Go to  Advanced Search

Magnetic resonance imaging lesion count as a surrogate endpoint in relapsing-remitting multiple sclerosis clinical trials

Show full item record

Files in this item

Files Size Format Description   View
ubc_2011_fall_qin_lang.pdf 420.6Kb Adobe Portable Document Format   View/Open
 
Title: Magnetic resonance imaging lesion count as a surrogate endpoint in relapsing-remitting multiple sclerosis clinical trials
Author: Qin, Lang
Degree Master of Science - MSc
Program Statistics
Copyright Date: 2011
Publicly Available in cIRcle 2011-09-01
Abstract: The count of active lesions based on magnetic resonance imaging (MRI) is often used as a potential surrogate endpoint in phase 2 clinical trials for relapsing-remitting multiple sclerosis (RRMS) patients. However, this surrogacy relationship has not been completely validated. In this report, we study whether at the trial level, the MRI lesion count is a good surrogate endpoint for the relapse rate, the usual clinical endpoint for RRMS clinical trials. Two different approaches to assess this surrogacy relationship are applied to the dataset used by Sormani et al. [1] (SBRCMB) which contains the summary results from 23 randomized, placebo-controlled clinical trials in RRMS. The SBRCMB approach uses simple linear regression with weighted least squares estimation, while our more comprehensive approach develops a detailed model for the endpoints and the treatment effects to take into account estimation errors and the correlated contrasts. Both approaches are based only on the summary results from each clinical trial. The shortcomings of the SBRCMB approach are discussed and the results from the two approaches are compared. Both approaches show that the MRI lesion count is a good surrogate endpoint, while our more comprehensive approach shows a nearly perfect surrogacy relationship. When the estimated surrogacy relationship is used to predict the true treatment effect on the clinical endpoint for the trials in the SBRCBM dataset, the approaches give similar point predictions, but the approximate 95% prediction intervals from the comprehensive approach are generally shorter. In practice, the estimated surrogacy relationship based on the comprehensive approach can give a precise prediction for the true treatment effect on the clinical endpoint if the treatment displays a large effect on the surrogate endpoint, but may otherwise lead to an inconclusive result.
URI: http://hdl.handle.net/2429/37092
Scholarly Level: Graduate

This item appears in the following Collection(s)

Show full item record

All items in cIRcle are protected by copyright, with all rights reserved.

UBC Library
1961 East Mall
Vancouver, B.C.
Canada V6T 1Z1
Tel: 604-822-6375
Fax: 604-822-3893