This study was a prospective, open-label, randomized, single-center trial, registered at ClinicalTrials.gov (NCT01869842). A total of 117 patients with 124 coronary lesions were enrolled between December 2011 and December 2012 and randomly assigned to receive either OCT- or angiography-guided implantation of a zotarolimus-eluting stent (Endeavor ResoluteTM, Medtronic CardioVascular; Santa Rosa, California, United States). Inclusion criteria were: a) age ≥ 20 years old and significant coronary de novo lesion(s) (≥ 70% diameter stenosis on visual estimation), and b) a native coronary artery with a reference vessel diameter between 2.5mm and 4.0mm that could be covered by a single stent. Exclusion criteria were the following: a) refusal to participate; b) participation in other study protocols; c) lesions with significant left main disease or chronic total occlusion; d) lesions in a grafted vessel, thrombosis, or bifurcation lesions requiring 2 stents; e) an ejection fraction ≤ 30%; f) allergy to either antiplatelet agents or the contrast dye; g) known renal failure with baseline creatinine level ≥ 2.0mg/dL, or end-stage renal disease; h) life expectancy < 1 year; i) prior DES treatment of a different vessel within 3 months; j) presence of an overlapping stent or long stent (> 3 0mm); k) lesion calcification visible on angiography, and l) current pregnancy, or women of childbearing potential. During the same study period, percutaneous coronary intervention (PCI) was performed in 1300 patients. Of these patients, 1183 were excluded. Twenty-five patients refused participation, 405 were participating in other study protocols, and 753 met the remaining criteria, as follows: left main disease in 75, chronic total occlusion in 55, graft vessel disease in 23, totally occluded thrombotic lesion in 125, bifurcation lesions requiring 2 stents in 45, ejection fraction < 30% in 48, chronic renal failure (creatinine level ≥ 2mg/dL) in 92, long lesion (> 30mm stent) or overlapping stents in 135, and anatomy not suitable for OCT procedure in 155. This randomized study was approved by the institutional review board of our institute and written consent was obtained from all enrolled patients.

All study participants fulfilling the enrollment criteria of this study were randomly assigned in a 1:1 ratio by an interactive web-based response system to receive OCT- or angiography-guided PCI. To preserve a balance between the 2 strategies, randomization was stratified according to the presence of diabetes mellitus, acute coronary syndrome, and the estimated length and diameter of the prospective DES implant. All patients received at least 75mg of acetylsalicylic acid and a loading dose of 300mg of clopidogrel at least 12hours pre-PCI. Unfractionated heparin was administered as needed to maintain the activated clotting time of > 250seconds. All PCI procedures were performed according to current standard techniques. In the OCT-guided arm, adjuvant postdilation was performed at operator discretion based on OCT findings. In the angiography-guided arm, stent optimization including adjuvant postdilation was based on a visual estimation of angiographic findings and procedural success was defined as ≤ 20% residual stenosis after stent placement by visual estimation. Postprocedure treatment included a 12-month prescription of dual antiplatelet therapy with 100mg acetylsalicylic acid and 75mg clopidogrel daily.

Quantitative coronary angiography analysis was performed before and after stent implantation, and at 6-month follow-up using an off-line quantitative coronary angiographic system (CASS system, Pie Medical Instruments; Maastricht, The Netherlands) in an independent core laboratory (Cardiovascular Research Center, Seoul, Korea). Reference vessel and minimal luminal diameters were obtained by comparison to the guidance catheter from diastolic frames in a single, matched view showing the smallest minimal luminal diameter: post-PCI and follow-up angiograms were evaluated in the same projection. Acute gain was defined as the difference between preprocedure and postprocedure minimal luminal diameter. Late loss was defined as the change in minimal luminal diameter between postprocedure and follow-up.

At postprocedure and 6-month follow-up after PCI, OCT imaging of the target lesion was performed using a frequency–domain OCT system (C7-XR OCT imaging system, LightLab Imaging, Inc., St. Jude Medical; St. Paul, Minnesota, United States). In this study, OCT cross-sectional images were recorded at 100 fps while a catheter was pulled back at 20mm/s within the stationary imaging sheath. All OCT images were analyzed blind by independent analysts at the core laboratory (Cardiovascular Research Center).

Cross-sectional OCT images were analyzed at 1mm intervals. Stent and luminal cross-sectional areas were measured, and neointimal hyperplasia cross-sectional area was assessed as the difference between the stent and luminal cross-sectional areas. Mean or median values are reported in this study. The neointimal hyperplasia thickness was measured as the distance between the endoluminal surface of the neointima and the strut, and an uncovered strut was defined as having a neointimal hyperplasia thickness of 0μm.8 A malapposed strut of zotarolimus-eluting stent was defined as detachment from the vessel wall by ≥ 110μm (stent strut thickness 91μm + abluminal polymer thickness 6μm, considering blooming artifact). The proportions of uncovered or malapposed struts were identified from OCT cross-sections, and are expressed as a percentage of total struts visible in the survey. In addition, uncovered struts associated with major adverse cardiovascular events, including cardiovascular death, nonfatal myocardial infarction, and stent thrombosis, were subclassified based on the degree of strut coverage as described (favorable coverage, < 5.9% uncovered; unfavorable coverage, ≥ 5.9% uncovered).9 Malapposed struts, which included any amount of malapposition, were further classified into persistent, resolved, or late stent malapposition, according to the change in malapposition visible in matching frames between postprocedure and follow-up OCT.10 Intrastent thrombi were defined as irregular masses protruding into the lumen, > 250μm at the thickest point.11

Clinical follow-up was performed at 1, 3, 6, and 12 months after PCI; angiographic and OCT follow-up was performed at 6 months. The primary end point was the percentage of uncovered struts in the 6-month follow-up OCT assessments. Secondary end points were: a) the presence of malapposed struts in the 6-month follow-up OCT; b) the occurrence of major adverse cardiac events at 12 months, defined as a composite of cardiac death, nonfatal myocardial infarction, or patients requiring target lesion revascularization, and c) stent thrombosis at 12 months, as defined by the Academic Research Consortium.12

This study was designed to compare the 6-month follow-up strut coverage of OCT-guided PCI vs angiography-guided PCI. We hypothesized that the improved stent apposition associated with high-resolution OCT guidance might reduce the incidence of uncovered struts by 50% at the 6-month follow-up OCT, compared to angiography guidance. Assuming 4.8% (4.3%) incidence of uncovered struts at 6-month follow-up, based on our prior studies of zotarolimus-eluting stents at 3 months (6.2%) and 9 months (3.4%),13,14 a sample size of 51 patients for each arm would provide 80% power and a 5% alpha error rate. Taking into account a 10% loss rate and exclusion due to poor OCT image quality, 114 patients were needed for the study. Statistical analysis was performed with the Statistical Analysis System software (SAS; 9.1.3., SAS Institute; North Carolina, United States). Categorical data are presented as numbers and percentages, and were compared with chi-square statistics or Fisher exact test. Continuous data are presented as the mean (standard deviation) (median) and compared with the Student t test. If the distributions were skewed, a nonparametric test was used. Hierarchical linear modeling was applied for the clustering problem. Specifically, data for lesions, cross-sections, and struts were modeled for each patient as a random effect variable. With this modeling, the clustering and correlation problems could be controlled. In case of clustered variables with nonnormal distributions, we integrated the results of hierarchical linear modeling with nonparametric analysis. A value of P < .05 denoted statistical significance.

Our study has some limitations. First, we cannot compare the incidence of acute stent malapposition associated with OCT and angiography guidance because postprocedure OCT examination was only performed in the OCT-guided arm. In addition, adjuvant postdilation for malapposition or stent expansion was not defined as a specific criterion but was performed at operator discretion in this trial. We need to investigate optimal OCT criteria to improve stent strut coverage, apposition, and clinical outcomes during follow-up. Second, our results are only informative for simple lesions, since patients with complex coronary lesions—diffuse long lesions, small (< 2.5mm) or large (> 4.0mm) vessel diseases, left main diseases, chronic total occlusion, and bifurcation lesions—were specifically excluded. Therefore, we cannot speculate that the impact of OCT guidance would be greater in complex lesions and cannot recommend the use of OCT in every patient with simple lesions because its cost-effectiveness is not fully identified yet. Currently this is actively being investigated, but other previous studies have already suggested a potential clinical benefit of OCT-guided PCI.24,27 Third, we only used zotarolimus-eluting stents and a follow-up of 6 months; therefore, any extrapolation of our findings to other DES and longer follow-up durations should be done cautiously. Finally, although the sample size was adequate to assess the stent strut coverage using OCT at 6 months follow-up, it was too small to evaluate the long-term clinical outcomes of OCT-guided PCI.