Noncontrast MRI in assessing venous reflux of legs using QFlow analysis and radial basis function neural network technique
Abstract Since venous reflux is difficult to quantify, triggered angiography non-contrast-enhanced (TRANCE)-magnetic resonance imaging (MRI) is a novel tool for objectively evaluating venous diseases in the lower extremities without using contrast media. This study included 26 pre-intervention patie...
Main Authors: | , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Nature Portfolio
2023-02-01
|
Series: | Scientific Reports |
Online Access: | https://doi.org/10.1038/s41598-023-30437-x |
_version_ | 1797864865146077184 |
---|---|
author | Min Yi Wong Chien-Wei Chen Yuan-Hsi Tseng Shao-Kui Zhou Yu-Hui Lin Yao-Kuang Huang Bor-Shyh Lin |
author_facet | Min Yi Wong Chien-Wei Chen Yuan-Hsi Tseng Shao-Kui Zhou Yu-Hui Lin Yao-Kuang Huang Bor-Shyh Lin |
author_sort | Min Yi Wong |
collection | DOAJ |
description | Abstract Since venous reflux is difficult to quantify, triggered angiography non-contrast-enhanced (TRANCE)-magnetic resonance imaging (MRI) is a novel tool for objectively evaluating venous diseases in the lower extremities without using contrast media. This study included 26 pre-intervention patients with superficial venous reflux in the lower extremities and 15 healthy volunteers. The quantitative flow (QFlow) analyzed the phase shift information from the pixels within the region of interest from MRI. The fast and simple radial basis function neural network (RBFNN) learning model is constructed by determining the parameters of the radial basis function and the weights of the neural network. The input parameters were the variables generated through QFlow, while the output variables were morbid limbs with venous reflux and normal limb classification. The stroke volume, forward flow volume, absolute stroke volume, mean flux, stroke distance, and mean velocity of greater saphenous veins from QFlow analysis could be used to discriminate the morbid limbs of pre-intervention patients and normal limbs of healthy controls. The neural network successfully classified the morbid and normal limbs with an accuracy of 90.24% in the training stage. The classification of venous reflux using the RBFNN model may assist physicians in clinical settings. |
first_indexed | 2024-04-09T22:58:51Z |
format | Article |
id | doaj.art-dcb327f13e344e47b74e27e577359d31 |
institution | Directory Open Access Journal |
issn | 2045-2322 |
language | English |
last_indexed | 2024-04-09T22:58:51Z |
publishDate | 2023-02-01 |
publisher | Nature Portfolio |
record_format | Article |
series | Scientific Reports |
spelling | doaj.art-dcb327f13e344e47b74e27e577359d312023-03-22T11:07:39ZengNature PortfolioScientific Reports2045-23222023-02-011311910.1038/s41598-023-30437-xNoncontrast MRI in assessing venous reflux of legs using QFlow analysis and radial basis function neural network techniqueMin Yi Wong0Chien-Wei Chen1Yuan-Hsi Tseng2Shao-Kui Zhou3Yu-Hui Lin4Yao-Kuang Huang5Bor-Shyh Lin6Division of Thoracic and Cardiovascular Surgery, Chiayi Chang Gung Memorial HospitalDepartment of Diagnostic Radiology, Chiayi Chang Gung Memorial HospitalDivision of Thoracic and Cardiovascular Surgery, Chiayi Chang Gung Memorial HospitalCollege of Photonics, National Yang Ming Chiao Tung UniversityDivision of Thoracic and Cardiovascular Surgery, Chiayi Chang Gung Memorial HospitalDivision of Thoracic and Cardiovascular Surgery, Chiayi Chang Gung Memorial HospitalCollege of Photonics, National Yang Ming Chiao Tung UniversityAbstract Since venous reflux is difficult to quantify, triggered angiography non-contrast-enhanced (TRANCE)-magnetic resonance imaging (MRI) is a novel tool for objectively evaluating venous diseases in the lower extremities without using contrast media. This study included 26 pre-intervention patients with superficial venous reflux in the lower extremities and 15 healthy volunteers. The quantitative flow (QFlow) analyzed the phase shift information from the pixels within the region of interest from MRI. The fast and simple radial basis function neural network (RBFNN) learning model is constructed by determining the parameters of the radial basis function and the weights of the neural network. The input parameters were the variables generated through QFlow, while the output variables were morbid limbs with venous reflux and normal limb classification. The stroke volume, forward flow volume, absolute stroke volume, mean flux, stroke distance, and mean velocity of greater saphenous veins from QFlow analysis could be used to discriminate the morbid limbs of pre-intervention patients and normal limbs of healthy controls. The neural network successfully classified the morbid and normal limbs with an accuracy of 90.24% in the training stage. The classification of venous reflux using the RBFNN model may assist physicians in clinical settings.https://doi.org/10.1038/s41598-023-30437-x |
spellingShingle | Min Yi Wong Chien-Wei Chen Yuan-Hsi Tseng Shao-Kui Zhou Yu-Hui Lin Yao-Kuang Huang Bor-Shyh Lin Noncontrast MRI in assessing venous reflux of legs using QFlow analysis and radial basis function neural network technique Scientific Reports |
title | Noncontrast MRI in assessing venous reflux of legs using QFlow analysis and radial basis function neural network technique |
title_full | Noncontrast MRI in assessing venous reflux of legs using QFlow analysis and radial basis function neural network technique |
title_fullStr | Noncontrast MRI in assessing venous reflux of legs using QFlow analysis and radial basis function neural network technique |
title_full_unstemmed | Noncontrast MRI in assessing venous reflux of legs using QFlow analysis and radial basis function neural network technique |
title_short | Noncontrast MRI in assessing venous reflux of legs using QFlow analysis and radial basis function neural network technique |
title_sort | noncontrast mri in assessing venous reflux of legs using qflow analysis and radial basis function neural network technique |
url | https://doi.org/10.1038/s41598-023-30437-x |
work_keys_str_mv | AT minyiwong noncontrastmriinassessingvenousrefluxoflegsusingqflowanalysisandradialbasisfunctionneuralnetworktechnique AT chienweichen noncontrastmriinassessingvenousrefluxoflegsusingqflowanalysisandradialbasisfunctionneuralnetworktechnique AT yuanhsitseng noncontrastmriinassessingvenousrefluxoflegsusingqflowanalysisandradialbasisfunctionneuralnetworktechnique AT shaokuizhou noncontrastmriinassessingvenousrefluxoflegsusingqflowanalysisandradialbasisfunctionneuralnetworktechnique AT yuhuilin noncontrastmriinassessingvenousrefluxoflegsusingqflowanalysisandradialbasisfunctionneuralnetworktechnique AT yaokuanghuang noncontrastmriinassessingvenousrefluxoflegsusingqflowanalysisandradialbasisfunctionneuralnetworktechnique AT borshyhlin noncontrastmriinassessingvenousrefluxoflegsusingqflowanalysisandradialbasisfunctionneuralnetworktechnique |