Fully automated identification of brain abnormality from whole-body FDG-PET imaging using deep learning-based brain extraction and statistical parametric mapping

Abstract Background The whole brain is often covered in [18F]Fluorodeoxyglucose positron emission tomography ([18F]FDG-PET) in oncology patients, but the covered brain abnormality is typically screened by visual interpretation without quantitative analysis in clinical practice. In this study, we aim...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Wonseok Whi, Hongyoon Choi, Jin Chul Paeng, Gi Jeong Cheon, Keon Wook Kang, Dong Soo Lee
Formato:	article
Lenguaje:	EN
Publicado:	SpringerOpen 2021
Materias:	Brain segmentation Quantitative PET analysis Deep learning Convolutional neural network FDG-PET Brain FDG-PET Medical physics. Medical radiology. Nuclear medicine R895-920
Acceso en línea:	https://doaj.org/article/a28bd4daf3b3432e8414859dba25bda2
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:a28bd4daf3b3432e8414859dba25bda2
record_format	dspace
spelling	oai:doaj.org-article:a28bd4daf3b3432e8414859dba25bda22021-11-21T12:26:53ZFully automated identification of brain abnormality from whole-body FDG-PET imaging using deep learning-based brain extraction and statistical parametric mapping10.1186/s40658-021-00424-02197-7364https://doaj.org/article/a28bd4daf3b3432e8414859dba25bda22021-11-01T00:00:00Zhttps://doi.org/10.1186/s40658-021-00424-0https://doaj.org/toc/2197-7364Abstract Background The whole brain is often covered in [18F]Fluorodeoxyglucose positron emission tomography ([18F]FDG-PET) in oncology patients, but the covered brain abnormality is typically screened by visual interpretation without quantitative analysis in clinical practice. In this study, we aimed to develop a fully automated quantitative interpretation pipeline of brain volume from an oncology PET image. Method We retrospectively collected 500 oncologic [18F]FDG-PET scans for training and validation of the automated brain extractor. We trained the model for extracting brain volume with two manually drawn bounding boxes on maximal intensity projection images. ResNet-50, a 2-D convolutional neural network (CNN), was used for the model training. The brain volume was automatically extracted using the CNN model and spatially normalized. For validation of the trained model and an application of this automated analytic method, we enrolled 24 subjects with small cell lung cancer (SCLC) and performed voxel-wise two-sample T test for automatic detection of metastatic lesions. Result The deep learning-based brain extractor successfully identified the existence of whole-brain volume, with an accuracy of 98% for the validation set. The performance of extracting the brain measured by the intersection-over-union of 3-D bounding boxes was 72.9 ± 12.5% for the validation set. As an example of the application to automatically identify brain abnormality, this approach successfully identified the metastatic lesions in three of the four cases of SCLC patients with brain metastasis. Conclusion Based on the deep learning-based model, extraction of the brain volume from whole-body PET was successfully performed. We suggest this fully automated approach could be used for the quantitative analysis of brain metabolic patterns to identify abnormalities during clinical interpretation of oncologic PET studies.Wonseok WhiHongyoon ChoiJin Chul PaengGi Jeong CheonKeon Wook KangDong Soo LeeSpringerOpenarticleBrain segmentationQuantitative PET analysisDeep learningConvolutional neural networkFDG-PETBrain FDG-PETMedical physics. Medical radiology. Nuclear medicineR895-920ENEJNMMI Physics, Vol 8, Iss 1, Pp 1-10 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Brain segmentation Quantitative PET analysis Deep learning Convolutional neural network FDG-PET Brain FDG-PET Medical physics. Medical radiology. Nuclear medicine R895-920
spellingShingle	Brain segmentation Quantitative PET analysis Deep learning Convolutional neural network FDG-PET Brain FDG-PET Medical physics. Medical radiology. Nuclear medicine R895-920 Wonseok Whi Hongyoon Choi Jin Chul Paeng Gi Jeong Cheon Keon Wook Kang Dong Soo Lee Fully automated identification of brain abnormality from whole-body FDG-PET imaging using deep learning-based brain extraction and statistical parametric mapping
description	Abstract Background The whole brain is often covered in [18F]Fluorodeoxyglucose positron emission tomography ([18F]FDG-PET) in oncology patients, but the covered brain abnormality is typically screened by visual interpretation without quantitative analysis in clinical practice. In this study, we aimed to develop a fully automated quantitative interpretation pipeline of brain volume from an oncology PET image. Method We retrospectively collected 500 oncologic [18F]FDG-PET scans for training and validation of the automated brain extractor. We trained the model for extracting brain volume with two manually drawn bounding boxes on maximal intensity projection images. ResNet-50, a 2-D convolutional neural network (CNN), was used for the model training. The brain volume was automatically extracted using the CNN model and spatially normalized. For validation of the trained model and an application of this automated analytic method, we enrolled 24 subjects with small cell lung cancer (SCLC) and performed voxel-wise two-sample T test for automatic detection of metastatic lesions. Result The deep learning-based brain extractor successfully identified the existence of whole-brain volume, with an accuracy of 98% for the validation set. The performance of extracting the brain measured by the intersection-over-union of 3-D bounding boxes was 72.9 ± 12.5% for the validation set. As an example of the application to automatically identify brain abnormality, this approach successfully identified the metastatic lesions in three of the four cases of SCLC patients with brain metastasis. Conclusion Based on the deep learning-based model, extraction of the brain volume from whole-body PET was successfully performed. We suggest this fully automated approach could be used for the quantitative analysis of brain metabolic patterns to identify abnormalities during clinical interpretation of oncologic PET studies.
format	article
author	Wonseok Whi Hongyoon Choi Jin Chul Paeng Gi Jeong Cheon Keon Wook Kang Dong Soo Lee
author_facet	Wonseok Whi Hongyoon Choi Jin Chul Paeng Gi Jeong Cheon Keon Wook Kang Dong Soo Lee
author_sort	Wonseok Whi
title	Fully automated identification of brain abnormality from whole-body FDG-PET imaging using deep learning-based brain extraction and statistical parametric mapping
title_short	Fully automated identification of brain abnormality from whole-body FDG-PET imaging using deep learning-based brain extraction and statistical parametric mapping
title_full	Fully automated identification of brain abnormality from whole-body FDG-PET imaging using deep learning-based brain extraction and statistical parametric mapping
title_fullStr	Fully automated identification of brain abnormality from whole-body FDG-PET imaging using deep learning-based brain extraction and statistical parametric mapping
title_full_unstemmed	Fully automated identification of brain abnormality from whole-body FDG-PET imaging using deep learning-based brain extraction and statistical parametric mapping
title_sort	fully automated identification of brain abnormality from whole-body fdg-pet imaging using deep learning-based brain extraction and statistical parametric mapping
publisher	SpringerOpen
publishDate	2021
url	https://doaj.org/article/a28bd4daf3b3432e8414859dba25bda2
work_keys_str_mv	AT wonseokwhi fullyautomatedidentificationofbrainabnormalityfromwholebodyfdgpetimagingusingdeeplearningbasedbrainextractionandstatisticalparametricmapping AT hongyoonchoi fullyautomatedidentificationofbrainabnormalityfromwholebodyfdgpetimagingusingdeeplearningbasedbrainextractionandstatisticalparametricmapping AT jinchulpaeng fullyautomatedidentificationofbrainabnormalityfromwholebodyfdgpetimagingusingdeeplearningbasedbrainextractionandstatisticalparametricmapping AT gijeongcheon fullyautomatedidentificationofbrainabnormalityfromwholebodyfdgpetimagingusingdeeplearningbasedbrainextractionandstatisticalparametricmapping AT keonwookkang fullyautomatedidentificationofbrainabnormalityfromwholebodyfdgpetimagingusingdeeplearningbasedbrainextractionandstatisticalparametricmapping AT dongsoolee fullyautomatedidentificationofbrainabnormalityfromwholebodyfdgpetimagingusingdeeplearningbasedbrainextractionandstatisticalparametricmapping
_version_	1718419021488979968

Fully automated identification of brain abnormality from whole-body FDG-PET imaging using deep learning-based brain extraction and statistical parametric mapping

Ejemplares similares