Objectives: This study aims to quantify longitudinal associations between widely used timed function tests (TFTs) (i.e., 10-meter walk run (TMW), supine to stand (STS), climb 4 stairs (CFS)) and magnetic resonance (MR) imaging biomarkers1,2 (i.e., MR spectroscopy fat fraction (FF) measures of soleus (SOL) and vastus lateralis (VL)), and characterize the sensitive age range and relationships with clinically relevant covariates for each measure.
Methods: Six multivariate disease progression models linking the 3 TFT velocity endpoints and 2 FF measures were developed using natural history data of the ImagingNMD study (NCT01484678). The models were validated with placebo data from 3 clinical trials.3 To explore the covariate effects, 8 subgroups were generated by combinations of dichotomous steroid use, baseline score categories (i.e., low and high) of the linked measure, and baseline age (i.e., young and old) groups. Covariate effects were assessed by plotting the mean value of the ages at which the changes are half of their maximum (〖DPT〗_50) on the y-axis against the continuous baseline values of the evaluated outcome measure on the x-axis, with curves generated separately for each subgroup.
Results: Multiplication of the Chapman-Richards growth and sigmoid Imax function4 and sigmoid Emax function provided the best structural models for TFT velocity and FF measures, respectively. Steroid use delayed disease progression in STS and CFS velocity and FF of SOL trajectories by about two years, but not in TMW and VL trajectories. Baseline STS velocity values were the most consistent predictor of the trajectories, compared to the other TFT velocity measures. Baseline FF values were only shown to be an important predictor of individualized trajectories in VL for participants with older baseline age.
Conclusions: The developed multivariate models can serve as informative tools for guiding the optimal usage of outcome measures and selecting inclusion/exclusion criteria for DMD clinical trials.