Non-Invasive Assessment of Pancreatic Iron Stores by Magnetic Resonance Imaging (MRI) in β-Thalassemic Patients.

Frequent transfusions of red blood cells are considered standard therapy for patients with β-thalassemia. However, this can lead to transfusional iron overload and subsequent end-organ damage with decrease in life-expectancy. Ferritin is the most widely available non-invasive method for assessing iron stores and iron overload in chronically transfused patients. However, it can also be elevated in inflammatory conditions. MRI has been proposed as a non-invasive method for detection and quantification of iron stores in specific organs. Most studies utilizing MRI for detection of iron overload have focused on the heart and liver, and it is unknown if MRI could satisfactory detect iron overload in other potentially involved organs such as pancreas.

To evaluate and correlate the level of iron accumulation in different organs and serum ferritin concentrations for 6 months before imaging studies in patients with β-thalassemia receiving chronic transfusion therapy.

MRI was used to asses iron content in three different organs (heart, liver, and pancreas) in patients with a diagnosis of β-thalassemia. Validation of the MRI technique was done by determining liver iron concentration (LIC) from 11 liver biopsies. LIC was determined by atomic absorption spectrometry and was correlated with liver T2* measurement obtained with MRI. There was a significant, curvilinear, inverse correlation between liver T2* MRI measurements and the LIC by Pearson′s method (r =-0.878, p=0.001). We used Pearson′s coefficient of correlation to assess association between T2* measurements among different organs (heart, liver and pancreas) and between organs and serum ferritin levels.

We evaluated 115 patients with a diagnosis of β-thalassemia that were receiving chronic transfusion therapy. Mean age was 21,25 years (range 7-54 years) and 43% were male. Mean T2* value in the liver was 3.91 ± 3.95 ms, indicating significant liver siderosis (T2*<6.3ms) in most patients (92.1%). Mean value of myocardial T2* was 24.96 ± 14.17 ms and the incidence of cardiac siderosis (T2*<20ms) was 36%. Additionaly, 19% of the patients (22/115) had severe cardiac siderosis (T2* <10ms). Mean T2* value in pancreas was 11.12 ± 11.20 ms, and pancreatic iron deposition (T2* < 21ms) was found in 83.5% of patients. There was no significant correlation between liver and pancreas iron overload (r =0.249), and liver and myocardial iron overload (r =0.149). There was a moderate correlation among pancreas and myocardial iron overload (r =0.546; p=0.001). Mean serum ferritin level was 2,676.5 +/- 2,051.7 ng/mL (range 59-12,362 ng/mL). There was no significant correlation among ferritin serum level and liver, heart and pancreas T2* values (r =-0.397; r =-0.220; r =-0.295).

Iron overload of liver, heart and pancreas, measured by MRI T2*, could not be predicted by ferritin levels in patients with β-thalassemia. Pancreatic iron overload can be measured by MRI, but we could not predict pancreatic hemosiderosis by detection of iron overload in others organs (except for a moderate correlation among pancreas and heart iron overload). Given that direct calibration of MRI with pancreas biopsies is not possible, further studies are necessary to validate this technique.

No relevant conflicts of interest to declare.