Dual-energy CT liver fat fraction as prognostic imaging biomarker in critically ill patients

OBJECTIVES: To analyze changes in the liver dual-energy CT fat fraction (liver DECT FF) at the intensive care unit (ICU), and its prognostic value. MATERIALS AND METHODS: ). The skeletal muscle index, muscle radiodensity attenuation, subcutaneous/visceral adipose tissue area, and waist circumference were assessed. Pre-existing diseases, reasons for ICU admission, ICU scoring systems, and in-hospital mortality were noted. t-tests, Wilcoxon tests, linear or Cox regression, Pearson correlation, and intraclass correlation coefficients (ICC) were employed. RESULTS: Of 76 total patients, 43.4% were female (age 61 ± 12 years) and 97.4% received parenteral nutrition. In-hospital mortality was 60.8%. The liver DECT FF at CT1 was the only parameter associated with in-hospital mortality (hazard ratio: 1.15 [95% confidence interval: 1.07-1.24], p < 0.001), and a shorter ICU stay (-3.66 [-6.29 to -1.02] days, p < 0.007). DECT FF was not associated with ICU scoring systems. It did not relevantly change within a median of 8 days (3.3 ± 5.5% at CT1; 2.7 ± 4.3% at CT2, p = 0.315). Subset inter-observer reproducibility was good (ICC: 0.87 [0.66-0.95]). CONCLUSION: The liver DECT FF may serve as an independent prognostic imaging biomarker of mortality in critically ill patients, with a superior value to single-energy body composition parameters. KEY POINTS: Question Although the risk of liver steatosis is increased in critically ill patients, it is unknown if liver fat content, when quantified using DECT, is of prognostic value. Findings The liver DECT FF is associated with in-hospital mortality in critically ill patients. Clinical relevance The liver DECT FF may be superior for survival prognosis than single-energy CT body composition parameters of muscle and fat tissue at the ICU.