Functional MRI to Identify High-risk CKD Phenotypes

PROJECT SUMMARY Patients with chronic kidney disease (CKD) are at increased risks of developing cardiovascular disease, progressing to end stage kidney disease, and dying prematurely. New strategies are needed to non-invasively identify targets to develop new therapies. Prior studies implicate decreased kidney perfusion and resultant chronic hypoxia in the pathogenesis of kidney fibrosis. The resulting chronic histopathologic lesions, which include interstitial fibrosis/tubular atrophy, global glomerulosclerosis, and microvascular disease lead to CKD progression. While chronic histopathologic lesions carry prognostic value, assessment requires a native kidney biopsy that is not performed in patients with common forms of CKD since a kidney biopsy carries risks. Non- invasive imaging biomarkers of chronic histopathologic lesions may enable selection of high-risk individuals to test new therapies or serve as surrogate endpoints to assess effectiveness of targeted interventions. However, non-invasive imaging remains underdeveloped in nephrology. Gadolinium-free kidney functional magnetic resonance imaging (fMRI) provides non-invasive assessments of kidney fibrosis, oxygenation, and perfusion. In prior work, our team incorporated kidney fMRI into a multicenter study and showed that kidney fMRI-derived biomarkers of fibrosis, hypoxia, and malperfusion associated with worse kidney function, differentiated patients with CKD from healthy volunteers, and may have prognostic value. Advancement of kidney fMRI in research and clinical practice requires further testing in a large cohort of patients with CKD that have gold standard assessments of histopathology. The Kidney Precision Medicine Project (KPMP) offers a unique opportunity to fill this gap. KPMP interrogates tissue obtained from research native kidney biopsy to identify critical pathways and targets that may ultimately lead to new therapies for patients with common forms of CKD. In preliminary data, we piloted kidney fMRI at a KPMP Recruitment Site (n=13) and showed high intra-reader, inter-reader, and within-participant agreement of the kidney fMRI biomarkers. In this proposal, we will perform kidney fMRI on 420 KPMP participants with CKD recruited from 7 Recruitment Sites and perform a 2-year follow-up kidney fMRI scan in a subset of participants (n=210). We will test the associations of kidney fMRI biomarkers with chronic histopathologic lesions and change in kidney function over time. We will extract kidney fMRI radiomic features to determine whether unsupervised clustering of radiomic features identifies imaging sub-phenotypes of CKD. To externally replicate our results, we will leverage de-identified data from a multicenter observational study of ~500 patients with diabetic kidney disease undergoing kidney fMRI in Europe. This proposal builds upon the work of an early-stage PI and fosters an ongoing cross-disciplinary collaboration among investigators experienced in clinical phenotyping, biomarker development and testing, and imaging. The results have the potential to inform whether kidney fMRI biomarkers identify sub-phenotypes of CKD or serve as surrogate markers of chronic histopathologic lesions in future drug development trials in patients with CKD.