Logo image
Better Together: Harmonizing Multicenter T2 MRI Data to Understand Brain Iron Deposition in Pediatric Patients with CKD: TH-PO1193
Abstract   Peer reviewed

Better Together: Harmonizing Multicenter T2 MRI Data to Understand Brain Iron Deposition in Pediatric Patients with CKD: TH-PO1193

Mckade Poirier, Ellen van der Plas, Timothy R. Koscik, Brian Becknell, Lyndsay Harshman and Emily JS Steinbach
Journal of the American Society of Nephrology, Vol.37(10S)
10/2026
DOI: 10.1681/ASN.2026enakn5s7

View Online

Abstract

Background: Chronic kidney disease (CKD) in children is associated with neurocognitive impairment and altered iron metabolism. Excess iron accumulation in the brain may contribute to oxidative stress and neurodevelopmental effects seen in CKD. T2* mapping with magnetic resonance imaging (MRI) provides a noninvasive marker of brain iron deposition, making it a tool for studying neurodevelopmental effects of CKD. However, pediatric CKD neuroimaging studies are limited by small sample sizes. Multicenter collaboration is necessary for adequately powered studies, but site related variability may limit generalization of imaging findings across sites. ComBat harmonization reduces inter-site variability while preserving biologically meaningful signal, enabling pooled multicenter analyses but has not been applied to pediatric T2* MRI. We evaluated the feasibility of ComBat harmonization in pediatric multicenter T2* MRI data and assessed regional brain iron differences in children with CKD. Methods: Children with CKD (n=49) and unaffected controls (n=69) underwent T2* MRI at the University of Iowa (n=110) and The Ohio State University (n=8). ComBat harmonization was applied to ROI-based T2* relaxation times to reduce scanner and site variability, and its effectiveness in reducing site effects was evaluated. Linear regression models examined associations between CKD status and T2* relaxation times across predefined ROIs, adjusting for age and sex. False discovery rate correction addressed multiple comparisons, and Cohen’s f estimated effect sizes. Results: Prior to ComBat harmonization, imaging site significantly influenced regional T2* relaxation times across multiple ROIs, indicating substantial inter-site variability. Following harmonization, site effects were no longer significant, supporting the feasibility of pooling multicenter pediatric T2* MRI data while preserving biologically relevant variation. Conclusion: These findings demonstrate that ComBat harmonization effectively reduced scanner- and site-related variability and enabled integrated analysis of pediatric multicenter T2* MRI data. This approach may facilitate adequately powered multicenter studies of rare pediatric diseases and improve the generalizability of neuroimaging findings across institutions. Larger longitudinal multicenter studies are needed to further evaluate CKD-related alterations in brain iron metabolism and neurodevelopment.

Details

Metrics

1 Record Views
Logo image