PMID- 34275582
OWN - NLM
STAT- MEDLINE
DCOM- 20210728
LR  - 20210728
IS  - 1532-2653 (Electronic)
IS  - 0967-5868 (Linking)
VI  - 90
DP  - 2021 Aug
TI  - Using deep learning convolutional neural networks to automatically perform 
      cerebral aqueduct CSF flow analysis.
PG  - 60-67
LID - S0967-5868(21)00218-6 [pii]
LID - 10.1016/j.jocn.2021.05.010 [doi]
AB  - Since the development of phase-contrast magnetic resonance imaging (PC-MRI), 
      quantification of cerebrospinal fluid (CSF) flow across the cerebral aqueduct has 
      been utilized for diagnosis of conditions such as normal pressure hydrocephalus 
      (NPH). This study aims to develop an automated method of aqueduct CSF flow 
      analysis using convolution neural networks (CNNs), which can replace the current 
      standard involving manual segmentation of aqueduct region of interest (ROI). 
      Retrospective analysis was performed on 333 patients who underwent PC-MRI, 
      totaling 353 imaging studies. Aqueduct flow measurements using manual ROI 
      placement was performed independently by two radiologists. Two types of CNNs, 
      MultiResUNet and UNet, were trained using ROI data from the senior radiologist, 
      with PC-MRI studies being randomly divided into training (80%) and validation 
      (20%) datasets. Segmentation performance was assessed using Dice similarity 
      coefficient (DSC), and CSF flow parameters were calculated from both manual and 
      CNN-derived ROIs. MultiResUNet, UNet and second radiologist (Rater 2) had DSCs of 
      0.933, 0.928, and 0.867, respectively, with p < 0.001 between CNNs and Rater 2. 
      Comparison of CSF flow parameters showed excellent intraclass correlation 
      coefficients (ICCs) for MultiResUNet, with lowest correlation being 0.67. For 
      UNet, lower ICCs of -0.01 to 0.56 were observed. Only 3/353 (0.8%) studies failed 
      to have appropriate ROIs placed by MultiResUNet, compared to 12/353 (3.4%) failed 
      cases for UNet. In conclusion, CNNs were able to measure aqueductal CSF flow with 
      similar performance to a senior neuroradiologist. MultiResUNet demonstrated fewer 
      cases of segmentation failure and more consistent flow measurements compared to 
      the widely adopted UNet.
CI  - Copyright (c) 2021 Elsevier Ltd. All rights reserved.
FAU - Tsou, Cheng-Hsien
AU  - Tsou CH
AD  - Department of Radiology, Taichung Veterans General Hospital, 1650 Sect. 4 Taiwan 
      Boulevard, Taichung 40705, Taiwan, ROC.
FAU - Cheng, Yun-Chung
AU  - Cheng YC
AD  - Department of Neuroradiology, Department of Radiology, Taichung Veterans General 
      Hospital, 1650 Sect. 4 Taiwan Boulevard, Taichung 40705, Taiwan, ROC; Department 
      of Industrial Engineering and Enterprise Information, Tunghai University, 
      Taichung 407224, Taiwan, ROC. Electronic address: iancheng@vghtc.gov.tw.
FAU - Huang, Chin-Yin
AU  - Huang CY
AD  - Department of Industrial Engineering and Enterprise Information, Tunghai 
      University, Taichung 407224, Taiwan, ROC.
FAU - Chen, Jeon-Hor
AU  - Chen JH
AD  - Department of Radiological Sciences, University of California, Irvine, CA, United 
      States; Department of Radiology, Eda Hospital, No. 1, Yida Rd., Jiaosu Village, 
      Yanchao District, Kaohsiung 82445, Taiwan, ROC.
FAU - Chen, Wen-Hsien
AU  - Chen WH
AD  - Department of Neuroradiology, Department of Radiology, Taichung Veterans General 
      Hospital, 1650 Sect. 4 Taiwan Boulevard, Taichung 40705, Taiwan, ROC.
FAU - Chai, Jyh-Wen
AU  - Chai JW
AD  - Department of Radiology, Taichung Veterans General Hospital, 1650 Sect. 4 Taiwan 
      Boulevard, Taichung 40705, Taiwan, ROC.
FAU - Chen, Clayton Chi-Chang
AU  - Chen CC
AD  - Department of Neuroradiology, Department of Radiology, Taichung Veterans General 
      Hospital, 1650 Sect. 4 Taiwan Boulevard, Taichung 40705, Taiwan, ROC.
LA  - eng
PT  - Journal Article
DEP - 20210524
PL  - Scotland
TA  - J Clin Neurosci
JT  - Journal of clinical neuroscience : official journal of the Neurosurgical Society 
      of Australasia
JID - 9433352
SB  - IM
MH  - Adolescent
MH  - Adult
MH  - Aged
MH  - Aged, 80 and over
MH  - Cerebral Aqueduct/*diagnostic imaging
MH  - Child
MH  - Child, Preschool
MH  - Cross-Sectional Studies
MH  - *Deep Learning
MH  - Female
MH  - Humans
MH  - Hydrocephalus, Normal Pressure/*cerebrospinal fluid/diagnostic imaging
MH  - Infant
MH  - Infant, Newborn
MH  - Magnetic Resonance Imaging/*methods
MH  - Male
MH  - Middle Aged
MH  - *Neural Networks, Computer
MH  - Retrospective Studies
MH  - Young Adult
OTO - NOTNLM
OT  - Cerebral aqueduct
OT  - Cerebrospinal fluid
OT  - Deep learning
OT  - Magnetic resonance imaging
COIS- Declaration of Competing Interest The authors declare that they have no known 
      competing financial interests or personal relationships that could have appeared 
      to influence the work reported in this paper.
EDAT- 2021/07/20 06:00
MHDA- 2021/07/29 06:00
CRDT- 2021/07/19 05:34
PHST- 2021/03/04 00:00 [received]
PHST- 2021/04/11 00:00 [revised]
PHST- 2021/05/01 00:00 [accepted]
PHST- 2021/07/19 05:34 [entrez]
PHST- 2021/07/20 06:00 [pubmed]
PHST- 2021/07/29 06:00 [medline]
AID - S0967-5868(21)00218-6 [pii]
AID - 10.1016/j.jocn.2021.05.010 [doi]
PST - ppublish
SO  - J Clin Neurosci. 2021 Aug;90:60-67. doi: 10.1016/j.jocn.2021.05.010. Epub 2021 
      May 24.