A Smart Platform for Management and Diagnosis of Multiple Sclerosis

Abstract

This chapter has described the concrete implementation of the proposed intelligent platform for the detection of Multiple Sclerosis and the management of patients. Starting from a datasets ,followed by the design and training of a compact convolutional neural network of approximately 1.2 million parameters, trained from scratch to perform a binary classification of the images and to localize the lesion regions through Grad-CAM. The model achieved very high performance on the available test set, and these results were discussed clearly, identifying the diversity of sources nature of the data and the slice-level partitions well as the lack of datasets concerning MS all of that will be considered as factors that must be considered when interpreting the metrics. The chapter also presented the complete MyelineCare platform through its main interfaces, demonstrating how the artificial intelligence engine is integrated into a usable maintainable clinical tool covering authentication, patient management, MRI analysis, explainable results, report generation, and longitudinal follow-up. Together, these elements demonstrate the technical feasibility of the proposed solution and its capacity to assist neurologists in the diagnosis and monitoring of Multiple Sclerosis, while remaining a decision-support tool placed under the responsibility of professional neurologict.