Diagnostic laparotomy is no longer routinely performed in Hodgkin's lymphoma and noninvasive diagnosis of spleen involvement remains uncertain. In order to assess the probability of splenic involvement based on clinical parameters, we retrospectively analyzed data on patients of the German Hodgkin's Lymphoma Study Group (GHSG) who underwent staging laparotomy and for whom splenic weight and size were available. Our study included 376 patients with Hodgkin's lymphoma who underwent staging laparotomy and splenectomy according to the treatment policy of the GHSG between February 1981 and January 1993. Univariate and multivariate analyses of pretherapeutic clinical characteristics and splenic weight were performed in order to predict the probability of splenic involvement. Computed tomographic (CT) images of 25 patients were available and used to correlate radiological splenic size and pathological splenic weight. In 171 of 376 patients spleen involvement was found. Average weight of the spleens was 258 g (+/-257) ranging from 55 to 3290 g. All spleens with a weight above 2000 g showed disease involvement, while those under 150 g were never involved. In the multivariate analysis, splenic weight ( p<0.001), erythrocyte sedimentation rate ( p<0.001), and clinical stage ( p<0.01) were found to be independently prognostic for spleen involvement. Splenic weight was highly correlated with a spleen index defined as the product of length, width, and thickness measured by CT (correlation coefficient: 0.93). By applying the identified risk factors in clinically staged patients spleen involvement can be determined. Spleen weight can be estimated with the help of a spleen index. Above an index of 1000 the probability of spleen involvement is higher than 90%. This might be of outstanding importance for patients being scheduled for involved field radiation.