P-A6 Validation of new informatics systems for routine HIV-1 genotypic and virtual phenotypic antiviral drug resistance analyses in clinical laboratories

Abstract

Clinical laboratories performing HIV-1 genotypic drug resistance testing need reliable and up-to-date software and database solutions to optimally analyze Sanger sequencing-generated data for interpretative reporting of results. Through a retrospective analysis of HIV-1 reverse transcriptase (RT) and protease (PR) sequences generated from the Trugene HIV-1 Genotyping assay (TG), we evaluated 2 analytical systems: a newly FDA-registered data processing module (DPM v1.0) based on several external validated drug resistance (DR) databases (including Stanford HIVdb (SD) and Geno2Pheno (G2P)), and the research-use-only ViroScore-HIV (VS) software. Results (NRTI, NNRTI and PI DR determinations) were compared with the respective interpretation (Guidelines v17.0) obtained from TG. HIV-1 tropism and DR to integrase inhibitors were not evaluated (not available in TG). Among 100 selected TG sequences obtained at the Mayo Clinic from March 2013 through May 2014, agreement of DR interpretative results between DPM v1.0 and VS was >99.9%. Agreement between TG and SD and between TG and G2P were both only 17%. Median agreement in DR interpretation between TG and SD, TG and G2P, SD and G2P were respectively 88.9%, 83.3%, and 82.6% for all drug classes, 86.5%, 83.5%, and 82% for PI class, 94.7%; 80.8%; and 91.6% for NRTI class (excluding AZT, DDI, D4T), and 94.8%, 81.7%, and 80.8% for NNRTI class (excluding RPV). With TG as the reference result, SD and G2P generated “positive” minor discordance (susceptible or S vs. intermediate or I, or I vs. resistant or R) to ≥1 drug in 66% and 56%, “negative” minor discordance (I vs. S, or R vs. I) in 32% and 54%, major discordance (S vs. R) in 6% and 15%, and major discordance (R vs. S) in 1% and 19% of subjects, respectively. DPM and VS were reliable for clinical laboratories to analyze RT and PR sequences for both research and routine clinical testing purposes. Differences in DR interpretation observed were the results of the up-to-date interpretive guidelines used by these applications.