Identifying Subgroups of Response to Biologics in Psoriasis using Large-scale Patient-level Data

Proposal
1693

Title of Proposed Research

Lead Researcher

Nophar Geifman

Affiliation

University of Manchester

Funding Source

None

Potential Conflicts of Interest

None

Data Sharing Agreement Date

19 February 2018

Lay Summary

Biologic therapies, a class of drugs modifying specific parts of the immune system's functioning, have led to remarkable improvements in outcomes for many patients suffering from Immune Mediated Inflammatory Diseases (IMID) such as psoriasis. However, these drugs are expensive, may result in serious adverse events and the response to treatment is variable. A precision medicine approach to psoriasis therefore needs to identify phenotypically distinctive subgroups of patients who may be more or less likely to respond to biologic therapies. We aim to target these therapies by using large-scale patient-level data to identify subgroups of patients for whom treatment with biologics will be most beneficial.
Using sophisticated computational methods, we aim to identify and characterise subgroups of patients demonstrating different patterns of response to treatment. To achieve this, we will apply our analytical methods to large-scale integrated patient-level data which will effectively provide sufficiently large enough cohorts to allow identification of clinically meaningful subgroups. These cohorts can be generated by employing a pooled meta-analysis approach where patients treated with the same medication or with similar disease characteristics are pooled together from a number of clinical trials and analyzed for patterns of response to treatment. By focusing on the distinguishing characteristics of these subgroups at baseline, the prediction of response to treatment might usefully be stratified. The results of our analyses will help gain a better understating of the mechanisms underlying response to biologics therapies in psoriasis and identify a responder subset which will benefit most from treatment. Analysis results will be disseminated in appropriate meetings, conferences and publications.

Study Data Provided

[{ "PostingID": 4444, "Title": "NOVARTIS-CAIN457A2302", "Description": "A Randomized, Double-blind, Placebo Controlled, Multicenter Study of Subcutaneous Secukinumab to Demonstrate Efficacy After Twelve Weeks of Treatment, and to Assess the Safety, Tolerability and Long-term Efficacy up to One Year in Subjects With Moderate to Severe Chronic Plaque-type Psoriasis

Medicine: secukinumab, Condition: Psoriasis, Phase: 3, Clinical Study ID: CAIN457A2302 , Sponsor: Novartis." },{ "PostingID": 4445, "Title": "NOVARTIS-CAIN457A2303", "Description": "A Randomized, Double-blind, Double-dummy, Placebo Controlled, Multicenter Study of Subcutaneous Secukinumab to Demonstrate Efficacy After Twelve Weeks of Treatment, Compared to Placebo and Etanercept, and to Assess the Safety, Tolerability and Long-term Efficacy up to One Year in Subjects With Moderate to Severe Chronic Plaque-type Psoriasis

Medicine: secukinumab , Condition: Psoriasis, Phase: 3, Clinical Study ID: CAIN457A2303 , Sponsor: Novartis." },{ "PostingID": 4453, "Title": "LILLY-I1F-MC-RHBA", "Description": "A Multicenter, Randomized, Double-Blind, Placebo-Controlled Study Comparing the Efficacy and Safety of LY2439821 to Etanercept and Placebo in Patients With Moderate-to-Severe Plaque Psoriasis. UNCOVER-2" },{ "PostingID": 4454, "Title": "LILLY-I1F-MC-RHBC", "Description": "A 12-Week Multicenter, Randomized, Double-Blind, Placebo-Controlled Study Comparing the Efficacy and Safety of LY2439821 to Etanercept and Placebo in Patients With Moderate to Severe Plaque Psoriasis With a Long-Term Extension Period UNCOVER-3" }]

Statistical Analysis Plan

Our analyses will focus on the identification of clinically meaningful subgroups with homogeneous treatment response. We will use latent class regression methods to identify these subgroups, after which we will use patient-level variables to define their distinguishing characteristics and investigate the interaction between patient characteristics and treatment. Principal component analysis will be used to assess the data and determine which measurements and outcomes to use for further investigation. Depending on the type and structure of the response data available, we will use either linear or logistic regression analysis, or survival analysis when the timing of outcomes is known. Where multiple responses over time have been recorded, we will use latent class mixed models with a random intercept and random coefficients for each individual. The Mixed-effects will be used to account for the likely correlation of repeated measurements within the same patient. The model will be adjusted for potential confounders such as age, BMI, gender, ethnicity etc. We will test the model for 1-10 latent classes and the optimal number of latent classes assessed using the Bayesian Information Criterion (BIC); the model which produces the lowest BIC and a reasonable number of patients in all the resulting classes (at least 5%) will be selected for further investigation.To preserve statistical power, analyses will be performed in a pooled dataset if possible (i.e. when there is sufficient overlap in structure and type of data across the analysis sets). Otherwise we will analyze each dataset separately and combine the results using random effects meta analysis.Finally, categorical patient characteristics (e.g. smoking status and specific therapy) will be contrasted for the resulting subgroups using Chi-square tests. Analyses of variance will be applied to the continuous variables (e.g. age at baseline and BMI). Association of comorbidities with the resulting subgroups will be explored with tests of binomial proportions.

Publication Citation

Nophar Geifman, Narges Azadbakht, Jiaping Zeng, Toby Wilkinson, Nick Dand, Iain Buchan, Deborah Stocken, Richard B. Warren, Jonathan Barker, Nick J. Reynolds, Michael R. Barnes, Catherine H. Smith, Christopher E. M. Griffiths, Niels Peek, and the BADBIR Study Group, on behalf of the PSORT Consortium. British Journal of Dermatology, Volume 185, Issue 4, 1 October 2021, Pages 825-835. Defining Treatment Response Trajectoriesin Psoriasis using Large-scale Patient-level Data.
DOI: 10.1111/bjd.20140