To assess the value of MR imaging signal changes of pancreas in predicting pancreas iron overload and impairment of the pancreatic function in secondary hemochromatosis.

Twenty patients with transfusion-dependent beta-thalassemia major were imaged at .5-T, with SE T1-weighted, FSE T2-weighted, and GRE T2*-weighted sequences. Image analysis was performed to assess pancreas-to-fat signal intensity ratios for all pulse sequences. Pancreatic exocrine and endocrine function were assessed. Twenty healthy volunteers underwent MR imaging with the same three sequences and served as a control group. Statistical analysis was performed using a 1-way ANOVA test to assess differences in MR pancreas-to-fat signal intensity ratios. A ROC analysis was performed to assess the pancreas-to-fat signal intensity threshold value in predicting pancreas iron overload for all pulse sequences.

The pancreas-to-fat signal intensity ratio was significantly decreased in all patients (100%) on FSE T2-weighted images (p < .0001), and GRE T2*-weighted images (p < .0001). On SE T1-weighted images pancreas-to-fat signal intensity ratio was decreased only in 17 of the 20 patients (85%, p < .0001). A significant correlation was found between increased pancreas-to-fat signal intensity ratios and decreased serum trypsin levels for SE T1-weighted sequences (r = -.77, p < .01).
At ROC analysis the GRE T2*-weighted sequences showed the best sensitivity and specificity to asses pancreas iron overload with a pancreas-to-fat signal intensity threshold level of 54.6 (sensitivity 82.4%, specificity 100 %, accuracy 91.2%).

The pancreas-to-fat signal intensity ratio measured on GRE T2*-weighted images is a valuable predictor of the degree of pancreas iron overload in secondary hemochromatosis. Decreased serum trypsin levels correlates with increased pancreas-to-fat signal intensity ratios on SE T1-weighted images suggesting the hypothesis of fatty replacement of the damaged parenchyma.