Global and Regional Deep Learning Models for Multiple Sclerosis Stratification From MRI.
classification
deep learning
input sampling
multiple sclerosis
structural MRI
Journal
Journal of magnetic resonance imaging : JMRI
ISSN: 1522-2586
Titre abrégé: J Magn Reson Imaging
Pays: United States
ID NLM: 9105850
Informations de publication
Date de publication:
06 Oct 2023
06 Oct 2023
Historique:
revised:
15
09
2023
received:
27
04
2023
accepted:
18
09
2023
medline:
7
10
2023
pubmed:
7
10
2023
entrez:
7
10
2023
Statut:
aheadofprint
Résumé
The combination of anatomical MRI and deep learning-based methods such as convolutional neural networks (CNNs) is a promising strategy to build predictive models of multiple sclerosis (MS) prognosis. However, studies assessing the effect of different input strategies on model's performance are lacking. To compare whole-brain input sampling strategies and regional/specific-tissue strategies, which focus on a priori known relevant areas for disability accrual, to stratify MS patients based on their disability level. Retrospective. Three hundred nineteen MS patients (382 brain MRI scans) with clinical assessment of disability level performed within the following 6 months (~70% training/~15% validation/~15% inference in-house dataset) and 440 MS patients from multiple centers (independent external validation cohort). Single vendor 1.5 T or 3.0 T. Magnetization-Prepared Rapid Gradient-Echo and Fluid-Attenuated Inversion Recovery sequences. A 7-fold patient cross validation strategy was used to train a 3D-CNN to classify patients into two groups, Expanded Disability Status Scale score (EDSS) ≥ 3.0 or EDSS < 3.0. Two strategies were investigated: 1) a global approach, taking the whole brain volume as input and 2) regional approaches using five different regions-of-interest: white matter, gray matter, subcortical gray matter, ventricles, and brainstem structures. The performance of the models was assessed in the in-house and the independent external cohorts. Balanced accuracy, sensitivity, specificity, area under receiver operating characteristic (ROC) curve (AUC). With the in-house dataset, the gray matter regional model showed the highest stratification accuracy (81%), followed by the global approach (79%). In the external dataset, without any further retraining, an accuracy of 72% was achieved for the white matter model and 71% for the global approach. The global approach offered the best trade-off between internal performance and external validation to stratify MS patients based on accumulated disability. 4 TECHNICAL EFFICACY: Stage 2.
Sections du résumé
BACKGROUND
BACKGROUND
The combination of anatomical MRI and deep learning-based methods such as convolutional neural networks (CNNs) is a promising strategy to build predictive models of multiple sclerosis (MS) prognosis. However, studies assessing the effect of different input strategies on model's performance are lacking.
PURPOSE
OBJECTIVE
To compare whole-brain input sampling strategies and regional/specific-tissue strategies, which focus on a priori known relevant areas for disability accrual, to stratify MS patients based on their disability level.
STUDY TYPE
METHODS
Retrospective.
SUBJECTS
METHODS
Three hundred nineteen MS patients (382 brain MRI scans) with clinical assessment of disability level performed within the following 6 months (~70% training/~15% validation/~15% inference in-house dataset) and 440 MS patients from multiple centers (independent external validation cohort).
FIELD STRENGTH/SEQUENCE
UNASSIGNED
Single vendor 1.5 T or 3.0 T. Magnetization-Prepared Rapid Gradient-Echo and Fluid-Attenuated Inversion Recovery sequences.
ASSESSMENT
RESULTS
A 7-fold patient cross validation strategy was used to train a 3D-CNN to classify patients into two groups, Expanded Disability Status Scale score (EDSS) ≥ 3.0 or EDSS < 3.0. Two strategies were investigated: 1) a global approach, taking the whole brain volume as input and 2) regional approaches using five different regions-of-interest: white matter, gray matter, subcortical gray matter, ventricles, and brainstem structures. The performance of the models was assessed in the in-house and the independent external cohorts.
STATISTICAL TESTS
METHODS
Balanced accuracy, sensitivity, specificity, area under receiver operating characteristic (ROC) curve (AUC).
RESULTS
RESULTS
With the in-house dataset, the gray matter regional model showed the highest stratification accuracy (81%), followed by the global approach (79%). In the external dataset, without any further retraining, an accuracy of 72% was achieved for the white matter model and 71% for the global approach.
DATA CONCLUSION
CONCLUSIONS
The global approach offered the best trade-off between internal performance and external validation to stratify MS patients based on accumulated disability.
EVIDENCE LEVEL
METHODS
4 TECHNICAL EFFICACY: Stage 2.
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Subventions
Organisme : 'la Caixa' Foundation
ID : 100010434
Organisme : 'la Caixa' Foundation
ID : LCF/BQ/PI20/11760008
Informations de copyright
© 2023 The Authors. Journal of Magnetic Resonance Imaging published by Wiley Periodicals LLC on behalf of International Society for Magnetic Resonance in Medicine.
Références
Thompson AJ, Baranzini SE, Geurts J, Hemmer B, Ciccarelli O. Multiple sclerosis. Lancet 2018;391(10130):1622-1636.
Tintore M, Rovira L, Río J, et al. Defining high, medium and low impact prognostic factors for developing multiple sclerosis. Brain 2015;138(Pt 7):1863-1874.
Haider L, Chung K, Birch G, et al. Linear brain atrophy measures in multiple sclerosis and clinically isolated syndromes: A 30-year follow-up. J Neurol Neurosurg Psychiatry 2021;92(8):839-846.
Popescu V, Agosta F, Hulst HE, et al. Brain atrophy and lesion load predict long term disability in multiple sclerosis. J Neurol Neurosurg Psychiatry 2013;84(10):1082-1091.
Cappelle S, Pareto D, Vidal-Jordana A, et al. A validation study of manual atrophy measures in patients with Multiple Sclerosis. Neuroradiology 2020;62(8):955-964.
Bonacchi R, Meani A, Pagani E, Marchesi O, Filippi M, Rocca MA. The role of cerebellar damage in explaining disability and cognition in multiple sclerosis phenotypes: A multiparametric MRI study. J Neurol 2022;269(7):3841-3857.
Lecun Y, Bengio Y, Hinton G. Deep learning. Nature 2015;521(7553):436-444.
Bernal J, Kushibar K, Asfaw DS, et al. Deep convolutional neural networks for brain image analysis on magnetic resonance imaging: A review. Artif Intell Med 2019;95:64-81.
Commowick O, Cervenansky F, Cotton F, Dojat M. MSSEG-2 challenge proceedings: Multiple sclerosis new lesions segmentation challenge using a data management and processing infrastructure. MICCAI 2021 - 24th Int Conf Med Image Comput Comput Assist Interv 2021;126-2021.
Shoeibi A, Khodatars M, Jafari M, et al. Applications of deep learning techniques for automated multiple sclerosis detection using magnetic resonance imaging: A review. Comput Biol Med 2021;136:104697.
Tousignant A, Lemaître P, Doina P, Arnold DL, Arbel T. Prediction of disease progression in multiple sclerosis patients using deep learning analysis of MRI data. Proc Mach Learn Res 2019;102:483-492.
Roca P, Attye A, Colas L, et al. Artificial intelligence to predict clinical disability in patients with multiple sclerosis using FLAIR MRI. Diagn Interv Imaging 2020;101(12):795-802.
Ahmed S, Kim BC, Lee KH, Yub H, Jung HY. Ensemble of ROI-based convolutional neural network classifiers for staging the Alzheimer disease spectrum from magnetic resonance imaging. PLoS One 2020;15(12):e0242712.
Kwak K, Niethammer M, Giovanello KS, Styner M, Dayan E. Differential role for hippocampal subfields in Alzheimer's disease progression revealed with deep learning. Cereb Cortex 2021;32(3):467-478.
Basheera S, Ram MSS. Deep learning based Alzheimer's disease early diagnosis using T2w segmented gray matter MRI. Int J Imaging Syst Technol 2021;31(3):1692-1710.
Mehmood A, Yang S, Feng Z, et al. A transfer learning approach for early diagnosis of Alzheimer's disease on MRI images. Neuroscience 2021;460:43-52.
Cao P, Gao J, Zhang Z. Multi-view based multi-model learning for MCI diagnosis. Brain Sci 2020;10(3):181.
Zhou P, Jiang S, Yu L, et al. Use of a sparse-response deep belief network and extreme learning machine to discriminate Alzheimer's disease, mild cognitive impairment, and normal controls based on amyloid PET/MRI images. Front Med 2021;7:621204.
Zhu W, Sun L, Huang J, Han L, Zhang D. Dual attention multi-instance deep learning for Alzheimer's disease diagnosis with structural MRI. IEEE Trans Med Imaging 2021;40(9):2354-2366.
Kurtzke JF. Rating neurologic impairment in multiple sclerosis. Neurology 1983;33(11):1444-1452.
Rizzo MA, Hadjimichael OC, Preiningerova J, Vollmer TL. Prevalence and treatment of spasticity reported by multiple sclerosis patients. Mult Scler J 2004;10(5):589-595.
Tintoré M, Rovira A, Río J, et al. Baseline MRI predicts future attacks and disability in clinically isolated syndromes. Neurology 2006;67(6):968-972.
Learmonth YC, Motl RW, Sandroff BM, Pula JH, Cadavid D. Validation of patient determined disease steps (PDDS) scale scores in persons with multiple sclerosis. BMC Neurol 2013;13:37.
Mowry EM, Bermel RA, Williams JR, et al. Harnessing real-world data to inform decision-making: Multiple sclerosis partners advancing technology and health solutions (MS PATHS). Front Neurol 2020;11:632.
Tustison NJ, Avants BB, Cook PA, et al. N4ITK: Improved N3 bias correction. IEEE Trans Med Imaging 2010;29(6):1310-1320.
Isensee F, Schell M, Pflueger I, et al. Automated brain extraction of multisequence MRI using artificial neural networks. Hum Brain Mapp 2019;40(17):4952-4964.
Jenkinson M, Beckmann CF, Behrens TEJ, Woolrich MW, Smith SM. FSL. Neuroimage 2012;62(2):782-790.
Schmidt P, Gaser C, Arsic M, et al. An automated tool for detection of FLAIR-hyperintense white-matter lesions in Multiple Sclerosis. Neuroimage 2012;59(4):3774-3783.
Prados F, Cardoso MJ, Kanber B, et al. A multi-time-point modality-agnostic patch-based method for lesion filling in multiple sclerosis. Neuroimage 2016;139:376-384.
Henschel L, Conjeti S, Estrada S, Diers K, Fischl B, Reuter M. FastSurfer - a fast and accurate deep learning based neuroimaging pipeline. Neuroimage 2020;219:117012.
Lungu O, Pantano P, Kumfor F, et al. Impaired self-other distinction and subcortical gray-matter alterations characterize socio-cognitive disturbances in multiple sclerosis. Front Neurol 2019;10:525.
He K, Zhang X, Ren S, Sun J. Deep residual learning for image recognition. Proc IEEE Comput Soc Conf Comput Vis Pattern Recognit 2016;770-778.
Huang SC, Pareek A, Seyyedi S, Banerjee I, Lungren MP. Fusion of medical imaging and electronic health records using deep learning: A systematic review and implementation guidelines. NPJ Digit Med 2020;3:136.
Sun X, Xu W. Fast implementation of DeLong's algorithm for comparing the areas under correlated receiver operating characteristic curves. IEEE Signal Process Lett 2014;21(11):1389-1393.
Paszke A, Gross S, Massa F, et al. PyTorch: An imperative style, high-performance deep learning library. Proceedings of the 33rd international conference on neural information processing systems. Red Hook, NY, USA: Curran Associates Inc.; 2019. p 8026-8037.
Amiri H, de Sitter A, Bendfeldt K, et al. Urgent challenges in quantification and interpretation of brain grey matter atrophy in individual MS patients using MRI. NeuroImage Clin 2018;19:466-475.
Eshaghi A, Marinescu RV, Young AL, et al. Progression of regional grey matter atrophy in multiple sclerosis on behalf of the MAGNIMS study group*. Brain 2018;141:1665-1677.
Eshaghi A, Prados F, Brownlee WJ, et al. Deep gray matter volume loss drives disability worsening in multiple sclerosis. Ann Neurol 2018;83(2):210-222.
Brown JWL, Pardini M, Brownlee WJ, et al. An abnormal periventricular magnetization transfer ratio gradient occurs early in multiple sclerosis. Brain 2017;140(2):387-398.