The OPTICO-Integration-II trial was a prospective, single-center, randomized, open label trial. Consecutive patients undergoing coronary angiography were screened for suitability at the Charité—University Medicine Berlin, Germany, between August 2017 and July 2018. Subjects were considered eligible when angiography revealed significant coronary artery disease (based on visual assessment in at least one native coronary artery together with a positive ischemia test in stable coronary artery disease or lesions causing acute coronary syndromes) [9]. Exclusion criteria comprised in-stent restenosis, ST-segment elevation acute coronary syndromes (STE-ACS), cardiogenic shock, left ventricular ejection fraction (LVEF) < 30%, estimated creatinine clearance (eGFR) < 40 ml/min, neoplasia on treatment or without a curative therapeutic approach, life expectancy < 24 months, pregnancy, participation in another investigational clinical trial, as well as inability or unwillingness to give written informed consent. Furthermore, lesions unsuitable for OCT imaging were excluded, such as severely calcified or extremely tortuous vessels, as well as very distal lesions or a reference vessel diameter of > 4 mm or < 2 mm. Finally, patients were randomly assigned in a 1:1:1 ratio to either ACR (group I), OCT (group II) or angiography-guided PCI (group III). Permuted block randomization with variable block size was performed by means of sealed opaque envelopes containing a computer-generated sequence. Randomization was conducted after crossing the target lesion with the coronary guidewire.

The study was approved by the local Ethics Committee at Charité—University Medicine Berlin, Germany (EA1/072/17) and was registered at ClinicalTrials.gov (NCT03646097). Written informed consent was obtained from all participating subjects.

For OCT imaging the OPTIS™ integrated system (St. Jude Medical, St. Paul, MN) with Dragonfly™ Duo (St. Jude Medical, St. Paul, MN, USA) catheters were used. After placement of a standard 0.014 inch coronary guidewire over a 6F guiding catheter and injection of 200 mg of intracoronary nitroglycerin an automatic OCT pullback was performed with a contrast flush (IMERON^® 350; Bracco Imaging Deutschland GmbH, Konstanz, Germany) using an automated power injector. OCT images covering the stenotic area and at least 10 mm of the vessel segments proximal and distal to the lesion and were recorded with an axial resolution of 200 μm.

For study groups I (ACR) and II (OCT), the OCT scan was immediately evaluated by the operator using the regular OCT acquisition console (OPTIS™ integrated system): In group I (ACR) the commercially available semi-automatic co-registration OPTIS™ integrated system was used for co-registration of coronary angiography and OCT [7]. This results in a matched side-by-side view of angiography and OCT was available for the operator, with a small white marker projected over the angiogram and indicating the exact localization (error margin about 1 mm [7]) of the displayed OCT frame on the angiogram. However, in study groups I (ACR) and II (OCT), operators were encouraged to choose a stent-landing zone, which ensures optimal target lesion coverage defined as untreated minimal lumen area (MLA) ≤ 60% of the adjacent reference segment lumen area up to 10 mm from the proximal and distal stent edges [5]. Furthermore, vessel-wall characteristics as plaque load were taken into account for selection of the optimal stent-landing zone. Stent size selection included assessment of the relation to the reference vessel size. Importantly, PCI optimization in the ACR and OCT group followed the ILUMIEN IV—OPTIMAL PCI OCT criteria (NCT03507777). Stent optimization in the angiography-guided PCI group was performed at the discretion of the operator. Details of the PCI strategy including need for pre-dilation, special lesion debarking, vessel reference diameter, intended lesion and stent length, intended stent diameter, intended stenting strategy and device landing zone were documented in every case.

The primary study endpoint was a composite imaging endpoint including major edge dissections and/or LGM as assessed by post-procedural OCT. Major edge dissections were defined as each dissection with flap angle ≥ 60° of the circumference of the vessel at the site of dissection and/or ≥ 3 mm in length after PCI according to the ILUMIEN IV—OPTIMAL PCI OCT criteria (NCT03507777). Longitudinal geographic mismatch (LGM) was defined as every untreated plaque with a minimal lumen area < 4.5mm² within 5 mm of the reference segment. If the segment of the lesion was fully covered and the stent protruded maximal 5 mm beyond the predetermined landing zone in the post-PCI OCT analysis it was considered as no geographic mismatch, again in line to ILUMIEN IV criteria. Secondary study endpoints included each individual component of the primary endpoint as well as the incidence of any edge dissection (any visible edge dissection < 60 degrees of the circumference of the vessel and < 3 mm in length), stent underexpansion [proximal minimal stent area (MSA) < 90% of the proximal reference lumen area, and/or distal MSA < 90% of the distal reference lumen area] and stent malapposition (stent struts separated from the vessel wall with a distance from the adjacent intima of ≥ 0.2 mm and not associated with any side branch). Furthermore, PCI result was assessed by determining MSA, length of the stented segment, number of implanted stents, as well as procedure time, fluoroscopy time, contrast volume and complications during hospital stay.

Sample size calculation was based on a retrospective evaluation of OCT-guided PCI cases (n = 350) within the Charite University Medicine Berlin using a system without ACR, where LGM was detectable in 35% and additional major edge dissections in 5% of all PCI cases. Under the assumption that a combined imaging endpoint including major edge dissections and/or LGM is detectable in 35% of the segments with OCT-guided PCI and can be markedly reduced to 5% of patients by ACR-guided PCI, a total of 48 patients (24 per group) yield a power of 80% (with a 2-sided α = 0.05) to detect a difference in the primary endpoint between patients with ACR-guided (study group I) and OCT-guided PCI (study group II). Furthermore, an equal group (n = 24) who underwent angiography-guided PCI (study group III) was included as an additional control group. Anticipating a drop-out rate of 12%—again based on retrospective image analysis at the study center—a total of 84 patients were considered sufficient. Given the fact that the primary endpoint can occur twice per patient, at the proximal and the distal stent edge, an edge-based analysis was performed to assess differences regarding the endpoints between groups.

Of 198 screened patients, a total of 84 patients were included and randomly assigned to undergo PCI guided by ACR (group I), OCT (group II) or angiography only (group III). The study flow is described in Fig. 1. A total of 14 patients were excluded due to impossibility for OCT imaging (pre-PCI = 2; post-PCI = 4) or low OCT imaging quality (pre-PCI = 3; post-PCI = 5). Baseline clinical characteristics are presented in Table 1. The population was predominantly male and characterized by a high prevalence of type 2 diabetes. Lesion characteristics (Table 2) were comparable among groups. Diameter stenosis as assessed by QCA was similar in the respective groups.

The primary endpoint was significantly reduced by ACR-guided PCI (4.2%, n = 2) as compared to OCT-guided PCI (19.1%, n = 9, p = 0.03) and angiography-guided PCI (25.5%, n = 12, p < 0.01) (Fig. 2, Table 3). There was no significant difference between OCT- and angiography- guided PCI (p = 0.46).

Other predefined secondary endpoints are reported in Table 3. The incidence of LGM was significantly reduced with ACR- (4.2%) as compared to OCT- (17.0%; p = 0.04) and angiography-guided PCI (22.9%; p = 0.02). Major edge dissections affected the proximal edge in 13.9% of the cases and the distal edge in 14.3% of the cases. Whereas OCT-guided PCI was associated with reduced rates of any edge dissections after PCI as compared to angiography-guided PCI (17.4% vs. 25.5%; p = 0.34), no edge dissection was detectable after ACR-guided PCI. Malapposition as defined in this study occurred in 60.0%, 62.5% and 58.3% without a statistically significant difference, respectively. The stent results with respect to MSA and stent expansion were comparable among the three study groups (Table 3). This was also true for procedural characteristics (Table 4) as numbers of implanted stents, stent length, radiation exposure, amount of contrast medium and procedural time did not differ among groups (Table 4). Finally, in-hospital clinical complications were observed in two patients, one in the OCT (repeat coronary angiography due to recurrent chest pain) and one in the angiography group (acute coronary syndrome).

Neither age, type 2 diabetes, nor acute coronary syndrome as index event or high lesion complexity (AHA class B2/C) was identified as significant predictors for the incidence of LGM or major edge dissections after PCI (Table 5). Importantly, the use of ACR- as compared to OCT-guided PCI [OR 0.17 (0.03–0.93); p = 0.04] as well as compared to angiographic-guided PCI [OR 0.12 (0.02–0.66); p = 0.02] was found to be inversely related with the primary endpoint (Table 5).