In the Washington Radiation for In-Stent restenosis Trial (WRIST), patients were first treated with conventional techniques and then randomized to either gamma-irradiation ((192)Ir) or placebo (dummy seeds). In the (192)Ir group with native coronary in-stent restenosis, we identified 8 patients with edge recurrence and compared them with 21 patients with no recurrence. Serial (postirradiation and follow-up) intravascular ultrasound analysis was performed according to conventional methods. When compared with nonrecurring lesions, lesions with distal edge recurrence had (1) greater decrease in mean distal lumen cross-sectional area (-3.0 +/- 1.2 vs -0.7 +/- 1.0 mm(2), p = 0.0002), (2) no change in mean distal external elastic membrane cross-sectional area versus an increase in mean distal cross-sectional area of 1.0 +/- 0.9 mm(2) in nonrecurring lesions (p = 0.0047), and (3) a greater increase in mean distal plaque + media cross-sectional area (2.9 +/- 1.2 mm vs 1.7 +/- 0.6 mm(2), p = 0.0103). Within the stented segment, the nonrecurring lesions had no decrease in mean lumen and no increase in mean intimal hyperplasia cross-sectional area. Conversely, lesions with distal edge recurrence had a significant decrease in mean intrastent lumen cross-sectional area (-1.7 +/- 1.7 mm(2)) and a significant increase in mean intrastent intimal hyperplasia cross-sectional area (1.6 +/- 1.6 mm(2)). Lesions with distal edge recurrence also had a greater decrease in mean proximal lumen cross-sectional area (-1.7 +/- 1.3 vs -0.3 +/- 0.8 mm(2), p = 0.0213), with a trend toward a greater increase in mean proximal plaque + media cross-sectional area. Thus, edge recurrence after (192)Ir treatment of in-stent restenosis is the result of neointimal hyperplasia (part of generalized treatment failure) and the absence of radiation-induced positive remodeling.