Israeli research giving HIV patients a fighting chance

An HIV infection is no longer a death sentence, as the number of approved treatments grows each year. But the virus is becoming resistant to an increasing number of drugs and is spinning off different strains. This anti-retroviral drug resistance …

An HIV infection is no longer a death sentence, as the number of approved treatments grows each year. But the virus is becoming resistant to an increasing number of drugs and is spinning off different strains. This anti-retroviral drug resistance is the primary reason why many treatments for HIV-infected patients fail.

That’s why the best way to use anti-HIV drugs is in combinations or ‘cocktails’ that are prescribed as the patient’s individual virus progresses and as resistance to the drugs change. Although there are new standardized systems to monitor the development of drug-resistant mutations, there is a vital need for a method that will help doctors decide which cocktail has the best chances of success for different individuals, each with their own unique variant of the virus.

That’s where researchers in Israel at the IBM Haifa Research Lab (HRL) come in. They were asked to take part in a EU consortium aimed at developing an integrated system for anti-HIV treatment. The goal? To perfect EuResist, a European integrated system for clinical management of antiretroviral drug resistance. The system will provide clinicians with a prediction of response to antiretroviral treatment for HIV patients, thus helping them choose the best drugs and drug combinations for any given HIV genetic variant.

“Monitoring the history of treatments and the progress of the virus itself is crucial to successful patient care,” said Boaz Carmeli, the manager of Healthcare and Life Science at the IBM Haifa Research Lab. “Tapping into knowledge garnered from a huge collection of data will help doctors take into account the patient, the virus, the viral mutations, and the current stage of the disease.”

Since 1981, approximately 1.7 million Americans have been infected with HIV, over half a million of them who have succumbed to the virus. There are approximately 40,000 new infections reported each year.

EuResist is using an innovative approach to predict the efficiency of anti-retroviral drug regimens against a specific HIV, based on viral genotype data integrated with treatment response data collected from some of the largest HIV databases in Europe. The project’s biomedical information integration technology gathers data from three large genotype-response databases, namely the Italian ARCA database (one of the biggest in the world), the German AREVIR database, and data coming from the Karolinska Infectious Diseases and Clinical Virology department. The data includes treatment histories, treatment response information, and the sequence of the relevant part of the HIV genome (genotype). The resulting EuResist integrated data set is expected to be the largest in the world.

Aside from the HRL team in Haifa, the other partners participating in this European Union 6th Framework project include: Informa S.r.l., Universit degli Studi di Siena, Italy; Karolinska Institute, Sweden; Max-Planck-Institute for Informatics, University Hospital of Cologne, Germany; RMKI, Hungry; Kingston University; and the European Federation of Pharmaceutical Industries and Associations (EFPIA).

“If we look closely at the current patient’s blood work, virus stage, family history, race, and so forth – and then compare it to the thousands of people who have been treated over the years, we can see what was done, what worked, and what didn’t,” saidMauricio Zazzi, Prof. Maurizio Zazzi, EuResist Scientific Coordinator and Professor of Microbiology at the University of Siena School of Medicine. On the basis of this history data, EuResist can predict how the virus will respond to a certain cocktail. “This method not only provides a huge savings in costs, it also means a patient’s chances for successful treatment are not dependant on their doctor’s individual knowledge.”

With EuResist, this interaction is done through the web, where physicians can input a patient’s information and status and then get a summary of what is known about this specific virus stage along with a prediction of what treatment has a good chance of helping the patient.

For example, a doctor in Bolivia – who may not have expertise in AIDS treatment or access to recent research – can use the knowledge accumulated in the EuResist system to treat patients. “This access to shared knowledge greatly increases our chances of fighting AIDS and can provide a vital contribution to world healthcare,” continues Zazzi.

HRB’s contributions to the project from Israel are two fold. The Healthcare and Life Science group has implemented a standardized biomedical information technology that processes and correlates clinical and genomic data from various data sources. And The lab’s Machine Learning group has developed a sophisticated model and training engine that helps predicts drug resistance.

The EuResist database currently has access to information from over 17,000 patients and is growing all the time. Although the different databases involved in the project all contain HIV data, each collection was stored using a different format and had its own unique understanding of what information is most important. The IBM team was challenged with creating order from the masses of data and developing an HIV-specific schema that would filter out the data needed for the system’s prediction engines.

Since it first opened as the IBM Scientific Center in 1972, the IBM Haifa Labs have conducted decades of research that has been vital to IBM’s success. R&D projects are being executed today in areas such as storage systems, verification technologies, multimedia, active management, information retrieval, programming environments, optimization technologies, and life sciences.

IBM Haifa researchers have a long history of leading industry based standards. They didn’t wait long before getting work underway for a specific clinical document architecture (CDA); this is basically an XML structure or template that defines the information used for HIV treatment of patients. When news of this document structure spread in Europe, the project gained visibility and many additional institutes volunteered to donate their data to the EuResist database.

“Delivering a standard CDA for HIV treatment is an important step in enabling various institutes and healthcare organizations to share their data,” explained Carmeli. “This paves the way for international collaboration and joint global efforts to fight disease.”

The EuResist project involves the development of several prediction engines of different flavors, with each engine designed to predict the efficiency of possible drug combinations and to recommend an optimal treatment. The different kinds of prediction engines include: evolutionary models, graph theoretical models, mutual-information based data mining, case-based reasoning, and machine learning.

The IBM Haifa team’s prediction engines are based on machine learning, an expertise for which the Haifa Lab is already widely recognized. “The system integrates the predictions output by the different engines and presents the physician with a consolidated set of results,” explained Shai Fine, manager of the Machine Learning group at the IBM Haifa Research Lab. “The challenge lies in understanding which engine is optimal for different patient scenarios.” So far, this method is producing results that are more accurate than those obtained from any single engine.

The project results so far are showing a tremendous success rate of 75%. The treatment predictions being output by the system are currently being compared with training sets and with actual patient data to further determine their accuracy.

“A system needs massive amounts of data in order to output accurate predictions,” reflects Zazzi. Combining the efforts and expertise of partners from across the European community has given EuResist the ingredients it needs for success. “This learning system will continue gaining knowledge about new treatments as they arise and adjust itself to give even better answers. We’re getting better all the time.”

Slashdot It!


*/