Artificial intelligence for early diagnosis and prognosis of multiple sclerosis
Publication date: 18-09-2021
Updated on: 22-06-2022
Estimated reading time: 1 min
The San Raffaele’s study demonstrates a possibility to recognize Multiple Sclerosis in its various forms starting from the analysis of a blood sample with help of AI
Multiple Sclerosis (MS) affects 2.3 million people worldwide and 120,000 in Italy every year, but its diagnosis remains challenging: there are different types of MS, with different evolution and prognosis, and it is not always easy distinguish them accurately and early.
A group of researchers from the San Raffaele Research Hospital, coordinated by Cinthia Farina, the Chief of the Immunobiology of Neurological Disorders laboratory, has demonstrated a possibility to recognize Multiple Sclerosis in its various forms through artificial intelligence tools starting from the analysis of a blood sample with advanced genomics techniques. This is the first time when artificial intelligence algorithms have been applied to this kind of data and led us to important clinical implications: an effective and early diagnosis will help patients to start appropriate treatments.
The results of the study were published in the scientific journal Cell Reports Medicine. The study was carried out with the support of the Ministry of Health of Italy and Merck-Serono.
Forms and the problem of diagnosis
Multiple Sclerosis is a chronic autoimmune disease that affects the central nervous system and causes the loss of myelin (the sheath that covers neurons) leading to irreversible neurodegenerative damage. It is a complex disease, which can present itself in different forms. The most common at the time of diagnosis is the Relapsing-Remitting (RR) form, characterized by relapses alternating with periods of absence of or mild symptoms. Unfortunately, there are no specific markers to understand if and when the disease evolves from this form to the progressive form, from which there is no turning back and which foresees a progressive clinical deterioration.
Aforesaid corresponds to a delay in the administration of adequate treatment: there are many drugs for RR form and only one approved drug for the progressive form. Some patients presented with the progressive form from the beginning: it is difficult to recognize through magnetic resonance and the only way is to monitor those patients clinically over time.
An early diagnosis is essential to intervene more effectively.
Information in blood
Published in the Cell Reports Medicine journal, the study demonstrated for the first time that there is enough information in the blood to classify this neurological disease.
"It has always been assumed that blood was 'the same' in all MS patients, but by analyzing the data from subjects affected by the different forms of the disease, we realized that this is not the case," explains Dr. Cinthia Farina, coordinator of the study.
Work on this study lasted for years: first, it was necessary to collect blood samples from patients that include the full spectrum of various forms of the MS and who had not yet started treatment (which would otherwise have altered the results). Data from healthy subjects and patients with other neurological diseases were used as the control group. In all, the study includes data from over 300 individuals.
Since MS is an autoimmune disease caused by an abnormal reaction of the immune defenses, the researchers analyzed the activation status of peripheral blood mononuclear cells (immune cells responsible for protecting our body) thanks to transcriptomics (a technique that identifies which genes are on or off inside the cells).
The goal is to be able to identify the clinical profile of the disease and its likely course, starting from the state of the immune system.
"We will have to continue the studies in order to understand which blood markers are most suitable for classifying this pathology. Moreover, the algorithm has to be further refined and trained. It’s important to have demonstrated the possibility to 'see' what happens at the immunological level in patients starting from a blood sample. Our ultimate goal is to develop an early and effective diagnosis system based on this technology, and to understand more of the role of the immune system in the different forms of MS," explains Cinthia Farina.