Just a moment, the page is loading...

Development of personalised predictive models for the evolution of atopic dermatitis severity in response to treatment

Development of personalised predictive models for the evolution of atopic dermatitis severity in response to treatment

Reiko Tanaka

Department of Bioengineering, Imperial College London

24 March 2020

Atopic dermatitis (AD) is a chronic skin disease characterised by recurrent itchy skin inflammation that can severely impact the patients' quality of life. Current treatments aim to manage chronic AD symptoms by preventing exacerbations, mainly using topical corticosteroids, emollients, and systemic therapies. However, successful control of AD symptoms is challenging due to considerable variance in responses to AD treatment between patients. Personalised treatment strategies are considered to be more beneficial to individual patients rather than a “one-size-fits-all” approach to therapy.A first step toward developing personalised treatment strategies is to predict the consequences of possible treatments at an “individual level”. Assessing the average effects at a population level cannot deal with the variability across patients, and is not sufficient for personalised medicine. Predicting the treatment effects earlier and switching to other treatments if needed will help improve the efficacy for successful control of AD symptoms. However, there is currently no systematic method to predict AD treatment effects at an individual level, although the community has a large amount of data already accumulated from clinical studies. This project aims to develop a method to predict treatment effects at an individual level and to stratify patients based on treatment effectiveness. This new method will be vital for personalised medicine. Several recent studies have tried to stratify patients based on the effectiveness of some AD treatments, by identifying an association between biomarkers measured prior to the treatment and treatment responses at a single time point in future. However, data from a single time point measurement is likely to be noisy and may not capture the dynamic heterogeneity in AD treatment effects. Some other groups have successfully stratified dynamic disease trajectories of AD occurrence using the longitudinal birth cohort data. However, their proposed stratification is not suitable for designing treatment strategies as they only considered binary outcomes (e.g. presence or absence of an itchy rash) measured on a yearly basis. In this project, we will address this gap and aim to develop a computational model that can predict the consequences of possible treatments at an “individual level”, i.e. the dynamic evolution of disease outcomes (e.g. AD severity scores such as EASI) that are measured over weeks or months, the usual time-scales of interest for the design of personalised treatment plans. We will develop a method to develop personalised predictive models and evaluate its effectiveness using the data to be shared.

[{ "PostingID": 19532, "Title": "GSK-FPC117291", "Description": "A randomized, open-label, comparative study to evaluate an intermittent dosing regimen of fluticasone propionate 0.05% cream (twice per week) in reducing the risk of relapse when added to regular daily moisturization using PHYSIOGEL Lotion in paediatric subjects with stabilized atopic dermatitis" },{ "PostingID": 19589, "Title": "GSK-203121", "Description": "Study 203121: A Randomized, Blinded, Vehicle-Controlled, Dose-Finding Study of GSK2894512 Cream for the Treatment of Atopic Dermatitis" },{ "PostingID": 20100, "Title": "GSK-205050", "Description": "Study 205050: A Multicenter, Randomized, Double-Blind, Placebo-Controlled, Parallel-Group Study to Investigate the Efficacy and Safety of Mepolizumab Administered Subcutaneously in Subjects with Moderate to Severe Atopic Dermatitis" }]

Statistical Analysis Plan