Cavernous hemangiomas in patients with chronic liver disease: MR imaging findings

The purpose of our study was to assess the difference in magnetic resonance imaging (MRI) features of cavernous hemangiomas in patients with chronic liver disease compared them with hemangiomas in normal livers. We retrospectively searched our records of MRI of the liver between October 1998 and June 2002, and identified 76 hemangiomas in 49 patients (18 men and 31 women; age range 29-81 years [mean, 57 years]). Hemangiomas were classified into 3 groups: patients with cirrhosis [group 1, 8 lesions in 8 patients], patients with chronic hepatitis [group 2, 6 lesions in 5 patients], and patients without underlying liver disease [group 3, 62 lesions in 36 patients]. Four radiologists, blinded to clinical information, retrospectively reviewed in consensus the MRI findings of hemangiomas for number, size, signal intensities on T1- and T2-weighted images, and enhancement patterns on early- and late-phase postcontrast images. The mean lesion numbers and sizes were 1.0 and 16.2 +/- 9.6 mm, 1.2 and 15.3 +/- 7.1 mm, and 1.7 and 26.1 +/- 24.7 mm in groups 1-3, respectively. There was a correlation (p < 0.05, coefficient: 0.35) between lesion number and severity of liver disease. Although there was no significant difference in lesion size among the 3 groups, all of 11 lesions larger than 4 cm in diameter belonged to group 3. Almost all lesions appeared moderately hypointense on T1-weighted images and moderately hyperintense on T2-weighted images. Twenty-seven lesions showed immediate homogeneous enhancement (pattern 1), and 49 showed peripheral nodular enhancement with centripetal enhancement progression (pattern 2). There was no difference in frequency of enhancement patterns among the 3 groups. Hemangiomas were more often solitary in livers with chronic liver disease, large lesions were exclusively seen in livers without chronic liver disease, and there was a trend for small lesions in patients with chronic liver disease. (C) 2004 Elsevier Inc. All rights reserved.