The need for accurate and noninvasive evaluation of liver iron stores prompted us to evaluate the reliability of high-field magnetic resonance imaging equipment in liver patients with low or moderate siderosis, given the poor results obtained using systems operating at low field strength in such cases. Twenty patients with sporadic porphyria cutanea tarda and 28 with comparable chronic liver diseases (chronic hepatitis or cirrhosis) and moderate siderosis were compared with 10 patients with idiopathic or secondary hemochromatosis and 10 healthy controls. Plasma iron profile, ferritin concentration and liver iron concentration, determined with atomic absorption spectroscopy, were matched with the magnetic resonance parameters - namely, transverse relaxation time and the signal intensity for a given proton amount, obtained with equipment operating at a field strength of 1.5 T. Hemochromatosis patients with mean liver iron concentrations of 550 μmol/gm dry wt (vs. 10 μmol of controls) exhibited an impressive reduction in the signal intensity with respect to the other three groups, and this reduction prevented any further comparison with the same porphyria cutanea tarda and chronic liver disease groups, whose liver iron level was twice that of the controls. The signal intensity remained almost unchanged in the latter groups, whereas the transverse relaxation time was significantly reduced. Moreover, correlation with liver iron was significantly inverse in the case of the transverse relaxation time (n = 17, r = 0.62, p = 0.008) and direct in the case of the transverse relaxation rate. The transverse relaxation time values returned to normal in five patients who had completed an iron-depletion program. No significant relationships between porphyrinuria (in porphyria cutanea tarda patients) or associated histological lesions (steatosis or inflammation) and the signal intensity or the transverse relaxation time were observed. We conclude that a magnetic resonance imaging system operating at higher field strength (with respect to previous studies) represents a useful, non-invasive diagnostic tool for the evaluation of even modest liver iron overload. Under these conditions, the decrease in the transverse relaxation time is proportional to the decrease in iron. Hemochromatosis cannot be investigated under the same conditions. (HEPATOLOGY 1993; 17:997–1002.)