A Deep Learning Approach to Predicting Disease Progression in Multiple Sclerosis Using Magnetic Resonance Imaging

Storelli, Loredana; Azzimonti, Matteo; Gueye, Mor; Vizzino, Carmen; Preziosa, Paolo; Tedeschi, G; Stefano, Nicola De; Pantano, P.; Filippi, Massimo; Rocca, Maria

doi:10.1097/rli.0000000000000854

Cited by 27 publications

(25 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This could be related to the intrinsic relationship between WM and GM damage burden and their reciprocal influence; as also shown by our cluster analysis and RF results, this underlines the complex and multifactorial relationship between CI and both WM and GM pathology. These results are in line with previous studies regarding predictive models considering MRI data with respect to CI [47,48].…”

Section: Discussionsupporting

confidence: 93%

Cortical lesions at diagnosis predict long‐term cognitive impairment in multiple sclerosis: A 20‐year study

Ziccardi

Pisani

Schiavi

et al. 2023

Euro J of Neurology

View full text Add to dashboard Cite

Background and purpose Although cognitive impairment (CI) is frequent in multiple sclerosis (MS) patients, few studies (and with conflicting results) have evaluated early predictors of CI in the long term. We aimed at determining associations between early clinical/neuroradiological variables with reference to CI after 20 years of MS. Methods We investigated in 170 MS patients the relationship between clinical/magnetic resonance imaging (MRI) data at diagnosis and cognitive status almost 20 years after MS onset. Among others, number and volume of both white matter lesions (WMLs) and cortical lesions (CLs) were evaluated at diagnosis and after 2 years. All MS patients were followed over time and underwent a comprehensive neuropsychological assessment at the end of study. Advanced statistical methods (unsupervised cluster analysis and random forest model) were conducted. Results CI patients showed higher focal cortical pathology at diagnosis compared to cognitively normal subjects (p < 0.001). Volumes of both WMLs and CLs emerged as the MRI metrics most associated with long‐term CI. Moreover, number of CLs (especially ≥3) was also strongly associated with long‐term CI (≥3 CLs: odds ratio [OR] = 3.7, 95% confidence interval = 1.8–7.5, p < 0.001), more than number of WMLs; the optimal cutoff of three CLs (area under the curve = 0.67, specificity = 75%, sensitivity = 55%) was estimated according to the risk of developing CI. Conclusions These results highlight the impact of considering both white and gray matter focal damage from early MS stages. Given the low predictive value of WML number and the poor clinical applicability of lesion volume estimation in the daily clinical context, the evaluation of number of CLs could represent a reliable prognostic marker of CI.

show abstract

Section: Discussionsupporting

confidence: 93%

Cortical lesions at diagnosis predict long‐term cognitive impairment in multiple sclerosis: A 20‐year study

Ziccardi

Pisani

Schiavi

et al. 2023

Euro J of Neurology

View full text Add to dashboard Cite

show abstract

“…On the basis of the promising results in image segmentation and classification, we tested the DL ability to predict disability progression in MS. Indeed, to predict clinical status and worsening in patients with MS, artificial intelligence algorithms were applied either on MRI [15][16][17][18] and/or clinical data [19,20]. In this pilot study, we hypothesized that the disability progression, as evaluated via the Expanded Disability Status Scale (EDSS) after 4 years from the baseline visit, may be predicted from structural MRI images acquired at baseline.…”

Section: Introductionmentioning

confidence: 99%

Evaluation of Disability Progression in Multiple Sclerosis via Magnetic-Resonance-Based Deep Learning Techniques

Taloni

Farrelly

Pontillo

et al. 2022

IJMS

Self Cite

View full text Add to dashboard Cite

Short-term disability progression was predicted from a baseline evaluation in patients with multiple sclerosis (MS) using their three-dimensional T1-weighted (3DT1) magnetic resonance images (MRI). One-hundred-and-eighty-one subjects diagnosed with MS underwent 3T-MRI and were followed up for two to six years at two sites, with disability progression defined according to the expanded-disability-status-scale (EDSS) increment at the follow-up. The patients’ 3DT1 images were bias-corrected, brain-extracted, registered onto MNI space, and divided into slices along coronal, sagittal, and axial projections. Deep learning image classification models were applied on slices and devised as ResNet50 fine-tuned adaptations at first on a large independent dataset and secondly on the study sample. The final classifiers’ performance was evaluated via the area under the curve (AUC) of the false versus true positive diagram. Each model was also tested against its null model, obtained by reshuffling patients’ labels in the training set. Informative areas were found by intersecting slices corresponding to models fulfilling the disability progression prediction criteria. At follow-up, 34% of patients had disability progression. Five coronal and five sagittal slices had one classifier surviving the AUC evaluation and null test and predicted disability progression (AUC > 0.72 and AUC > 0.81, respectively). Likewise, fifteen combinations of classifiers and axial slices predicted disability progression in patients (AUC > 0.69). Informative areas were the frontal areas, mainly within the grey matter. Briefly, 3DT1 images may give hints on disability progression in MS patients, exploiting the information hidden in the MRI of specific areas of the brain.

show abstract

“…Some MS detection studies presented in the literature are given below. Storelli et al [21] analyzed the MRIs of 373 MS patients using a CNN model and attained accuracy rates of 83.3%, 67.7%, and 85.7% for clinical, cognitive, and combined clinical plus cognitive diagnoses, respectively. The parameter values and optimization methods used in the CNN architecture negatively affected the classification results.…”

Section: Introductionmentioning

confidence: 99%

“…The datasets in the studies presented above [22,[28][29][30] are small. In Storelli et al [21] and Narayana et al [24], the dataset is large, but has a low accuracy rate. In [21][22][23][24][25][26][27][28]30], computational complexity is high.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

An Accurate Multiple Sclerosis Detection Model Based on Exemplar Multiple Parameters Local Phase Quantization: ExMPLPQ

Macin

Taşçı

et al. 2022

Applied Sciences

View full text Add to dashboard Cite

Multiple sclerosis (MS) is a chronic demyelinating condition characterized by plaques in the white matter of the central nervous system that can be detected using magnetic resonance imaging (MRI). Many deep learning models for automated MS detection based on MRI have been presented in the literature. We developed a computationally lightweight machine learning model for MS diagnosis using a novel handcrafted feature engineering approach. The study dataset comprised axial and sagittal brain MRI images that were prospectively acquired from 72 MS and 59 healthy subjects who attended the Ozal University Medical Faculty in 2021. The dataset was divided into three study subsets: axial images only (n = 1652), sagittal images only (n = 1775), and combined axial and sagittal images (n = 3427) of both MS and healthy classes. All images were resized to 224 × 224. Subsequently, the features were generated with a fixed-size patch-based (exemplar) feature extraction model based on local phase quantization (LPQ) with three-parameter settings. The resulting exemplar multiple parameters LPQ (ExMPLPQ) features were concatenated to form a large final feature vector. The top discriminative features were selected using iterative neighborhood component analysis (INCA). Finally, a k-nearest neighbor (kNN) algorithm, Fine kNN, was deployed to perform binary classification of the brain images into MS vs. healthy classes. The ExMPLPQ-based model attained 98.37%, 97.75%, and 98.22% binary classification accuracy rates for axial, sagittal, and hybrid datasets, respectively, using Fine kNN with 10-fold cross-validation. Furthermore, our model outperformed 19 established pre-trained deep learning models that were trained and tested with the same data. Unlike deep models, the ExMPLPQ-based model is computationally lightweight yet highly accurate. It has the potential to be implemented as an automated diagnostic tool to screen brain MRIs for white matter lesions in suspected MS patients.

show abstract

A Deep Learning Approach to Predicting Disease Progression in Multiple Sclerosis Using Magnetic Resonance Imaging

Cited by 27 publications

References 53 publications

Cortical lesions at diagnosis predict long‐term cognitive impairment in multiple sclerosis: A 20‐year study

Cortical lesions at diagnosis predict long‐term cognitive impairment in multiple sclerosis: A 20‐year study

Evaluation of Disability Progression in Multiple Sclerosis via Magnetic-Resonance-Based Deep Learning Techniques

An Accurate Multiple Sclerosis Detection Model Based on Exemplar Multiple Parameters Local Phase Quantization: ExMPLPQ

Contact Info

Product

Resources

About