We read with great interest the case report by Udell and colleagues 1 in which they allude to the difficulties and pitfalls in making the diagnosis of neonatal hemochromatosis (NH) in 2 of their patients. The report demonstrates once more how much there still is to learn about this probable alloimmune disorder. 2,3 Establishing the diagnosis of NH remains a challenge, not least because biochemical iron studies in suspected patients only provide limited information. Demonstration of abnormal extrahepatic iron deposition is essential, and modern magnetic resonance (MR) imaging technology has a particular appeal for this.
However, little evidence has been gathered in this field, and to date, only a handful of cases with MR images have been published. [4][5][6][7][8] Despite this, the technique is advocated regularly, and there seems to be reasonable consensus as to what the necessary diagnostic requirements are. On T 2 -weighted sequences, tissues with increased iron content have low signal intensity. In NH, the reticuloendothelial system is spared so that the spleen retains a normal, higher signal intensity compared with affected tissues. Skeletal muscle is used by some authors as a reference for tissue signal intensity because tissues with increased iron content have a lower signal intensity compared with skeletal muscle. Furthermore, the signal intensity of the pancreas in these case reports varies from slightly lower than that of the spleen, to low signal intensity comparable with the liver, to very low signal (black). It is also important to be aware that siderosis or increased iron content of the liver is physiological in the third trimester of pregnancy and in the neonatal period, so that the liver will have a lower signal intensity than the spleen or skeletal muscle on T 2 -weighted sequences even in normal neonates. 9,10 The authors unfortunately do not make this vital point in their report and have highlighted low signal intensity of the liver as diagnostic of NH. The MR scan of patient 1 does not convincingly show iron deposition in the pancreas, which has slightly lower signal intensity than the spleen, but not as low as the liver or skeletal muscle. In patient 2, the pancreas has not been demonstrated for comparison. Therefore, although the buccal mucosal biopsy findings support a diagnosis of NH, the MR scans in both patients appear to be nondiagnostic. More information on this new technique is needed, particularly to establish the sensitivity and specificity of such MR imaging. We hope that in the future, pediatric radiologists and hepatologists feel encouraged to pool experience and perhaps even contribute to some form of registry for the MR imaging in this rare pediatric disease.