P Roca1, A Attye2, L Colas3, A Tucholka4, P Rubini4, S Cackowski5, J Ding3, J-F Budzik3, F Renard6, S Doyle4, E L Barbier5, I Bousaid7, R Casey8, S Vukusic8, N Lassau9, S Verclytte3, F Cotton10. 1. Pixyl, Research and Development Laboratory, 38000 Grenoble, France. Electronic address: contact@pixyl.ai. 2. Grenoble Alpes University, 38000 Grenoble, France; Sydney Imaging Lab, Sydney University, 2006 Sydney, NSW, Australia. 3. Imaging Department, Lille Catholic Hospitals, Lille Catholic University, 59000 Lille, France. 4. Pixyl, Research and Development Laboratory, 38000 Grenoble, France. 5. University Grenoble Alpes, Inserm, U1216, Grenoble Institute Neurosciences, 38000 Grenoble, France. 6. University Grenoble Alpes, CNRS, Grenoble INP, LIG, 38000 Grenoble, France; University Grenoble Alpes, AGEIS, 38000 Grenoble, France. 7. Direction Transformation Numérique et Systèmes d'Information, Institut Gustave Roussy, 94805 Villejuif, France. 8. Department of Neurology-Multiple Sclerosis, Pathologies de la myéline et neuro-inflammation, Hôpital Pierre Wertheimer, Hospices Civils de Lyon, 69500 Bron, France; Université Claude Bernard Lyon 1, Université de Lyon, 69622 Villeurbanne, France; Observatoire Français de la Sclérose en Plaques, Centre de Recherche en Neurosciences de Lyon, INSERM 1028 et CNRS UMR 5292, 69003 Lyon, France; Eugène Devic EDMUS Foundation Against Multiple Sclerosis, 69500 Bron, France. 9. Radiology Department, Institut Gustave Roussy, 94805 Villejuif, France; BIOMAPS, UMR1281, Université Paris-Saclay, Inserm, CNRS, CEA, Laboratoire d'Imagerie Biomédicale Multimodale Paris-Saclay, 94800 Villejuif, France. 10. Observatoire Français de la Sclérose en Plaques, Centre de Recherche en Neurosciences de Lyon, INSERM 1028 et CNRS UMR 5292, 69003 Lyon, France; Department of Radiology, Centre Hospitalier Lyon-Sud, Hospices Civils de Lyon, 69310 Pierre-Bénite, France; CREATIS, CNRS UMR 5220, INSERM U1044, 69622 Villeurbanne, France.
Abstract
PURPOSE: The purpose of this study was to create an algorithm that combines multiple machine-learning techniques to predict the expanded disability status scale (EDSS) score of patients with multiple sclerosis at two years solely based on age, sex and fluid attenuated inversion recovery (FLAIR) MRI data. MATERIALS AND METHODS: Our algorithm combined several complementary predictors: a pure deep learning predictor based on a convolutional neural network (CNN) that learns from the images, as well as classical machine-learning predictors based on random forest regressors and manifold learning trained using the location of lesion load with respect to white matter tracts. The aggregation of the predictors was done through a weighted average taking into account prediction errors for different EDSS ranges. The training dataset consisted of 971 multiple sclerosis patients from the "Observatoire français de la sclérose en plaques" (OFSEP) cohort with initial FLAIR MRI and corresponding EDSS score at two years. A test dataset (475 subjects) was provided without an EDSS score. Ten percent of the training dataset was used for validation. RESULTS: Our algorithm predicted EDSS score in patients with multiple sclerosis and achieved a MSE=2.2 with the validation dataset and a MSE=3 (mean EDSS error=1.7) with the test dataset. CONCLUSION: Our method predicts two-year clinical disability in patients with multiple sclerosis with a mean EDSS score error of 1.7, using FLAIR sequence and basic patient demographics. This supports the use of our model to predict EDSS score progression. These promising results should be further validated on an external validation cohort.
PURPOSE: The purpose of this study was to create an algorithm that combines multiple machine-learning techniques to predict the expanded disability status scale (EDSS) score of patients with multiple sclerosis at two years solely based on age, sex and fluid attenuated inversion recovery (FLAIR) MRI data. MATERIALS AND METHODS: Our algorithm combined several complementary predictors: a pure deep learning predictor based on a convolutional neural network (CNN) that learns from the images, as well as classical machine-learning predictors based on random forest regressors and manifold learning trained using the location of lesion load with respect to white matter tracts. The aggregation of the predictors was done through a weighted average taking into account prediction errors for different EDSS ranges. The training dataset consisted of 971 multiple sclerosispatients from the "Observatoire français de la sclérose en plaques" (OFSEP) cohort with initial FLAIR MRI and corresponding EDSS score at two years. A test dataset (475 subjects) was provided without an EDSS score. Ten percent of the training dataset was used for validation. RESULTS: Our algorithm predicted EDSS score in patients with multiple sclerosis and achieved a MSE=2.2 with the validation dataset and a MSE=3 (mean EDSS error=1.7) with the test dataset. CONCLUSION: Our method predicts two-year clinical disability in patients with multiple sclerosis with a mean EDSS score error of 1.7, using FLAIR sequence and basic patient demographics. This supports the use of our model to predict EDSS score progression. These promising results should be further validated on an external validation cohort.
Authors: Pedro Alves; Eric Green; Michelle Leavy; Haley Friedler; Gary Curhan; Carl Marci; Costas Boussios Journal: Mult Scler J Exp Transl Clin Date: 2022-06-22