Practical Implications of Non-Invasive MRI-Based Liver and Cardiac Iron Assessments in Children and Adults with Transfusion-Dependent Anemias.

BACKGROUND: MRI measures of hepatic iron content (HIC) in patients with transfusional iron overload correlate well with invasive liver biopsy- based HIC measures. HIC is the major component of, and a good surrogate for total body iron burden, and has been used worldwide as an index for chelation management [Olivieri&Brittenham, Blood 1997]. However, cardiac iron toxicity, not liver iron per se, causes the most mortality in transfusional iron overload, particularly in thalassemia major (TM). The objective of this study was to evaluate in adults and children with transfusional iron overload, for whom these measures were obtained by MRI as standard clinical care, the relationship between cardiac and liver iron assessments and how the combined information might influence clinical practice decisions.

METHODS: IRB approved, single-center retrospective chart review of patients age 2–45y with transfusional iron overload who had MRI liver and heart iron assessments from Jan 02 to Mar 07. Sixty-six patients were evaluated. This was a cross-sectional analysis of one MRI per patient. Iron in myocardium and liver was measured by R2* (1/T2*), an MR relaxation parameter which varies proportionally with tissue iron concentration, using a 1.5 T scanner. Myocardial T2* was assessed from one midpapillary ventricular short-axis slice using a cardiac-gated, segmented, multiecho gradient sequence obtained in a single breath-hold, similar to the method of Westwood et al [J Magn Reson Imaging, 2003]. Left Ventricular Ejection Fraction (LVEF) was measured by standard cardiac MRI methods. We express cardiac iron status as T2* values (normal >20 msec), and liver iron values as predicted HIC (mg/g dry wt) based on standard curves from [Wood et al, Circulation 2005].

RESULTS: Among the 34 pediatric patients (65% male), diagnoses were: TM (n= 17), transfused sickle cell disease (SS) (n= 8), and marrow failure (MF) (n=9). Mean age was 14.3 ± 4.4 y (range 5–20). Adult patients (n=32, 47% male), mean age 33.5 ± 12.3 y (range 21–78) had these diagnoses: TM (n= 21), thal intermedia (n= 7), 4-gene alpha thal (n= 1), MF or MDS (n=3). Cardiac T2* and the MRI-based HIC are correlated in neither age cohort (adults: Spearman coeff. ρ 0.265, p = 0.14; children ρ-0.061, p=0.73). We identified a subset of subjects (2/34 children (6%), 7/32 adults (23%)) with severely low cardiac T2* but HIC within target range for therapy (predicted HIC <7 mg/g dry wt), which would be missed by liver assessment alone. LVEF correlated with the cardiac T2* in adults (ρ0.412, p = .02) but not in children (ρ.277, p = .12). There is a cluster in both age cohorts of TM patients with ‘normal’ EF (>60%) but abnormal T2*, at risk for cardiac disease. We observed a significant difference in LVEF in children with TM vs. SS (59.5 ± 3.9 vs. 64.3±3.6%, p=0.008), and also a less significant difference between these groups in mean T2* measures (TM 26.11±3.55, SS 34.76±2.12 msec p = 0.1136).

CONCLUSION: We demonstrate that in children and adults with transfusional iron overload (and non-transfused thal intermedia), combined T2* cardiac and R2* hepatic iron determination detects a population with severe cardiac iron deposition without matched increased HIC. As MRI assessments of cardiac and liver iron become prevalent, prospective studies will be needed to understand the best ways to use these data for management.