Poster Presentation at:
45th ICAAC, Washington DC, USA, 16-19 December 2005.

Treatment history but not previous genotype improves the accuracy of predicting virologic response to HIV therapy

B.A. Larder 1 , D. Wang 1 , A. Revell 1 , R. Harrigan 2 , J. Montaner 2 , S. Wegner 3 , and C. Lane 4 .

1: HIV Resistance Response Database Initiative, London, UK; 2: BC Centre for Excellence in HIV/AIDS Vancouver, Canada; 3: US Military HIV Research Program, Rockville, MD, USA; 4: National Institute of Allergy & Infectious Diseases, Bethesda, MD, USA.

Background: Previous exposure to HAART can result in archived resistant virus, undetectable by conventional genotyping, which may undermine the success of new treatment regimens. We assessed the influence of previous treatment or genotype history on the accuracy of artificial neural networks (ANN) predicting virologic response to HAART.

Methods: Study 1: Basic ANN models were trained using 2,559 treatment change episodes (TCEs), i.e. current genotype, baseline viral load (VL), new drugs, follow-up time and VL response. Treatment history models were trained with 4 additional input variables - previous treatment with any of the following: AZT, 3TC, any NNRTI and any PI.

Study 2: Basic models were trained as above and the genotype history models were trained using 301 TCEs containing cumulative data from all available genotypes up to and including baseline.

The ANNs were tested using the input variables of independent test TCEs, which produced a predicted VL response that was compared to the actual response.

Results: In all cases, averaged output from 10 ANN models was used. Correlations between predicted and actual VL change for basic and treatment history models gave values of 0.30 and 0.45 respectively (p<0.01). The correlations for the basic and genotypic history models gave values of 0.39 and 0.31 respectively (ns).

Conclusions: The addition of treatment history data significantly improved the accuracy of ANN confirming that historical exposure to antiretroviral drugs can influence response to a new regimen. This suggests that treatment history may act as a surrogate for minority mutant populations enabling ANN to overcome this shortcoming of current genotyping. The use of cumulative historical genotype data did not improve the accuracy of the ANN.