Dissociation Between Hepatic Steatosis and Paraspinal Myosteatosis on Quantitative MRI

Abstract

Background:

Hepatic steatosis and skeletal muscle fat infiltration represent key features of metabolic dysfunction, yet their interrelationship remains poorly defined. Paraspinal muscles are routinely visualized during abdominal MRI, offering an accessible site for quantifying myosteatosis using proton density fat fraction (PDFF).

Objectives:

To identify metabolic determinants of hepatic and paraspinal muscle PDFF and assess whether paraspinal myosteatosis provides independent diagnostic value for diabetes mellitus.

Methods:

In this retrospective study, 266 adults who underwent abdominal MRI with quantitative fat mapping were evaluated. Hepatic and paraspinal PDFF were measured using MRI-PDFF, and metabolic associations were analyzed using correlation, regression, and logistic models. ROC curves assessed whether paraspinal PDFF improved diabetes discrimination beyond age, sex, body mass index (BMI), and hepatic PDFF.

Results:

Hepatic PDFF correlated strongly with BMI (r = 0.46) and ALT (r = 0.29), while showing no association with paraspinal PDFF (r = 0.01). Paraspinal PDFF was higher in diabetes (14.0% vs. 10.6%) but was not an independent predictor after adjusting for age, sex, BMI, and hepatic PDFF. Adding hepatic PDFF minimally improved diabetes prediction (AUC = 0.794→0.799), and paraspinal PDFF offered no additional improvement. An exploratory threshold of 11.8% paraspinal PDFF yielded moderate sensitivity (63%) and specificity (70%).

Conclusions:

Hepatic steatosis and paraspinal myosteatosis are metabolically distinct ectopic fat depots. Although paraspinal fat is elevated in diabetes, it does not enhance metabolic risk prediction beyond BMI. These findings support organ-specific lipid pathways and highlight the need for longitudinal and volumetric analyses to clarify their independent clinical roles.