Thrombus removal using percutaneous rotational thrombectomy (PRT), followed by tissue plasminogen activator (t-PA), was studied by contrast angiography and fiberoptic angioscopy in a canine femoral artery model of thrombosis. After thrombus induction and following each treatment, comparisons were made between angioscopy and angiography for the detection of thrombus and subintimal dissection. Angioscopic images were scored in a blinded fashion for lining, protruding, or occlusive thrombus (class 1,2, or 3) as well as estimated wall coverage by thrombus. Angiograms were studied for percent diameter stenosis and the presence of flaps. Following external forceps crush injury of 18 arteries, two hour occlusion, and injection of thrombin, mean angiographic stenosis was 66%, thrombus coverage by angioscopy was 81%, and mean angioscopy class was 2.5. Following PRT, stenosis decreased to 27% (p less than 0.008), thrombus coverage was reduced to 49% (p less than 0.02), and angioscopy class dropped to 2.0 (p less than 0.07). After t-PA treatment, these values were further reduced to 25% (p = NS), 26% (p less than 0.02), and 1.3 (p less than 0.008), respectively. In comparison to angiography, subintimal dissection (seen as flaps) and thrombus (lining, protruding, or occlusive) were present significantly more often by angioscopy (p less than 0.001). It is concluded that PRT results in significant thrombolysis, apparent by angiography and angioscopy. Follow-up t-PA can produce additional, incremental thrombolysis, apparent only by angioscopy. A beneficial role for t-PA following mechanical thrombolysis is suggested by this model. The superior sensitivity of angioscopy for detection of flaps and thrombus is underscored by this study.