We used the serial intravascular ultrasound (IVUS) data from the Serial Ultrasound REstenosis trial to explain why positive remodeling lesions have a higher rate of clinical restenosis after non-stent interventions. Serial IVUS was performed before intervention and immediately and 1 and 6 months after percutaneous transluminal coronary angioplasty (n = 35) or directional coronary atherectomy (n = 26). External elastic membrane, lumen, and plaque + media (external elastic membrane minus lumen) areas were measured at the reference and stenosis. Stenoses were divided into 3 groups: positive remodeling (lesion greater than proximal reference external elastic membrane), intermediate remodeling (lesion external elastic membrane smaller than proximal reference but larger than distal reference), and negative remodeling (lesion equal to or less than distal reference external elastic membrane). The early (postintervention to 1 month) and late (1- to 6-month) changes in lesion external elastic membrane and plaque + media areas were compared. An early increase in plaque + media area was associated with an equal or greater increase in external elastic membrane area in positive (r = 0.78, p <0.0001), intermediate (r = 0.69, p <0.0001), and negative (r = 0.59, p = 0.0003) remodeling lesions. A late (1- to 6-month) decrease in external elastic membrane area correlated inversely with the early increase in plaque + media area in positive (r = 0.77, p = 0.0002) and intermediate (r = 0.45, p = 0.0003), but not in negative (r = 0.02, p = 0.9) remodeling lesions. In positive remodeling lesions, the early increase in plaque + media area was associated with both an exaggerated early increase and late decrease in external elastic membrane area. Positive remodeling lesions have an exaggerated early increase in external elastic membrane area and, especially, an exaggerated late decrease in external elastic membrane area after percutaneous transluminal coronary angioplasty and directional coronary atherectomy. This may explain that the increased clinical restenosis after positive remodeling lesions is treated with non-stent interventions.