BACKGROUND Discordant results have been reported regarding morphological predictors of restenosis after percutaneous transluminal coronary angioplasty (PTCA). These discrepancies may be related to the limitations of angiography in the study of plaque morphology. RESULTS We studied 117 consecutive patients who underwent successful PTCA and who underwent coronary angioscopy before and immediately after the procedure. Angiographic follow-up was performed in 99 (85%) patients. We analyzed the relationship between angioscopic variables at the time of PTCA and the occurrence of restenosis assessed by quantitative coronary angiography. Plaque shape and color had no effect on late loss in luminal diameter (late loss: smooth lesions, 0.55 +/- 0.68 mm; complex lesions, 0.76 +/- 0.60 mm; white plaques, 0.51 +/- 0.56 mm; yellow plaques, 0.65 +/- 0.72 mm; P = NS). An angioscopic protruding thrombus at the PTCA site was associated with significantly greater loss in luminal diameter (late loss: no thrombus, 0.47 +/- 0.54 mm; lining thrombus, 0.59 +/- 0.67 mm; protruding thrombus, 1.07 +/- 0.77 mm; P < .05). Dissection assessed by angioscopy immediately after PTCA had no effect on late loss in luminal diameter (late loss: no dissection, 0.60 +/- 0.60 mm; simple dissection, 0.82 +/- 0.75 mm; complex dissection, 0.57 +/- 0.80 mm; P = NS). CONCLUSIONS These results show that coronary angioscopy may be helpful in predicting the risk of restenosis after PTCA. The high rate of angiographic recurrence observed when PTCA is performed at thrombus-containing lesions supports a role for thrombus in the process of luminal renarrowing after PTCA.