Currently there is no single specific parameter for multiple sclerosis (MS) enabling definite diagnosis. Despite diagnostic criteria, based mainly on MRI, it is mandatory to exclude numerous other diseases mimicking the clinical picture of MS. To date no tool has been designed which facilitates the diagnostic process of MS by combining complete medical data. Diagnosis is based mostly on doctors' experience instead of quantifiable indicators. In case of many patients protracted observation delaying therapeutic decision is necessary. Importantly, in case up to 13% patients initially diagnosed with MS final diagnosis differs. Project aim is to elaborate an analytical tool based on deep learning (DL) network, expediting differential diagnosis of MS. Data from MR, optic coherence tomography (OCT), immune and neurodegeneration markers examinations, neuropsychological and clinical data will be collected in the system. Based on 5000 cases (MS patients ca. 50% of the study group) primary patterns. Clinical paths defined by neurologists and radiologists will include parameters and with assigned weight describing the influence of each factor on final diagnosis. Correlation between parameters will also be assessed.
The final project results: