Background: Previous reports have described the computed tomographic (
CT) appearance of collateral veins following splenic vein occlusion (S
VO). This retrospective study was performed to determine the etiology,
clinical manifestations, and accuracy of CT diagnosis in patients wit
h this entity. Methods: A computer search of radiology reports for a 1
-year period found 52 patients with SVO diagnosed by absence of visual
ization of the splenic vein accompanied by the formation of the expect
ed perigastric collateral veins. Clinical data were reviewed for seque
la of SVO and clinical impact of the diagnosis. Results: In 12 cases,
other studies confirmed the CT diagnosis of SVO. In no case was the CT
diagnosis proved to be incorrect by other imaging studies. Angiograph
ic records found five additional cases with SVO not diagnosed by CT, b
ut two of five had convincing CT evidence of SVO noted upon reevaluati
on by the authors. Review of clinical data showed heme-positive stool
in six, of which three had significant gastrointestinal hemorrhage. Sp
lenic infarction occurred in two cases. Conclusions: Our data indicate
that SVO is more common than previously suspected and usually remains
clinically silent, but CT appears to be highly specific and fairly se
nsitive for its diagnosis.