Quantitative coronary angiography (QCA) is routinely used for assessment of strategies aimed at reducing in-stent restenosis. Yet QCA enables only the measurement of luminal variation of stented segments and, unlike intravascular ultrasound (IVUS), provides only an indirect estimation of late in-stent neointimal formation, which has a key role in the process of in-stent restenosis. The aims of the present study were to correlate the IVUS measurement of in-stent intimal hyperplasia (IH) with QCA indexes of restenosis, to find out whether QCA is an adequate surrogate of IVUS, and, using either QCA and IVUS data, to define the sample sizes needed to demonstrate the effectiveness of strategies to reduce in-stent restenosis. The database of the European Imaging Laboratory was used to screen 154 stents implanted between 1997 and 2001 and studied by IVUS at 6 +/- 1 months of follow-up. All cases underwent serial QCA assessment (preintervention, postintervention, and follow-up). Only 131 cases with single stent implantation in native coronary arteries were included in the study. Stent restenosis, defined as percent diameter stenosis (DS) > 50%, was present at QCA in 69 out of 131 cases (53%). Linear regression analyses were performed to correlate the amount of IH, calculated by IVUS as the average of all cross-section areas (CSA; mean % IH CSA) and QCA indexes of restenosis (late loss and % DS). A positive significant correlation was found between IVUS mean % IH CSA and QCA % DS (r = 0.74; P < 0.0001) and between IVUS mean % IH CSA and QCA late loss (r = 0.72; P < 0.0001). Based on IVUS measurements of mean % IH CSA, a total sample size of 74 stents would be required in a two-arm comparison to have 0.80 power to detect at 0.05 significant level a 30% difference between two compared groups. Alternatively, adopting the QCA late loss, 230 stents would be required. QCA measurements of late in-stent restenosis are well correlated with IVUS calculation of in-stent neointimal formation. IVUS assessment of IH allows smaller sample sizes than QCA to document significant reductions of in-stent restenosis. Therefore, the use of IVUS should be encouraged in comparison studies aimed at revealing significant neointimal differences in small sample size populations.