3084-LB: Pancreatic Fat Infiltration and Diabetic Kidney Disease: CT Imaging and Genetic Analysis

Introduction and Objective: Pancreatic fat infiltration plays a significant role in the progression of diabetic microvascular complications (DMCs). Methods: Data were derived from a type 2 diabetes cohort study, including 666 patients from China, aged 70.9±7.7 years. Abdominal CT scans were used to assess organ-specific fat infiltration and volume, and a 3D U-Net convolutional neural network was applied for visceral segmentation. Continuous glucose monitoring (CGM) was used to measure glycemic fluctuations. Results: Our 3D U-Net model demonstrated high segmentation accuracy with dice similarity coefficients (DSC) of 0.96 for the liver, 0.78 for the pancreas, 0.95 for the kidneys, and 0.71 for the adrenal glands. Pancreatic fat infiltration was significantly associated with diabetic kidney disease (OR = 1.05, P = 0.02) and diabetic foot (OR = 1.04, P = 0.03). These associations were mediated by insulin resistance, with mediation effects of 29.8% and 5.3%, respectively. Excessive pancreatic fat was association with increased time of hypoglycemia, as measured by CGM. GWAS identified 9 significant SNPs associated with pancreatic fat deposition, including loci in MORC1, SDK1, and EYA1, which are involved in fat metabolism and vascular dysfunction. Conclusion: Pancreatic fat infiltration was associated with diabetic kidney disease, diabetic foot and time of hypoglycemia. Genetic loci associated with pancreatic fat provide new insights into the molecular mechanisms. Disclosure Y. Lu: None. Funding National Natural Science Foundation of China (82170870, 82470901)