Multi-task glioma segmentation and IDH mutation and 1p19q codeletion classification via a deep learning model on multimodal MRI

Objectives: To develop a deep learning model for simultaneous segmentation of glioma lesions and classification of IDH mutation and 1p/19q codeletion status using multimodal MRI. Materials and Methods: We employed a CNN model with Encoder-Decoder architecture for segmentation, followed by fully conn...

Full description

Saved in:
Bibliographic Details
Main Authors: Erin Beate Bjørkeli, Morteza Esmaeili
Format: Article
Language:English
Published: KeAi Communications Co., Ltd. 2025-06-01
Series:Meta-Radiology
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2950162825000207
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849717514302390272
author Erin Beate Bjørkeli
Morteza Esmaeili
author_facet Erin Beate Bjørkeli
Morteza Esmaeili
author_sort Erin Beate Bjørkeli
collection DOAJ
description Objectives: To develop a deep learning model for simultaneous segmentation of glioma lesions and classification of IDH mutation and 1p/19q codeletion status using multimodal MRI. Materials and Methods: We employed a CNN model with Encoder-Decoder architecture for segmentation, followed by fully connected layers for classification. The model was trained and validated using the BraTS 2020 dataset (132 examinations with known molecular status, split 80/20). Four MRI sequences iamges (T1, T1ce, T2, FLAIR) were used for analysis. Segmentation performance was evaluated using mean Dice Score (mDS) and mean Intersection over Union (mIoU). Classification was assessed using accuracy, sensitivity, and specificity. Results: The model achieved the best segmentation performance with all four modalities (mDS validation ​= ​0.73, mIoU validation ​= ​0.62). Among single modalities, FLAIR performed best (mDS validation ​= ​0.56, mIoU validation ​= ​0.44). For classification, the combined four modalities achieved an overall accuracy of 0.98. However, classification precision for IDH and 1p19q was potentially limited by class imbalance. Conclusion: Our CNN-based Encoder-Decoder model demonstrates the benefit of multimodal MRI for accurate glioma segmentation and shows promising results for molecular subtype classification. Future work will focus on addressing class imbalance and exploring feature integration to enhance classification performance.
format Article
id doaj-art-b2c22dc7793548b08a7a7c9eee1583f3
institution DOAJ
issn 2950-1628
language English
publishDate 2025-06-01
publisher KeAi Communications Co., Ltd.
record_format Article
series Meta-Radiology
spelling doaj-art-b2c22dc7793548b08a7a7c9eee1583f32025-08-20T03:12:38ZengKeAi Communications Co., Ltd.Meta-Radiology2950-16282025-06-013210015210.1016/j.metrad.2025.100152Multi-task glioma segmentation and IDH mutation and 1p19q codeletion classification via a deep learning model on multimodal MRIErin Beate Bjørkeli0Morteza Esmaeili1Department of Diagnostic Imaging, Akershus University Hospital, Lørenskog, Norway; Institute of Clinical Medicine, University of Oslo, Oslo, NorwayDepartment of Diagnostic Imaging, Akershus University Hospital, Lørenskog, Norway; Department of Electrical Engineering and Computer Science, University of Stavanger, Stavanger, Norway; Corresponding author.Objectives: To develop a deep learning model for simultaneous segmentation of glioma lesions and classification of IDH mutation and 1p/19q codeletion status using multimodal MRI. Materials and Methods: We employed a CNN model with Encoder-Decoder architecture for segmentation, followed by fully connected layers for classification. The model was trained and validated using the BraTS 2020 dataset (132 examinations with known molecular status, split 80/20). Four MRI sequences iamges (T1, T1ce, T2, FLAIR) were used for analysis. Segmentation performance was evaluated using mean Dice Score (mDS) and mean Intersection over Union (mIoU). Classification was assessed using accuracy, sensitivity, and specificity. Results: The model achieved the best segmentation performance with all four modalities (mDS validation ​= ​0.73, mIoU validation ​= ​0.62). Among single modalities, FLAIR performed best (mDS validation ​= ​0.56, mIoU validation ​= ​0.44). For classification, the combined four modalities achieved an overall accuracy of 0.98. However, classification precision for IDH and 1p19q was potentially limited by class imbalance. Conclusion: Our CNN-based Encoder-Decoder model demonstrates the benefit of multimodal MRI for accurate glioma segmentation and shows promising results for molecular subtype classification. Future work will focus on addressing class imbalance and exploring feature integration to enhance classification performance.http://www.sciencedirect.com/science/article/pii/S2950162825000207GlioblastomaMolecular subtypeBraTS datasetTCGAIsocitrate dehydrogenase
spellingShingle Erin Beate Bjørkeli
Morteza Esmaeili
Multi-task glioma segmentation and IDH mutation and 1p19q codeletion classification via a deep learning model on multimodal MRI
Meta-Radiology
Glioblastoma
Molecular subtype
BraTS dataset
TCGA
Isocitrate dehydrogenase
title Multi-task glioma segmentation and IDH mutation and 1p19q codeletion classification via a deep learning model on multimodal MRI
title_full Multi-task glioma segmentation and IDH mutation and 1p19q codeletion classification via a deep learning model on multimodal MRI
title_fullStr Multi-task glioma segmentation and IDH mutation and 1p19q codeletion classification via a deep learning model on multimodal MRI
title_full_unstemmed Multi-task glioma segmentation and IDH mutation and 1p19q codeletion classification via a deep learning model on multimodal MRI
title_short Multi-task glioma segmentation and IDH mutation and 1p19q codeletion classification via a deep learning model on multimodal MRI
title_sort multi task glioma segmentation and idh mutation and 1p19q codeletion classification via a deep learning model on multimodal mri
topic Glioblastoma
Molecular subtype
BraTS dataset
TCGA
Isocitrate dehydrogenase
url http://www.sciencedirect.com/science/article/pii/S2950162825000207
work_keys_str_mv AT erinbeatebjørkeli multitaskgliomasegmentationandidhmutationand1p19qcodeletionclassificationviaadeeplearningmodelonmultimodalmri
AT mortezaesmaeili multitaskgliomasegmentationandidhmutationand1p19qcodeletionclassificationviaadeeplearningmodelonmultimodalmri