Automated detection of critical findings in multi-parametric brain MRI using a system of 3D neural networks

Abstract With the rapid growth and increasing use of brain MRI, there is an interest in automated image classification to aid human interpretation and improve workflow. We aimed to train a deep convolutional neural network and assess its performance in identifying abnormal brain MRIs and critical in...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Kambiz Nael, Eli Gibson, Chen Yang, Pascal Ceccaldi, Youngjin Yoo, Jyotipriya Das, Amish Doshi, Bogdan Georgescu, Nirmal Janardhanan, Benjamin Odry, Mariappan Nadar, Michael Bush, Thomas J. Re, Stefan Huwer, Sonal Josan, Heinrich von Busch, Heiko Meyer, David Mendelson, Burton P. Drayer, Dorin Comaniciu, Zahi A. Fayad
Formato: article
Lenguaje:EN
Publicado: Nature Portfolio 2021
Materias:
R
Q
Acceso en línea:https://doaj.org/article/1dfb4bbd8eff4505b1963e02abf093e9
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:1dfb4bbd8eff4505b1963e02abf093e9
record_format dspace
spelling oai:doaj.org-article:1dfb4bbd8eff4505b1963e02abf093e92021-12-02T17:04:05ZAutomated detection of critical findings in multi-parametric brain MRI using a system of 3D neural networks10.1038/s41598-021-86022-72045-2322https://doaj.org/article/1dfb4bbd8eff4505b1963e02abf093e92021-03-01T00:00:00Zhttps://doi.org/10.1038/s41598-021-86022-7https://doaj.org/toc/2045-2322Abstract With the rapid growth and increasing use of brain MRI, there is an interest in automated image classification to aid human interpretation and improve workflow. We aimed to train a deep convolutional neural network and assess its performance in identifying abnormal brain MRIs and critical intracranial findings including acute infarction, acute hemorrhage and mass effect. A total of 13,215 clinical brain MRI studies were categorized to training (74%), validation (9%), internal testing (8%) and external testing (8%) datasets. Up to eight contrasts were included from each brain MRI and each image volume was reformatted to common resolution to accommodate for differences between scanners. Following reviewing the radiology reports, three neuroradiologists assigned each study to abnormal vs normal, and identified three critical findings including acute infarction, acute hemorrhage, and mass effect. A deep convolutional neural network was constructed by a combination of localization feature extraction (LFE) modules and global classifiers to identify the presence of 4 variables in brain MRIs including abnormal, acute infarction, acute hemorrhage and mass effect. Training, validation and testing sets were randomly defined on a patient basis. Training was performed on 9845 studies using balanced sampling to address class imbalance. Receiver operating characteristic (ROC) analysis was performed. The ROC analysis of our models for 1050 studies within our internal test data showed AUC/sensitivity/specificity of 0.91/83%/86% for normal versus abnormal brain MRI, 0.95/92%/88% for acute infarction, 0.90/89%/81% for acute hemorrhage, and 0.93/93%/85% for mass effect. For 1072 studies within our external test data, it showed AUC/sensitivity/specificity of 0.88/80%/80% for normal versus abnormal brain MRI, 0.97/90%/97% for acute infarction, 0.83/72%/88% for acute hemorrhage, and 0.87/79%/81% for mass effect. Our proposed deep convolutional network can accurately identify abnormal and critical intracranial findings on individual brain MRIs, while addressing the fact that some MR contrasts might not be available in individual studies.Kambiz NaelEli GibsonChen YangPascal CeccaldiYoungjin YooJyotipriya DasAmish DoshiBogdan GeorgescuNirmal JanardhananBenjamin OdryMariappan NadarMichael BushThomas J. ReStefan HuwerSonal JosanHeinrich von BuschHeiko MeyerDavid MendelsonBurton P. DrayerDorin ComaniciuZahi A. FayadNature PortfolioarticleMedicineRScienceQENScientific Reports, Vol 11, Iss 1, Pp 1-10 (2021)
institution DOAJ
collection DOAJ
language EN
topic Medicine
R
Science
Q
spellingShingle Medicine
R
Science
Q
Kambiz Nael
Eli Gibson
Chen Yang
Pascal Ceccaldi
Youngjin Yoo
Jyotipriya Das
Amish Doshi
Bogdan Georgescu
Nirmal Janardhanan
Benjamin Odry
Mariappan Nadar
Michael Bush
Thomas J. Re
Stefan Huwer
Sonal Josan
Heinrich von Busch
Heiko Meyer
David Mendelson
Burton P. Drayer
Dorin Comaniciu
Zahi A. Fayad
Automated detection of critical findings in multi-parametric brain MRI using a system of 3D neural networks
description Abstract With the rapid growth and increasing use of brain MRI, there is an interest in automated image classification to aid human interpretation and improve workflow. We aimed to train a deep convolutional neural network and assess its performance in identifying abnormal brain MRIs and critical intracranial findings including acute infarction, acute hemorrhage and mass effect. A total of 13,215 clinical brain MRI studies were categorized to training (74%), validation (9%), internal testing (8%) and external testing (8%) datasets. Up to eight contrasts were included from each brain MRI and each image volume was reformatted to common resolution to accommodate for differences between scanners. Following reviewing the radiology reports, three neuroradiologists assigned each study to abnormal vs normal, and identified three critical findings including acute infarction, acute hemorrhage, and mass effect. A deep convolutional neural network was constructed by a combination of localization feature extraction (LFE) modules and global classifiers to identify the presence of 4 variables in brain MRIs including abnormal, acute infarction, acute hemorrhage and mass effect. Training, validation and testing sets were randomly defined on a patient basis. Training was performed on 9845 studies using balanced sampling to address class imbalance. Receiver operating characteristic (ROC) analysis was performed. The ROC analysis of our models for 1050 studies within our internal test data showed AUC/sensitivity/specificity of 0.91/83%/86% for normal versus abnormal brain MRI, 0.95/92%/88% for acute infarction, 0.90/89%/81% for acute hemorrhage, and 0.93/93%/85% for mass effect. For 1072 studies within our external test data, it showed AUC/sensitivity/specificity of 0.88/80%/80% for normal versus abnormal brain MRI, 0.97/90%/97% for acute infarction, 0.83/72%/88% for acute hemorrhage, and 0.87/79%/81% for mass effect. Our proposed deep convolutional network can accurately identify abnormal and critical intracranial findings on individual brain MRIs, while addressing the fact that some MR contrasts might not be available in individual studies.
format article
author Kambiz Nael
Eli Gibson
Chen Yang
Pascal Ceccaldi
Youngjin Yoo
Jyotipriya Das
Amish Doshi
Bogdan Georgescu
Nirmal Janardhanan
Benjamin Odry
Mariappan Nadar
Michael Bush
Thomas J. Re
Stefan Huwer
Sonal Josan
Heinrich von Busch
Heiko Meyer
David Mendelson
Burton P. Drayer
Dorin Comaniciu
Zahi A. Fayad
author_facet Kambiz Nael
Eli Gibson
Chen Yang
Pascal Ceccaldi
Youngjin Yoo
Jyotipriya Das
Amish Doshi
Bogdan Georgescu
Nirmal Janardhanan
Benjamin Odry
Mariappan Nadar
Michael Bush
Thomas J. Re
Stefan Huwer
Sonal Josan
Heinrich von Busch
Heiko Meyer
David Mendelson
Burton P. Drayer
Dorin Comaniciu
Zahi A. Fayad
author_sort Kambiz Nael
title Automated detection of critical findings in multi-parametric brain MRI using a system of 3D neural networks
title_short Automated detection of critical findings in multi-parametric brain MRI using a system of 3D neural networks
title_full Automated detection of critical findings in multi-parametric brain MRI using a system of 3D neural networks
title_fullStr Automated detection of critical findings in multi-parametric brain MRI using a system of 3D neural networks
title_full_unstemmed Automated detection of critical findings in multi-parametric brain MRI using a system of 3D neural networks
title_sort automated detection of critical findings in multi-parametric brain mri using a system of 3d neural networks
publisher Nature Portfolio
publishDate 2021
url https://doaj.org/article/1dfb4bbd8eff4505b1963e02abf093e9
work_keys_str_mv AT kambiznael automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT eligibson automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT chenyang automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT pascalceccaldi automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT youngjinyoo automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT jyotipriyadas automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT amishdoshi automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT bogdangeorgescu automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT nirmaljanardhanan automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT benjaminodry automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT mariappannadar automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT michaelbush automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT thomasjre automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT stefanhuwer automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT sonaljosan automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT heinrichvonbusch automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT heikomeyer automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT davidmendelson automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT burtonpdrayer automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT dorincomaniciu automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
AT zahiafayad automateddetectionofcriticalfindingsinmultiparametricbrainmriusingasystemof3dneuralnetworks
_version_ 1718381835640111104