Predicting the outcomes of HIV treatment interruptions using computational modelling
Hillmann, AndreasORCID: 0000-0002-6517-1615
(2021)
Predicting the outcomes of HIV treatment interruptions using computational modelling.
PhD thesis, Dublin City University.
In the past 30 years, HIV infection made a transition from fatal to chronic disease due to the emergence of potent treatment largely suppressing viral replication. However, this medication must be administered life-long on a
regular basis to maintain viral suppression and is not always well tolerated. Any interruption of treatment causes residual virus to be reactivated and infection to progress, where the underlying processes occurring and
consequences for the immune system are still poorly understood. Nonetheless, treatment interruptions are common due to adherence issues or limited access to antiretroviral drugs. Early clinical studies, aiming at
application of treatment interruptions in a structured way, gave contradictory results concerning patient safety, discouraging further trials. In-silico models potentially add to knowledge but a review of the Literature indicates most
current models used for studying treatment interruptions (equation-based), neglect recent clinical findings of collagen formation in lymphatic tissue due to HIV and its crucial role in immune system stability and efficacy. The aim
of this research is the construction and application of so-called ‘Bottom-Up’ models to allow improved assessment of these processes in relation to HIV treatment interruptions. In this regard, a novel computational model based on
2D Cellular Automata for lymphatic tissue depletion and associated damage to the immune system was developed. Hence, (i) using this model, the influence of spatial distribution of collagen formation on HIV infection
progression speed was evaluated while discussing aspects of computational performance. Further, (ii) direct Monte Carlo simulations were employed to explore the accumulation of tissue impairment due to repeated treatment interruptions and consequences for long-term prognosis. Finally, (iii) an inverse Monte Carlo approach was used to reconstruct yet unknown characteristics of patient groups. This is based on sparse data from past
clinical studies on treatment interruptions with the aim of explaining their contradictory results.
Metadata
Item Type:
Thesis (PhD)
Date of Award:
March 2021
Refereed:
No
Supervisor(s):
Crane, Martin and Ruskin, Heather J.
Uncontrolled Keywords:
HIV Modelling; Monte-Carlo Simulation; Differential Equations Models