Andalusian researchers are using artificial intelligence models to successfully predict 85% of hepatitis C cases among the undetected general population.

Researchers from the Progreso y Salud Foundation, in collaboration with the Digestive System Department of the Virgen del Rocío University Hospital, have developed a precision screening based on artificial intelligence, the algorithm of which has been trained and tested using more than one hundred thousand clinical cases to identify people. risk of contracting chronic hepatitis C in Andalusia and thus be able to reduce the number of tests carried out. The results of this work, Prediction of undiagnosed hepatitis C in patients were presented at the spring conference of the Spanish Association for the Study of the Liver (AEEH), which took place this Thursday and Friday in Seville and at which the progress towards the I consensus on personalized precision medicine in hepatology was presented.

The work, coordinated by the President of the AEEH, Manolo Romero, and the Director of the Andalusian Platform for Computational Medicine of the Progreso y Salud Foundation, Joaquín Dopaso, was presented this Thursday. The doctor assured that the results achieved so far are “very encouraging”, since this pre-screening will reduce the target population to be tested for hepatitis C much more than any other method used so far “with the prevalence of current hepatitis.” If we had to screen 143 people to identify a case using this prediction strategy, we see that we only need 7,” Dopazo explained. Data suggests that the use of AI could be of enormous value in diagnosing the estimated 20,000 cases that remain hidden in Spain, thereby accelerating the goal of hepatitis C elimination.

Identify patients much earlier

The work presented in Seville was carried out by selecting patients diagnosed with hepatitis C and matched control groups between 2017 and 2022. They were obtained from the Population Health Database, a unique resource of the Andalusian health system that contains data from more than 15 million people and can be used for clinical research. The algorithm was derived by analyzing a total of 6,000 hepatitis C cases diagnosed during this period, as well as approximately 120,000 controls with similar characteristics. Using artificial intelligence models such as XGBoost, they successfully identified in 85% of cases those patients infected with the hepatitis C virus who had not previously been diagnosed.

According to the researcher, “the use of models that can predict disease before it manifests represents a radical transformation of the healthcare system, because if we screen in this way, we will have a significant ability to detect and identify patients much earlier than they actually There is.

This study on precision prevention, pioneering in Andalusia regarding hepatitis C, is one example of the optimization of diagnostics provided by personalized precision medicine, a topic that was the focus of the AEEH spring conferences, in the application of these new technologies and also considered predictive tools for pathologies such as fatty liver , liver cirrhosis, hepatitis B, liver tumors or liver transplantation.

Plan to combat hepatitis C in Andalusia

Andalusia is one of the few autonomous communities that has a Hepatitis C Elimination Plan, in line with WHO’s goals of eliminating viral hepatitis by 2030. Together with Galicia, it is the only community that provides screening to the general population, which hepatologists say is necessary to eliminate this chronic viral infection for which there is no vaccine, but there is a treatment that cures it. According to Manuel Romero, President of the AEEH: “The project developed by the Progreso y Salud Foundation greatly opens the way to make this screening among the general population in the most effective way.”

Galicia is currently the only locality where hepatitis C is diagnosed among the general population. This is done by age cohorts (those groups in which the majority of infections are concentrated) and using a method already tested with COVID19, called sample pooling, which also will significantly reduce the number of PCR tests that need to be performed to obtain a result. diagnosis of HCV. “In Andalusia, which has a quarter of Spain’s population, solving the problem of screening the general population and doing it effectively requires even more, which is why the use of AI is being explored,” concludes Romero.