Liver Diffusion Weighted MRI: Effect of Iron Overload on Apparent Diffusion Coefficient

ABSTRACT

We assessed the effect of hepatic iron levels on liver apparent diffusion coefficient (ADC) values assessed by magnetic resonance imaging (MRI) and the influence of different b ‐values on the extent of this association. We prospectively enrolled 110 patients (60 women, 33.05 ± 7.86 years) with beta‐thalassemia. Single‐shot echoplanar diffusion‐weighted imaging (DWI) with different b values was acquired, and ADC_200 ( b ‐values 0/200 s/mm ² ), ADC_600 ( b ‐values 0/600 s/mm ² ), and ADC_1000 ( b ‐values 0/1000 s/mm ² ) were obtained. Liver T2* values were measured with a gradient–echo multiecho sequence and were converted into liver iron concentration (LIC) values. There was a significant difference among the three different ADC values ( p  < 0.0001). Mean MRI LIC values were 7.99 ± 10.09 mg/g dw, and 64 (58.2%) patients had liver iron overload (LIC > 3 mg/g dw). Significant negative correlation was found between MRI LIC values and ADC_200 values ( R  = −0.284, p  = 0.003), ADC_600 values ( R  = −0.645, p  < 0.00001), and ADC_1000 values ( R  = −0.842, p  < 0.0001). MRI LIC was more strongly correlated with both ADC_600 and ADC_1000 than with ADC_200 ( p  < 0.00001 for both comparisons) and with ADC_600 than with ADC_200 ( p  < 0.0001). ADC_200 values were comparable between patients without and with liver iron overload (2.11 ± 0.53 vs. 1.87 ± 0.89 × 10–3 mm ² /s, p  = 0.081), while patients with liver iron overload had significantly decreased ADC_600 values (0.91 ± 0.72 vs. 1.52 ± 0.49 × 10–3 mm ² /s, p  < 0.0001), and ADC_1000 values (0.48 ± 0.38 vs. 1.23 ± 0.23 × 10–3 mm ² /s, p  < 0.0001). In conclusion, the demonstrated inverse correlation between LIC and liver ADC values, accentuated at higher diffusion b ‐values, has relevant clinical implications, indicating that liver iron overload acts as a confounding factor in quantitative DWI and should be accounted for to avoid misinterpretation of liver ADC measurements in diagnostic and therapeutic workflows.