In an experimental animal model of femoral artery thrombosis, contrast angiography was compared to intravascular angioscopy. Additionally, the effect of mechanical, rotational thrombectomy and the additive benefit of the administration of intravascular streptokinase were assessed by means of both procedures. After external forceps crush injury alone, contrast angiograms were generally normal (6 of 14) or showed minimal luminal irregularity (3 of 14), and 5 of 14 had 30% to 50% stenosis. With angioscopy, none appeared normal, and 14 of 14 showed thrombi layered along the wall, as well as intimal flaps, and 6 of 14 had partially occlusive thrombi (p less than 0.001 angiography vs angioscopy). After 2-hour occlusion and injection of thrombin into the injured segment, angiographic total (5 of 14), subtotal (3 of 14), or partial thrombotic occlusions (5 of 14) were created. Angioscopy showed similar results, except that total occlusions were classed as subtotal occlusions. After rotational thrombectomy, most arteries again appeared normal by contrast angiography (6 of 11) but none were angioscopically normal (p less than 0.006). Streptokinase, administered after rotational thrombectomy in seven arteries, normalized one 30% angiographic stenosis; there were no other angiographic changes. Findings with angioscopy were also unchanged. We conclude that in the diagnosis and treatment of intravascular thrombosis, angioscopy is generally more sensitive in the detection of intravascular thrombi, with the exception of total thrombotic occlusions. Angioscopy was uniquely effective in identifying subintimal flaps, which were never identified by angiography. In this model, streptokinase provided little or no additional thrombolytic benefit to mechanical thrombectomy alone.