Pattern identification of blood analytes for the discrimination and characterization of ALS and lower MND patients: a machine learning approach

Background: Although many studies demonstrate the involvement of altered metabolic pathways, the molecular mechanisms underlying the onset and progression of ALS/ MND are for the most part unknown. Biomarker-based monitoring tools for ALS/MND are not yet appropriate for routine clinical practice. The possibility to early predict the disease form - namely, patients with the selective involvement of the upper or the lower motor neuron (uMND and l-MND, respectively) or those with the involvement of both upper and lower motor neurons (ALS) is still far to be achieved. We are interested in finding out correlations between clinical and blood data, both synchronically and diachronically: we have built over time a database including more than a hundred patients with their clinical and biochemical data. Objectives: To look for specific patterns of blood analytes in ALS and l-MND patients, by using statistical and machine learning methods, to support the diagnosis and prognosis of ALS/MND patients.