OBJECTIVE Radiolabelled anti-fibrin antibodies have not yet enabled reliable and practical diagnosis of venous thromboembolism. However, previous unsuccessful clinical trials were performed with anti-fibrin beta-chain antibodies that do not optimally bind to thrombi during anticoagulation. The current experiments were performed to determine if radiolabelled anti-D-dimer antibodies more reliably allowed nuclear medicine imaging of deep venous thrombi during anticoagulation than anti-beta-chain antibodies. METHODS Dogs with pre-existing unilateral femoral vein thrombi were anticoagulated with heparin (300 units.kg (-1) bolus followed by 90 units.kg(-1).h(-1) continuous infusion). They were then allocated to receive one of three (111)In labelled antibodies: anti-D-dimer, anti-beta or a non-specific control antibody. Gamma scans of the legs were performed at regular intervals for 24 h. Scans were interpreted in a blinded fashion and the number of gamma counts from the femoral area on the thrombosed side was compared to the contralateral side. Clot/blood isotope density ratios were determined post-mortem. RESULTS Leg thrombi were 100% detectable in the anti-D-dimer group, but only 60% detectable in the anti-beta group. Mean +/- SD relative counts in the thrombosed femoral area was 137 +/- 10% compared to the contralateral side in the anti-D-dimer group, but only 116 +/- 20% in the anti-beta group. The clot/blood ratio was 24.5 +/- 2.8 in the anti-D-dimer group, but only 7.8 +/- 2.0 in the anti-beta group. CONCLUSIONS In labelled anti-D-dimer provides superior nuclear medicine images of thrombi during intensive anticoagulation compared to anti-beta. Clinical results with radiolabelled anti-D-dimer may be more promising than those previously observed with other anti-fibrin antibodies.