While government consultants predict yet another catastrophic wave of Covid 19 that will force us to other lockdowns to be spent indoors, there are scientists, in Milan, committed to understanding why the virus has attacked some patients so violently that kill them and others has passed without leaving too many consequences. These are professors Carlo Tacchetti and Antonio Esposito, both professors of the Vita-Salute San Raffaele University, respectively director and vice-director of the Experimental Imaging Center of the IRCCS San Raffaele hospital in Milan, who although they are neither epidemiologists nor virologists , however, they found themselves forced to face the Covid emergency in the most affected region of Italy with thousands of people hospitalized right in their hospital. For this reason, the two doctors devised an autonomous learning platform which they called “AI-SCore” (acronym for Artificial Intelligence – Sars Covid Risk Evaluation), in summary they hypothesized that an algorithm could be able to calculate for each individual – on the basis of a series of clinical and diagnostic indicators – the probability of developing the most severe forms of Covid, thus allowing targeted health interventions and reducing the impact on the health system. Tacchetti, who specializes in Oncology, and his colleague Esposito, radiologist, started asking themselves some questions, the simplest of which is: who are the subjects who, once infected, even in the absence of serious symptoms, need timely treatment? Why are they the most affected by Covid?
It is now known that Coronavirus, despite having a low lethality rate, was fatal for the elderly, average age 80 years, with at least two previous pathologies. However, cases of young people who didn’t make it, who had to face months in intensive care, who were intubated and reacted worse than “grandparents” were not uncommon. This not knowing, in Phase 1, who were the most vulnerable or most predisposed to the virus certainly complicated the doctors’ work in the first weeks of the epidemic. The project began with the observation of the radiologist who had noted that in some swab negative subjects, but with symptoms attributable to Covid, the CT scans showed evident pulmonary impairment. In search of additional parameters that could help determine the prognosis of each patient, the two researchers soon understood that to relate all the numerous variables and observations that came from the work of the many doctors and scientists engaged in the struggle worldwide to Coronavirus, more computing power was needed than a single individual can field. For this reason, Tacchetti and Esposito asked two world giants of information technology such as Microsoft and Nvidia for help, the San Raffaele Omics Sciences Center directed by Dr. Giovanni Tonon, and the support of two companies: Orobix, an engineering company, production and governance of artificial intelligence systems and Porini, center of excellence for advanced analyzes. The data collection of over 2000 patients, recruited from San Raffaele, Bolognini hospital of Seriate and Centro Cardiologico Monzino, has already started, as well as the construction of the software infrastructure on which the algorithm will rest. In practice, it is as if between 140 and 150 parameters were entered into a big brain for each patient: the outcome of the CT scan, the hematoclinical examinations, the patient’s medical history, all associated also with the genome and with the utmost respect for privacy. “AI-SCore” processes and classifies each entry entered to provide medical personnel with all the information needed to respond to the various phases of the emergency. The algorithm is “trained” to predict the risk of the individual patient and works on the symptomatic, but it is not excluded that in the future (if the costs allow it) it can be used even on the asymptomatic, thus avoiding an enormous risk of contagion in the population . If all goes well it will be at full capacity between October and November and will benefit not only Italy but also abroad. Furthermore, given the cutting-edge results, it can serve in many other contexts in which it is essential to stratify the health risk and save lives.