[The scope and limitations of cavography in tumour diagnosis (author's transl)].

Fourteen cavograms showing abnormal findings were selected from a total of 500 examinations; in these the diagnosis had been confirmed surgically, histologically, at autopsy or by the clinical course. The aim of selecting this material was to determine retrospectively the accuracy to cavography in localising retroperitoneal masses, to determine their extent, whether it is possible to diagnose the type of tumour and whether the presence of thrombi can be regarded as a sign of invasion of the cava by malignant tumour. The abnormal cavograms were evaluated with respect to certain radiological signs and the diagnostic value of individual signs was calculated statistically. It was found that the only fairly reliable sign of tumour thrombus consists of an ill-defined impression with breaks in the continuity of the vascular margin. All other radiological signs examined were found to be non-specific.