18th International HIV Drug Resistance Workshop Basic Principles & Clinical Implications June 9–13 2009, Fort Myers, Florida, USA

COMPUTATIONAL MODELS DEVELOPED WITHOUT A GENOTYPE FOR RESOURCE-POOR COUNTRIES PREDICT RESPONSE TO HIV TREATMENT WITH 82% ACCURACY

Antivir Ther 2009; 14 Suppl 1:A38 (abstract no. 36)

AD Revell¹, D Wang¹, R Harrigan², J Gatell³, L Ruiz⁴, S Emery⁵, MJ Pérez-Elías⁶, C Torti⁷, J Baxter⁸, F DeWolf⁹, Brian Gazzard¹⁰, AM Geretti¹¹, S Staszewski¹², R Hamers¹³, AMJ Wensing¹⁴, J Lange¹⁵, JM Montaner² and BA Larder^1
1HIV Resistance Response Database Initiative (RDI), London, UK; ²BC Centre for Excellence in HIV/AIDS, Vancouver, BC, Canada; ³Hospital Clinic of Barcelona, Barcelona, Spain; ⁴Fundació irsiCaixa, Badalona, Spain; ⁵National Centre in HIV Epidemiology and Clinical Research, Sydney, Australia; ⁶Ramón y Cajal Hospital, Madrid, Spain; ⁷University of Brescia, Brescia, Italy; ⁸Cooper University Hospital, Camden, NJ, USA; ⁹Netherlands HIV Monitoring Foundation, Amsterdam, the Netherlands; ¹⁰Chelsea and Westminster Hospital, London, UK; ¹¹Royal Free Hospital, London, UK; ¹²Hospital of the Johann Wolfgang Goethe-University, Frankfurt, Germany; ¹³PharmAccess Foundation, Academic Medical Centre, Amsterdam, the Netherlands; ¹⁴University Medical Centre, Utrecht, the Netherlands; ¹⁵Academic Medical Centre, Amsterdam, the Netherlands

BACKGROUND: Optimizing therapy following initial treatment failure in resource-poor countries is challenging due to limited treatment options and the relative unavailability of resistance testing. We have demonstrated that computational models trained with clinical data can predict virological response from genotype, viral load, CD4+ T-cell count and treatment history and may be a useful treatment selection tool. Here, we model response without genotype for potential use in resource-poor countries.

METHODS: Two random forest models were trained to predict the probability of the follow-up viral load being <50 copies/ml after a treatment change. Model 1 was trained with 8,214 treatment change episodes (TCEs), representative of clinical practice in resource-poor countries (no previous protease inhibitors [PIs], T-20, raltegravir or maraviroc, with PIs allowed in the new regimen). The input variables were baseline viral load, CD4+ T-cell count, treatment history (previous AZT, 3TC or any NNRTI), the drugs in the new regimen and time to follow-up. Model 2 was as Model 1 except 11 individual drug treatment history variables were used. Receiver operator characteristic curves were plotted for the models’ performance with an independent test set of 400 TCEs from different patients to those used for training. The models were used to identify alternative three-drug regimens available in resource- poor countries that were predicted to reduce the viral load to <50 copies/ml for the 205 test cases with incomplete response.

RESULTS: Both models achieved an accuracy of 82% and an area under the curve of 0.88. Sensitivity was 77% and 79% and specificity 86% and 85% for Models 1 and 2, respectively. The optimal operating points were 0.44 and 0.47. The most important variable in the modelling was baseline viral load. The two models identified alternative regimens that were predicted to be fully suppressive for 94 (46%) and 98 (48%) of the failures. When all currently licensed protease inhibitors were included, this rose to 100 (49%) and 106 (52%).

CONCLUSIONS: Models trained with sufficiently large, representative datasets can predict virological response accurately without a genotype. The results highlight the value of viral load information. This approach has potential for optimizing antiretroviral therapy in resource-poor countries.

2009-06-09
36

Copyright © 2009 - International Medical Press Ltd.. Reproduction of this abstract (other than one copy for personal reference) must be cleared through the International Medical Press Ltd. 2-4 Idol Lane, London EC3R 5DD UK.