New predictive computer models designed to optimise HIV therapy in countries with limited healthcare resources have been published online. The models, which were developed with data from tens of thousands of patients around the world, accurately predict how an individual on failing therapy will respond to any new combination of HIV drugs.
The publication describes two new sets of models: one that does not require the genetic code of the virus, for use settings where HIV genotyping tests are unavailable, and another that includes this information for use in well-resourced settings. Both sets of models were developed with relaxed requirements for input data, again to suit low to middle income countries.
Both sets of models predicted the responses to the new regimen introduced in the clinic with approximately 80% accuracy. They were significantly more accurate than using genotyping, with state of the art interpretation, to predict responses. Both sets of models were able to identify combinations of locally available drugs that were predicted to produce a response in 90% or more of the cases that failed the new combination introduced in the clinic.
"These models represent a significant step forward towards the individualisation of HIV therapy in countries where genotyping is unavailable, treatment options are limited, and the selection of the best combination is particularly critical," commented Dr Brendan Larder, scientific chair of The HIV Resistance Response Database Initiative (RDI) and an author on the paper.
Currently, drug changes are not generally individualized but made according to set protocols, which can lead to sub-optimal treatments being introduced that can enable the development of drug resistance. Resistance is on the increase in many low to middle income countries, which poses a threat not only to the individual but to whole populations through the increased risk of onward transmission of drug-resistant virus.
The new models are now available to be used by healthcare professionals as part of the RDI's HIV Treatment Response Prediction System (HIV-TRePS), which is freely available online at www.hivrdi.org/treps.
The RDI's participation in this project is through a sub-contract with Leidos Biomedical Research, the prime contractor for the Frederick National Laboratory for Cancer Research, sponsored by the National Cancer Institute.
Objectives: Optimizing antiretroviral drug combination on an individual basis can be challenging, particularly in settings with limited access to drugs and genotypic resistance testing. Here we describe our latest computational models to predict treatment responses, with or without a genotype, and compare their predictive accuracy with that of genotyping.
Methods: Random forest models were trained to predict the probability of virological response to a new therapy introduced following virological failure using up to 50 000 treatment change episodes (TCEs) without a genotype and 18 000 TCEs including genotypes. Independent data sets were used to evaluate the models. This study tested the effects on model accuracy of relaxing the baseline data timing windows, the use of a new filter to exclude probable non-adherent cases and the addition of maraviroc, tipranavir and elvitegravir to the system.
Results: The no-genotype models achieved area under the receiver operator characteristic curve (AUC) values of 0.82 and 0.81 using the standard and relaxed baseline data windows, respectively. The genotype models achieved AUC values of 0.86 with the new non-adherence filter and 0.84 without. Both sets of models were significantly more accurate than genotyping with rules-based interpretation, which achieved AUC values of only 0.55–0.63, and were marginally more accurate than previous models. The models were able to identify alternative regimens that were predicted to be effective for the vast majority of cases in which the new regimen prescribed in the clinic failed.
Conclusions: These latest global models predict treatment responses accurately even without a genotype and have the potential to help optimize therapy, particularly in resource-limited settings.
Andrew D Revell, Dechao Wang, Maria-Jesus Perez-Elias, Robin Wood, Dolphina Cogill, Hugo Tempelman, Raph L Hamers, Peter Reiss, Ard I van Sighem, Catherine A Rehm, Anton Pozniak, Julio S G Montaner, H Clifford Lane, Brendan A Larder