ABSTRACT
Thirty-eight patients (76 legs) with suspected deep vein thrombosis were examined with 99Tcm-fibrinogen study (Tc-F), which consists of three parts: (1) radionuclide venography (RNV) of both legs after injection of 150 MBq of 99Tcm-fibrinogen into the dorsal veins of each foot, (2) blood pool imaging (BPI) at 5 to 20 min and (3) uptake imaging (UI) at 4 to 6 and 17 to 26 h. The clinical diagnosis was deep vein thrombosis in 27 legs. Twenty-nine patients (with 18 thrombi) were also studied with contrast venography (CV). Comparison of RNV with BPI suggests that the time-consuming and laborious RNV could, without affecting the diagnosis, be substituted by BPI, which is possible to perform after a single antecubital injection. When all patients were taken into consideration the sensitivity of Tc-F to detect fresh thrombosis was 100% but the specificity was lower (73%) because of the false positive findings in RNV/BPI. However, the specificity of UI was 100%. Previous anticoagulation did not prevent the uptake of fibrinogen within thrombi. There was good agreement of RNV/BPI and CV in the popliteal, femoral and iliac veins but the calf veins were better delineated in CV. Seventeen of the 18 thrombi examined both with Tc-F and CV were positive with RNV/BPI, 14 with UI and 13 with CV. Sensitivity, specificity, accuracy, positive predictive value and negative predictive value (%) were with RNV/BPI: 83, 73, 86, 90, 83, with UI: 78, 100, 86, 100, 74 and with CV: 73, 100, 83, 100, 69. Our results suggest that Tc-F could be used not only as a screening test but also instead of CV in the detection of venous thrombosis.
