Treatment, Prevention, Remission and Cure: Statistical Challenges and Advances in HIV Drug Development
Organizers: Qiming Liao (ViiV), Cliburn Chan (Duke)
Chair: Qiming Liao (ViiV)
Vice Chair: Ralph DeMasi (ViiV)
Kevin Wiehe (Duke)
Ralph DeMasi (ViiV)
Hengrui Cai (NCSU)
Qiming Liao (ViiV)
Title: HIV Vaccine Design Strategies Based on Statistical Inference of B Cell Maturation
Speaker: Kevin Wiehe (Duke)
HIV-1 vaccine design is greatly hindered by the virus’ rapid mutation rate which allows it to evade the host immune response and results in the extraordinary diversity of global HIV strains. In rare cases, HIV infected individuals make broadly neutralizing antibodies (bnAbs) that can neutralize 70-99% of all global HIV strains. If similar types of antibodies could be induced by vaccination, they would provide protection. Current efforts in HIV vaccine design have therefore coalesced around a strategy of recapitulating development of bnAb lineages similar to those observed in infection. These “lineage-based” approaches present unique challenges in statistical inference including inferring the unmutated state of a B cell receptor, partitioning antibody sequences into clonally related groups, and reconstructing B cell clonal genealogies to estimate maturation pathways. This talk will review these challenges and approaches to address them. In addition, we will discuss our development of a computational tool that simulates somatic hypermutation to estimate the probability of antibody mutations prior to antigenic selection. We will demonstrate the application of this tool to studying HIV-1 bnAb lineages and show that rarely-targeted mutations critical for gain of neutralization during maturation can act as rate-limiting steps in the development of HIV-1 bnAbs. Finally, we will show how targeting such “improbable mutations” for selection by specifically designed immunogens relieves maturational bottlenecks in immunized mice and is a potentially generalizable vaccine strategy for engineering B cell maturation towards a desired antibody response.
Title: Current Initiatives in HIV Clinical Development: A Statistical Perspective
Speaker: Ralph DeMasi (ViiV Healthcare)
Since the approval of AZT monotherapy for HIV treatment in 1987, continuous improvements of HIV treatment and prevention strategies have made HIV a chronic, but manageable, disease and has resulted in a dramatic decline in the incidence of new HIV infections. Despite this, there is a significant unmet medical need for development of additional safe and efficacious HIV medicines to reach the WHO/UNAID/DHHS goal of eradicating the HIV epidemic by the end of the decade. To meet this goal, HIV research and development continues to be very active, and several medicines from a range of therapeutic classes targeting different stages of the viral life cycle are currently under development, as well as medicines and strategies for HIV prevention and remission/cure. Within this backdrop many statistical challenges remain, offering considerable opportunity for novel and innovative methods for statistical design and analysis. This talk will highlight some current statistical challenges with emphasis on initiatives related to new classes of HIV treatments, modes of administration, prevention modalities/strategies and special populations.
Title: GEAR: On Optimal Treatment Decision Making by Auxiliary Data with Application to AIDs Study
Speaker: Hengrui Cai (NCSU)
Personalized optimal treatment decision making, which finds the optimal treatment decision rule (ODR) based on individual characteristics, has attracted increasing attention recently in many fields, such as medicine, education, and economics. Current ODR methods usually require the primary outcome of interest in samples for assessing treatment effects, namely the experimental sample. However, in many studies, treatments may have a long-term effect, and as such the primary outcome of interest cannot be observed in the experimental sample due to the limited duration of experiments, which makes the estimation of ODR impossible. We address this challenge by making use of an auxiliary data source, namely the auxiliary sample, to facilitate the estimation of ODR in the experimental sample. Both samples have common baseline covariates and intermediate outcomes that are highly related to the long-term primary outcome, while the auxiliary sample contains the primary outcome of interest but not the treatment information that is available in the experimental sample. We propose an auGmented inverse propensity weighted Experimental and Auxiliary sample-based decision Rule (GEAR) by maximizing the augmented inverse propensity weighted estimator of the value function over a class of decision rules using the experimental sample, with the primary outcome being imputed based on the auxiliary sample. The proposed GEAR is applied to the AIDS Clinical Trials Group Protocol 175 data.