Liver MRI–derived Proton Density Fat Fraction Predicts Diabetes Risk in Patients with Obesity: A Dose-Response Study Ming Deng, Zhen Li, Li Sun, Huawei Wang, Jun Tang , Yanxin Liu, Xunuo Chen, Haibo Xu , Zhe Dai Radiology. 2025 Oct;317(1):e250026. doi: 10.1148/radiol.250026.
This study tried to assess the relationship between Proton Density Fat Fraction and risk of prediabetes or type 2 diabetes in patients with obesity and liver enzymes. Metabolic dysfunction–associated steatotic/fatty liver disease (MASLD/MAFLD) is a growing health problem linked to type 2 diabetes mellitus, but its relationship with a risk of prediabetes or type 2 diabetes in patients with obesity remains unclear. MRI-derived proton density fat fraction (PDFF) is a biomarker and accurately quantify liver fat within 20 seconds. This was a retrospective study which included 683 adult patients with obesity (BMI greater than 28) who underwent MRI to determine PDFF and who were administered an oral glucose tolerance test between September 2020 and December 2023. A validation cohort of 100 patients was analyzed independently. Multinomial logistic regression assessed the association between liver PDFF and prediabetes or type 2 diabetes. Mediation analysis evaluated the levels of alanine aminotransferase and aspartate aminotransferase. The findings were validated in an independent cohort. A dose-response relationship was found between liver PDFF and abnormal glucose metabolism in the stuyd group. Severe steatosis (PDFF ≥ 25%) was associated with higher odds of prediabetes (OR, 3.49, P = .01) and type 2 diabetes (OR, 3.45;P = .01). The association was partially mediated by aspartate aminotransferase (mediation proportion, 27.3%) and alanine aminotransferase (mediation proportion, 38.8%) (P < .001 for both). These findings were validated because patients with dysglycemia had higher mean liver PDFF than did patients with normoglycemia (19.7% vs 12.9%, respectively; P < .001). Higher liver fat is associated with metabolic dysfunction. Increasing PDFF correlated with: higher fasting and 2-hour glucose, higher insulin and C-peptide levels and greater insulin resistance (HOMA-IR). Limitations of the study were: being a single center study, the cross-sectional design preventing proof of causality, absence of liver biopsies, which is very common in such studies. | References
|

