Transient Hemodynamics Prediction Using an Efficient Octree-Based Deep Learning Model

被引:2
作者
Maul, Noah [1 ,2 ]
Zinn, Katharina [1 ,2 ]
Wagner, Fabian [1 ]
Thies, Mareike [1 ]
Rohleder, Maximilian [1 ,2 ]
Pfaff, Laura [1 ,2 ]
Kowarschik, Markus [2 ]
Birkhold, Annette [2 ]
Maier, Andreas [1 ]
机构
[1] FAU Erlangen Nurnberg, Pattern Recognit Lab, Erlangen, Germany
[2] Siemens Healthcare GmbH, Forchheim, Germany
来源
INFORMATION PROCESSING IN MEDICAL IMAGING, IPMI 2023 | 2023年 / 13939卷
关键词
Hemodynamics; Octree; Operator learning; RATE WAVE-FORMS; BLOOD-FLOW;
D O I
10.1007/978-3-031-34048-2_15
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Patient-specific hemodynamics assessment could support diagnosis and treatment of neurovascular diseases. Currently, conventional medical imaging modalities are not able to accurately acquire high-resolution hemodynamic information that would be required to assess complex neurovascular pathologies. Instead, computational fluid dynamics (CFD) simulations can be applied to tomographic reconstructions to obtain clinically relevant information. However, three-dimensional (3D) CFD simulations require enormous computational resources and simulation-related expert knowledge that are usually not available in clinical environments. Recently, deep-learning-based methods have been proposed as CFD surrogates to improve computational efficiency. Nevertheless, the prediction of high-resolution transient CFD simulations for complex vascular geometries poses a challenge to conventional deep learning models. In this work, we present an architecture that is tailored to predict high-resolution (spatial and temporal) velocity fields for complex synthetic vascular geometries. For this, an octree-based spatial discretization is combined with an implicit neural function representation to efficiently handle the prediction of the 3D velocity field for each time step. The presented method is evaluated for the task of cerebral hemodynamics prediction before and during the injection of contrast agent in the internal carotid artery (ICA). Compared to CFD simulations, the velocity field can be estimated with a mean absolute error of 0.024 m s(-1), whereas the run time reduces from several hours on a high-performance cluster to a few seconds on a consumer graphical processing unit.
引用
收藏
页码:183 / 194
页数:12
相关论文
共 31 条
[21]   Rapid wall shear stress prediction for aortic aneurysms using deep learning: a fast alternative to CFD [J].
Faisal, Md. Ahasan Atick ;
Mutlu, Onur ;
Mahmud, Sakib ;
Tahir, Anas ;
Chowdhury, Muhammad E. H. ;
Bensaali, Faycal ;
Alnabti, Abdulrahman ;
Yavuz, Mehmet Metin ;
El-Menyar, Ayman ;
Al-Thani, Hassan ;
Yalcin, Huseyin Cagatay .
MEDICAL & BIOLOGICAL ENGINEERING & COMPUTING, 2025,
[22]   Deep learning-based hemodynamic prediction of carotid artery stenosis before and after surgical treatments [J].
Wang, Sirui ;
Wu, Dandan ;
Li, Gaoyang ;
Zhang, Zhiyuan ;
Xiao, Weizhong ;
Li, Ruichen ;
Qiao, Aike ;
Jin, Long ;
Liu, Hao .
FRONTIERS IN PHYSIOLOGY, 2023, 13
[23]   Deep learning acceleration of iterative model-based light fluence correction for photoacoustic tomography [J].
Liang, Zhaoyong ;
Zhang, Shuangyang ;
Liang, Zhichao ;
Mo, Zongxin ;
Zhang, Xiaoming ;
Zhong, Yutian ;
Chen, Wufan ;
Qi, Li .
PHOTOACOUSTICS, 2024, 37
[24]   A reduced-order model based on the coupled 1D-3D finite element simulations for an efficient analysis of hemodynamics problems [J].
Eduardo Soudah ;
Riccardo Rossi ;
Sergio Idelsohn ;
Eugenio Oñate .
Computational Mechanics, 2014, 54 :1013-1022
[25]   A reduced-order model based on the coupled 1D-3D finite element simulations for an efficient analysis of hemodynamics problems [J].
Soudah, Eduardo ;
Rossi, Riccardo ;
Idelsohn, Sergio ;
Onate, Eugenio .
COMPUTATIONAL MECHANICS, 2014, 54 (04) :1013-1022
[26]   A Deep Learning Approach to Using Wearable Seismocardiography (SCG) for Diagnosing Aortic Valve Stenosis and Predicting Aortic Hemodynamics Obtained by 4D Flow MRI [J].
Ebrahimkhani, Mahmoud ;
Johnson, Ethan M. I. ;
Sodhi, Aparna ;
Robinson, Joshua D. ;
Rigsby, Cynthia K. ;
Allen, Bradly D. ;
Markl, Michael .
ANNALS OF BIOMEDICAL ENGINEERING, 2023, 51 (12) :2802-2811
[27]   Prediction of renal allograft chronic rejection using a model based on contrast-enhanced ultrasonography [J].
Yang, Cheng ;
Wu, Shengdi ;
Yang, Ping ;
Shang, Guoguo ;
Qi, Ruochen ;
Xu, Ming ;
Rong, Ruiming ;
Zhu, Tongyu ;
He, Wanyuan .
MICROCIRCULATION, 2019, 26 (06)
[28]   Deep learning-based surrogate model for three-dimensional patient-specific computational fluid dynamics [J].
Du, Pan ;
Zhu, Xiaozhi ;
Wang, Jian-Xun .
PHYSICS OF FLUIDS, 2022, 34 (08)
[29]   Machine learning-based segmentation of ischemic penumbra by using diffusion tensor metrics in a rat model [J].
Kuo, Duen-Pang ;
Kuo, Po-Chih ;
Chen, Yung-Chieh ;
Kao, Yu-Chieh Jill ;
Lee, Ching-Yen ;
Chung, Hsiao-Wen ;
Chen, Cheng-Yu .
JOURNAL OF BIOMEDICAL SCIENCE, 2020, 27 (01)
[30]   Machine-Learning-Based Prediction of Photobiomodulation Effects on Older Adults With Cognitive Decline Using Functional Near-Infrared Spectroscopy [J].
Lee, Kyeonggu ;
Chun, Minyoung ;
Jung, Bori ;
Kim, Yunsu ;
Yang, Chaeyoun ;
Choi, Jongkwan ;
Cha, Jihyun ;
Lee, Seung-Hwan ;
Im, Chang-Hwan .
IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING, 2024, 32 :3710-3718