We report the experience of our general hospital in selecting the patients for orthotopic liver transplantation (OLT). Fifty-one patients with cirrhosis were examined and 20 of them submitted to OLT from August, 1992, to November, 1995. For liver studies, the 20 transplant recipients were examined with US and plain and dynamic CT; 15/20 were submitted to CTAP, 10/20 to Lipiodol CT and 17/20 to angiography. The accuracy of these techniques in HCC detection was assessed by correlation with resected whole livers. The accuracy of duplex Doppler and color flow Doppler for portal and/or mesenteric vein thrombosis was evaluated by correlation with resected livers, CT and angiographic findings. Pathologic examinations diagnosed HCC in 5/20 transplant recipients: 2 lesions (1.5 cm and 2 cm; 2 cm and 3.5 cm) were found in 2 resected specimens (total hepatectomy) and 1 lesion was found in 3 cases (2.5 cm, 1.5 cm, 1 cm). The sensitivity of US, plain and dynamic CT in identifying HCC patients was 20%; US and CT specificity rates were 100% and 87%, respectively. CTAP sensitivity was 75% and the sensitivity of Lipiodol CT and angiography was 100%. Therefore, in our series, US was poorly sensitive in the detection of liver cancers, which may depend on the small number of patients, lesion size (< or = 3.5 cm) and the radiologists ignoring clinical and laboratory data on purpose. Nevertheless, the patients with a single HCC not exceeding 5 cm phi or with no more than 3 tumors, none of them exceeding 3 cm phi, are generally considered eligible for transplantation: therefore, our patients chosen for OLT on the basis of US and CT findings were actually eligible for transplantation in spite of US and CT false negative results. At US, the portal vein had an average caliber of 13.5 +/- 2.5 mm in 21/51 patients; the average caliber of the common hepatic artery was 6 +/- 1.5 mm in 49/51 patients; average spleen length was 174 +/- 38 mm. US showed ascites in 28/51 cases. In conclusion, considering also the long stand-by list for OLT, the first selection of transplant candidates could be performed with US and color flow Doppler, plain and dynamic CT. The patients who are ruled out as candidates for OLT on the basis of the findings of these imaging techniques and of clinical and laboratory findings are submitted to no further examination and referred to the transplantation unit. Otherwise, if conventional and color flow Doppler US and conventional CT are not enough to exclude a patient from OLT, the subject is submitted to more invasive (angiography, CTAP, Lipiodol CT) or less widespread (spiral CT, MRI) techniques.