Evaluation of myocardial iron during iron chelation therapy is not feasible by repeated endomyocardial biopsies owing to the heterogeneity of iron distribution and the risk of complications. Recently, we described a noninvasive method based on magnetic resonance imaging. Here, the method was used for repeated estimation of the myocardial iron content during iron chelation with deferrioxamine in 14 adult nonthalassemic patients with transfusional iron overload. We investigated the repeatability of the method and the relationship between the myocardial iron estimates and iron status. The repeatability coefficient (2sD) was 2.8 micromol/g in the controls (day-to-day) and 4.0 micromol/g in the patients (within-day). Myocardial iron estimates were elevated in 10 of all 14 patients at first examination, but normalized in 6 patients after 6 to 18 months of treatment. If liver iron declined below 350 micromol/g all but one of the myocardial iron estimates were normal or nearly normal. At start (R2 = 0.69, P =.0014) and still after 6 months of iron chelation (R2 = 0.76, P =.001), the estimates were significantly and more closely related to the urinary iron excretion than to liver iron or serum ferritin levels. In conclusion, our preliminary data, which may only pertain to patients with acquired anemias, suggest the existence of a critical liver iron concentration, above which elevated myocardial iron is present, but its extent seems related to the size of the chelatable iron pool, as reflected by the urinary iron excretion. This further supports the concept of the labile iron pool as the compartment directly involved in transfusional iron toxicity.