 2024, 120 (02): : E251 - E251
  • [30] Using deep learning to predict cardiovascular magnetic resonance findings from echocardiography videos
    Sahashi, Y.
    Ouyang, D.
    Kwan, A.
    EUROPEAN HEART JOURNAL, 2024, 45