PT-014 - UTILIZING A DISEASE PROGRESSION MODEL TO DEVELOP A CLINICAL TRIAL SIMULATION TOOL THAT OPTIMIZES TRIAL DESIGN FOR AUTOSOMAL DOMINANT TUBULOINTERSTITIAL KIDNEY DISEASE.

S. Ramesh¹, M. Rogge², J. Kim³, S. Kang⁴, K. Kidd⁵, A. Williams⁶, J. Roignot⁷, K. Blakeslee⁸, A. Bleyer⁶, S. Kim⁹; ¹University of Florida, University of Florida - Orlando, FL, USA, ²University of Florida, Orlando, FL, USA, ³University of Central Florida, Orlando, FL, USA, ⁴University of Florida College of Pharmacy Department of Pharmaceutics, FL, USA, ⁵Wake Forest University School of Medicine, Winston-Salem, NC, USA, ⁶Wake Forest Baptist Health, Winston-salem, NC, USA, ⁷Sailbio, Cambridge, MA, USA, ⁸Sailbio, Somerville, MA, USA, ⁹University of Florida, FL, USA.

Background: Autosomal dominant tubulointerstitial kidney disease (ADTKD), is a rare genetic disease often caused by a mutation in the protein uromodulin (UMOD). The rarity of subjects, high variability in onset, and slow progression of the disease will challenge recruitment, conduct, and design of future clinical trials that differentiate drug effect from the natural history of the disease. This work aimed to develop a Clinical Trial Simulation (CTS) tool based on our previously developed disease progression model and to provide a case example of identifying an ideal trial design for ADTKD.

Methods: R (version 4.2.3) and Simulx were used to develop the tool. A “high-risk patients” case study was formulated using chronic kidney disease trial literature data (i.e., end-stage kidney disease ≤ 40 years, 18 ≤ Age ≤ 40 years, 20 ≤ eGFR ≤ 45 mL/min/1.73m2, and 12 ≤ age of onset of gout ≤40 years). Only 8.9% of the subjects were qualified under this category in the original data set. To increase the number of high-risk subjects, a ctree synthesizing method was used while maintaining the distribution of covariates. Along with the existing surrogate endpoint, changes from baseline to last measures of estimated Glomerular Filtration Rate (eGFR), an additional plausible surrogate metric, the last measures of eGFR were also tested. A two-sided t-test (α = 0.05) was used to compare treatment vs. placebo groups in simulated trials.

Results: The developed CTS tool included three main components: i) selection criteria, ii) trial design, and iii) power curve. In simulation, at least a 15% increase in DPT50 (age at which eGFR is half of its maximum change) was necessary to differentiate the assumed drug effect from the natural history of the disease. To achieve 80% power in a clinical trial, the change from baseline to the last measure of eGFR required 100 subjects in each treatment group with 3 years of observation duration, whereas comparing the last measures of eGFR needed just 50 subjects per treatment group with 2.25 years of observation duration. The figure illustrates the power curve comparison between these metrics.

Conclusion: The model-informed CTS tool was successfully developed, which holds great promise in informing ADTKD trial designs with optimal sample size and trial duration through simulation before actual execution.