SUN-294 FUNCTIONAL MRI IN PREDICTING DIABETIC KIDNEY DISEASE

Abstract

There are no known studies evaluating functional renal Magnetic Resonance Imaging (MRI) changes in Type 1 Diabetes (T1DM) compared to healthy controls. This is a prospective pilot study with a primary aim to evaluate if multi-parametric MRI can detect renal hypoxia (utilizing BOLD, Blood Oxygen Level Dependent) and structural changes (utilizing diffusion tensor imaging, DTI) in patients with T1DM, comparing patients with different stages of diabetic kidney disease (DKD) to healthy matched controls. 34 T1DM patients with variable degrees of renal impairment and 10 controls with intact renal function were recruited. Following an overnight fast, participants underwent physical examination, blood tests, 24-hour urine collection and functional MRI with controlled water intake. MRI parameters from two independent readers were obtained and analyzed. A value of T2*, from which resonance transverse relaxation rate (R2*) is calculated (1/T2*) is obtained from BOLD images. Functional anisotropy (FA) and apparent diffusion coefficient (ADC) values were obtained from DTI images. To determine measured Glomerular Filtration Rate (mGFR), each patient with T1DM underwent 99m-technetium- diethylene-triamine-penta-acetic acid (99mTc-DTPA) plasma clearance with Brochner-Mortensen correction. Univariate analysis was conducted with Mann-Whitney tests. The median mGFR for participants with T1DM was 105 (77,120) ml/min/1.73m2. 31% had hyperfiltration arbitrarily defined as an mGFR³120ml/ min/1.73m2. Participants with hyperfiltration were younger than the normofiltering group (31.5 years vs. 49.5 years, p=0.006). There were no other significant baseline characteristic differences between the two groups. There were also no significant baseline differences between participants with T1DM and healthy controlsR2* was significantly lower both in the cortex (14.7 [13.7-15.8] vs 15.7 [15.1-16.6], p<0.001) and medulla (24.8 [21.8,28.2] vs 29.3 [24.3,32.4], p<0.001) in participants with diabetes compared to healthy controls, respectively. Medullary FA was higher; while cortical FA was lower in diabetes compared to controls. On the other hand, there were no significant differences in both cortical and medullary ADC in participants with and without diabetes. Univariate comparison of MRI parameters between normofiltration and hyperfiltration showed no significant difference for cortical R2*, cortical FA and medullary FA. There was a marginally higher medullary R2* in the hyperfiltration 25.8(23.3,28.7)(1/s) compared to normofiltration 24.4(21.4,27.9)(1/s) (p=0.01) group. ADC in the cortex 2.61(2.53,2.74) (103mm2/s) vs. 2.40(2.29,2.53) (103mm2/s) (p<0.001) and medulla 2.46(2.34,2.58) vs. 2.27(2.08,2.48) (p=0.02) were also observed to be higher in the hyperfiltration group. A lower R2* in our diabetes cohort reflect higher intra-renal oxygen content in diabetes. Higher medullary R2* suggest that hypoxic changes in hyperfiltration to those without hyperfiltration. FA changes in T1DM compared to controls; as well as ADC changes in hyperfiltration also suggests that renal structural changes in DKD may be detected with functional MRI. Further larger studies are required to evaluate the utility of functional MRI in T1DM and hyperfiltration.