This study was undertaken to evaluate ultrasonography (US) and computed tomography (CT) in the diagnostics of portomesenteric vein thrombosis (PMT). Between 1972-1988 we found 40 cases with PMT, all verified in angiography and/or autopsy. Between 1972-79 US and CT were never used, when 19 cases with PMT were found (1 premortem diagnosis with splenoportography). Between 1980-88 there were 21 patients with PMT. US was performed in 13, detecting PMT in 2. CT was performed in 5 patients confirming US diagnosis in 2 cases and revealing one further PMT. One PMT was diagnosed in the venous phase of an arteriography. The sensitivities of US and CT to detect PMT were 15% (2/13) and 60% (3/5) (p less than 0.05). The delay until diagnosis decreased from 27 +/- 28 days in 1972-79 to 12 +/- 9 days in 1980-88 (p less than 0.02). However, neither the rate of correct antemortem diagnosis (5% vs 19%) nor the survivals (68% vs 65%, 95% vs 81%, 95% vs 100% at 1, 12 and 60 months) were improved. Because only the patients with correct diagnosis received appropriate therapy, followed by longest survivals, the importance of the diagnostic attempts should not be questioned. CT may be more sensitive than US for detecting PMT, but neither method have yet replaced angiographic techniques.