The relationship between cardiac and liver iron evaluated by MR imaging in haematological malignancies and chronic liver disease

Abstract

Although iron overload is clinically significant, only limited data have been published on iron overload in haematological diseases. We investigated cardiac and liver iron accumulation by magnetic resonance imaging (MRI) in a cohort of 87 subjects who did not receive chelation, including 59 haematological patients. M-HIC (MRI-based hepatic iron concentration, normal values <36 μmol/g) is a non-invasive, liver biopsy-calibrated method to analyse iron concentration. This method, calibrated to R2 (transverse relaxation rate), was used as a reference standard (M-HIC(R2)). Transfusions and ferritin were evaluated. Mean M-HIC(R2) and cardiac R(*) of all patients were 142 μmol/g (95% CI, 114-170) and 36.4 1/s (95% CI, 34.2-38.5), respectively. M-HIC(R2) was higher in haematological patients than in patients with chronic liver disease or normal controls (P<0.001). Clearly elevated cardiac R2(*) was found in two myelodysplastic syndrome (MDS) patients with severe liver iron overload. A poor correlation was found between liver and cardiac iron (n=82, r=0.322, P=0.003), in contrast to a stronger correlation in MDS (n=7, r=0.905, P=0.005). In addition to transfusions, MDS seemed to be an independent factor in iron accumulation. In conclusion, the risk for cardiac iron overload in haematological diseases other than MDS is very low, despite the frequently found liver iron overload.

Figure 1

Cardiac short-axis left ventricular MR images of three patients with similar parameters (TE=8 ms). The images illustrate the different degrees of hepatic and cardiac iron accumulation, which is seen as a signal drop. A signal intensity of myocardium was measured with constant region of interest (oval line) from MR images in each patient. Subphrenic liver, marked with an asterisk, is situated at the lower part of the image. (a) A patient with ALL; normal level of cardiac R2^* (30 1/s) with only slightly elevated hepatic iron concentration (40 μmol/g). There is no visible signal drop found in the liver or the heart. (b) A patient with AML; cardiac R2^* (37 1/s) was normal, although there was heavy hepatic iron overload (380 μmol/g). The liver signal dropped owing to the severe iron overload, giving more contrast to the normal cardiac tissue. (c) A patient with MDS; pathologically low R2^* (85 1/s), indicating cardiac iron overload with simultaneous severe hepatic iron overload (488 μmol/g). The figure shows the signal drop of both cardiac and hepatic tissue due to excess iron.

Figure 2

A correlation between cardiac R2^* and hepatic iron concentration. (a) In MDS, the cardiac iron (R2^*) increases as hepatic iron concentrations increase, showing a strong dependence. Two of the MDS patients had the highest R2^* values among the entire cohort. (b) All subjects (n=82) showed a weak dependence between cardiac and hepatic iron concentrations.

Figure 3

In haematological patients, both transfusion load (RBC) and ferritin can be used to estimate iron concentrations (HICest). The number of RBC units was more accurate, with better correlations in estimating liver iron concentration with linear fit. (a) The correlation between hepatic iron concentration and RBC (range 0–144) is demonstrated with the equation HICest(transfusion)=4.6RBC+2.1. (b) The correlation between hepatic iron concentration and ferritin is demonstrated with the equation HICest(ferritin)=0.12ferritin−77.