Advancing AIDS Treatment Through Machine Learning

Share this post:

According to estimates, 17 million people worldwide are currently being treated for AIDS – that’s more than ever before despite the fact that the number of people being infected with the virus has gone down. The secret is the ‘cocktail’ that is helping AIDS patients live longer, which is essentially increasing the population of AIDs patients over the years.

Treatment by physicians today is ad-hoc, involving a search for the cocktail that is best suited for the current mutation of the AIDS virus in the specific patient. This new generation of artificially intelligent algorithms is being develop to go beyond predicting the success of the cocktail itself. Its goal is to forecast the period of time until the virus develops immunity to the cocktail. This allows a significant enhancement to how we manage and plan the patient’s treatments and care.

Using machine learning, IBM researchers at the company’s lab in Haifa Israel are contributing to a significant breakthrough in AIDS treatment by predicting how long it will take for the virus to develop immunity to a specific cocktail. Today, treatment for HIV patients involves a cocktail, a combination of medications that treat the person’s current viral load. These cocktails must be continuously updated as the virus changes and as the patient’s reactions to the treatment change over time.

The new technology is aimed at helping doctors provide AIDS patients with the optimal sequence of cocktails at every stage of the virus. This research is being done within the framework of an EU project, in which Israel collaborated with scientists from five EU countries.

Our work to fight AIDS began back in 2008, in cooperation with the laboratories of five European Union countries. The tool, called EuResist, helps doctors optimize the treatment of HIV patients by formulating the most effective cocktail for each specific patient. This tool, offered as a free service in the EU, is used by thousands of doctors from different countries.

Analyzing clinical, laboratory and demographic data accumulated over the years from across Europe and Africa allows the system to predict the right combination of drugs that works for the maximum amount of time, and to let physicians know when that time period is likely to end. These innovations are being provided as a free tool that can help extend the lives of millions of people who are the victims of this terrible disease.

An article from 2010 reported an experiment based on EuResist technology for formulating the cocktail. Results showed that the computerized system could outperform doctors when it came to predicting whether the treatment would succeed or fail. Nine out of 10 world-renowned doctors predicted success/failure of treatment less accurately that the system developed by the researchers.

The new research, now in advanced stages of development, is intended to help physicians formulate the best match between the cocktail and the current mutation of the virus. The significance of this development lies in its potential to greatly improve the planning and management of the treatment, and the ability to extend the patient’s life.

This development is expected to reduce patient suffering, to help the patient’s body cope more effectively with immune system failure, and save costly resources in the treatment process.