Objectives: a) Describe hepatocellular semiology in magnetic resonance imaging and lipiodol computerized tomography in patients with cirrhosis, who are candidates for surgery; b) Clarify the respective roles of magnetic resonance imaging and lipiodol computerized tomography in hepatocellular detection.
Methods: Twenty four patients with suspected hepatocellular carcinoma underwent successive magnetic resonance imaging and lipiodol computerized tomography. Thirty-four of the 67 lesions seen by lipiodol computerized tomography and 28 of 52 lesions seen by magnetic resonance imaging were confirmed histologically.
Results: In lipiodol computerized tomography, 44% of hepatocellular carcinomas had a dense and homogeneous pattern; 24% had a homogeneous but slightly dense pattern. Sixteen distinct deposits were described: 4 were confirmed as hepatocellular carcinoma and 12 were not controlled histologically. In magnetic resonance imaging 57% of hepatocellular carcinomas have a high intensity on T1 and T2 weighted spin echo images, 38% were hyperintense on T2 and hypo or isointense on T1 weighted images. Eighty-six percent of hyperintense T1 and T2 weighted images were hepatocellular carcinoma. When the gold standard was histology, lipiodol computerized tomography sensitivity (81%) was higher than magnetic resonance imaging (68%). When the gold standard was lipiodol computerized tomography, the sensitivity of magnetic resonance imaging was 47 +/- 12%.
Conclusions: a) The sensitivity of lipiodol computerized tomography was better than resonance magnetic imaging; b) the homogeneous and slightly dense pattern corresponded to a hepatocellular carcinoma in 50% of cases; c) on magnetic resonance imaging any lesions with high intensity on T1 and T2 spin echo images strongly suggests hepatocellular carcinoma; d) if surgical resection after ultrasonography is being considered, the second step should be an magnetic resonance imaging.