Articles

www.thelancet.com Publushed online October 30, 2016 http://dx.doi.org/10.1016/S0140-6736(16)31922-5 1

Optical coherence tomography compared with intravascular ultrasound and with angiography to guide coronary stent implantation (ILUMIEN III: OPTIMIZE PCI): a randomised controlled trialZiad A Ali, Akiko Maehara, Philippe Généreux, Richard A Shlofmitz, Franco Fabbiocchi, Tamim M Nazif, Giulio Guagliumi, Perwaiz M Meraj, Fernando Alfonso, Habib Samady, Takashi Akasaka, Eric B Carlson, Massoud A Leesar, Mitsuaki Matsumura, Melek Ozgu Ozan, Gary S Mintz, Ori Ben-Yehuda, Gregg W Stone, for the ILUMIEN III: OPTIMIZE PCI Investigators*

SummaryBackground Percutaneous coronary intervention (PCI) is most commonly guided by angiography alone. Intravascular ultrasound (IVUS) guidance has been shown to reduce major adverse cardiovascular events (MACE) after PCI, principally by resulting in a larger postprocedure lumen than with angiographic guidance. Optical coherence tomography (OCT) provides higher resolution imaging than does IVUS, although findings from some studies suggest that it might lead to smaller luminal diameters after stent implantation. We sought to establish whether or not a novel OCT-based stent sizing strategy would result in a minimum stent area similar to or better than that achieved with IVUS guidance and better than that achieved with angiography guidance alone.

Methods In this randomised controlled trial, we recruited patients aged 18 years or older undergoing PCI from 29 hospitals in eight countries. Eligible patients had one or more target lesions located in a native coronary artery with a visually estimated reference vessel diameter of 2·25–3·50 mm and a length of less than 40 mm. We excluded patients with left main or ostial right coronary artery stenoses, bypass graft stenoses, chronic total occlusions, planned two-stent bifurcations, and in-stent restenosis. Participants were randomly assigned (1:1:1; with use of an interactive web-based system in block sizes of three, stratified by site) to OCT guidance, IVUS guidance, or angiography-guided stent implantation. We did OCT-guided PCI using a specific protocol to establish stent length, diameter, and expansion according to reference segment external elastic lamina measurements. All patients underwent final OCT imaging (operators in the IVUS and angiography groups were masked to the OCT images). The primary efficacy endpoint was post-PCI minimum stent area, measured by OCT at a masked independent core laboratory at completion of enrolment, in all randomly allocated participants who had primary outcome data. The primary safety endpoint was procedural MACE. We tested non-inferiority of OCT guidance to IVUS guidance (with a non-inferiority margin of 1·0 mm²), superiority of OCT guidance to angiography guidance, and superiority of OCT guidance to IVUS guidance, in a hierarchical manner. This trial is registered with ClinicalTrials.gov, number NCT02471586.

Findings Between May 13, 2015, and April 5, 2016, we randomly allocated 450 patients (158 [35%] to OCT, 146 [32%] to IVUS, and 146 [32%] to angiography), with 415 final OCT acquisitions analysed for the primary endpoint (140 [34%] in the OCT group, 135 [33%] in the IVUS group, and 140 [34%] in the angiography group). The final median minimum stent area was 5·79 mm² (IQR 4·54–7·34) with OCT guidance, 5·89 mm² (4·67–7·80) with IVUS guidance, and 5·49 mm² (4·39–6·59) with angiography guidance. OCT guidance was non-inferior to IVUS guidance (one-sided 97·5% lower CI –0·70 mm²; p=0·001), but not superior (p=0·42). OCT guidance was also not superior to angiography guidance (p=0·12). We noted procedural MACE in four (3%) of 158 patients in the OCT group, one (1%) of 146 in the IVUS group, and one (1%) of 146 in the angiography group (OCT vs IVUS p=0·37; OCT vs angiography p=0·37).

Interpretation OCT-guided PCI using a specific reference segment external elastic lamina-based stent optimisation strategy was safe and resulted in similar minimum stent area to that of IVUS-guided PCI. These data warrant a large-scale randomised trial to establish whether or not OCT guidance results in superior clinical outcomes to angiography guidance.

Funding St Jude Medical.

Published Online October 30, 2016 http://dx.doi.org/10.1016/S0140-6736(16)31922-5

See Online/Comment http://dx.doi.org/10.1016/S0140-6736(16)32062-1

*Investigators listed in the appendix

New York Presbyterian Hospital and Columbia University, New York, NY, USA (Z A Ali MD, A Maehara MD, T M Nazif MD, O Ben-Yehuda MD, Prof G W Stone MD); Cardiovascular Research Foundation, New York, NY, USA (Z A Ali, A Maehara, P Généreux MD, T M Nazif, M Matsumura BS, M O Ozan MS, G S Mintz MD, O Ben-Yehuda, Prof G W Stone); St Francis Hospital, Roslyn, New York, NY, USA (R A Shlofmitz MD); Centro Cardiologico Monzino Istituto di Ricovero e Cura a Carattere Scientifico, Milan, Italy (F Fabbiocchi MD); Ospedale Papa Giovanni XXIII, Bergamo, Italy (G Guagliumi MD); Northwell Health, Manhasset, New York, NY, USA (P M Meraj MD); Hospital Universitario de La Princesa, Madrid, Spain (F Alfonso MD); Emory University Hospital, Atlanta, GA, USA (Prof H Samady MD); Wakayama Medical University, Wakayama, Japan (Prof T Akasaka MD); Eastern Cardiology, Greenville, NC, USA (E B Carlson MD); and University of Alabama, Birmingham, AB, USA (Prof M A Leesar MD)

Correspondence to: Prof Gregg W Stone, New York Presbyterian Hospital and Columbia University, New York, NY 10032, USA gs2184@columbia.edu

IntroductionAngiography is the most commonly used imaging method to guide procedural decision making during percutaneous coronary intervention (PCI). However, angiography has

various well recognised limitations.1 Angiography pro­vides a two­dimensional representation of a complex three­dimensional structure and only displays luminal dimensions and characteristics, without information

Articles

2 www.thelancet.com Publushed online October 30, 2016 http://dx.doi.org/10.1016/S0140-6736(16)31922-5

about plaque morphology, vascular remodelling, or atherosclerosis burden. Angiography is also suboptimal in identifying stent underexpansion, malapposition, residual dissection, thrombus, and plaque protrusion. These limitations are overcome, in part, by intravascular ultra­sound (IVUS), which allows tomographic cross­sectional imaging of the vessel wall. Findings from large observational cohort studies, randomised trials, and meta­analyses have shown that by achieving larger luminal dimensions than does angiography guidance, IVUS­guided drug­eluting stent implantation might reduce major adverse cardiovascular events (MACE), including target lesion re vascularisation (TLR) and stent thrombosis.2–8 Despite these data and guideline recommendations,9 IVUS­guided PCI is infrequently used in most countries. Low axial resolution (150–200 µm), poor discrimination of plaque subtypes, slow pullback, and absence of approp­riately powered randomised trials are often cited as the causes for low levels of adoption.

Optical coherence tomography (OCT) is a newer intravascular imaging method than IVUS that provides

rapid acquisition of higher resolution (10–20 µm) images capable of more accurately identifying thrombus, lipid, calcium, fibrous cap thickness, dissections, plaque prolapse, stent malapposition, and strut coverage.10 Nonetheless, few studies of OCT­guided stenting have been done, and OCT is not widely used for this purpose. In the ILUMIEN I study,11 OCT guidance changed PCI procedural planning in 57% of cases. Surprisingly, in 31% of cases, OCT led to choice of a smaller diameter stent size than would have occurred with angiography alone. Investigators of a small randomised comparison12 of OCT­guided versus IVUS­guided PCI also reported smaller acute luminal gains using OCT than using IVUS, a finding potentially due to the low­depth penetration of OCT compared with IVUS at lipid­rich lesion sites.10 Thus, whether or not OCT guidance of stent implantation can achieve similar luminal dimensions or clinical outcomes as can IVUS guidance is unknown.

By contrast with the lesion site, the true vessel size is assessable with OCT in disease­free reference segments.10 We therefore developed an OCT­specific technique for

Research in context

Evidence before this studyWe searched PubMed using the search terms “optical coherence tomography”, “intravascular ultrasound”, “angiography”, and “percutaneous coronary intervention” using MeSH terms and appropriate variations up to Dec 5, 2014, with no language restrictions, before designing our study. We found no previous randomised controlled trials that compared all three methods in terms of either acute procedural success or clinical outcomes. Several studies, mostly observational, but some randomised controlled trials compared intravascular ultrasound (IVUS) with angiography. Meta-analyses of these studies identified reductions in target lesion failure, stent thrombosis, major adverse cardiovascular events, and death with use of IVUS, largely as a result of larger minimum stent area. Investigators of observational studies comparing optical coherence tomography (OCT) with angiography suggested that although OCT guidance frequently led physicians to change their percutaneous coronary intervention (PCI) strategy, it often led to implantation of smaller diameter stents than with angiography. Only one head-to-head study of 70 patients compared OCT with IVUS directly, in which IVUS led to larger postprocedural minimum stent area than did OCT.

Added value of this studyThis is the first randomised controlled trial to compare OCT-guided, IVUS-guided, and angiography-guided percutaneous coronary intervention. The study was done in eight countries in 29 hospitals with liberal inclusion criteria representing the global practice of interventional cardiology. Despite its better resolution than that of IVUS, OCT technology has various limitations, including incomplete visualisation of the vessel wall in some lipid-rich lesions, resulting in use of smaller

stents based on luminal dimensions than with IVUS and angiography. To overcome these limitations, we introduced a standardised method for stent sizing based on preintervention OCT measurements of the external elastic lamina. OCT guidance using this stent optimisation protocol was non-inferior to IVUS guidance for the primary endpoint of postprocedure minimum stent area. OCT guidance led to greater stent expansion and acute procedural success than did angiography guidance, with fewer untreated major dissections and malappositions than with IVUS guidance.

Implications of all the available evidenceSince the inception of our study, the 800 patient multicentre randomised controlled OPINION trial showed that target vessel failure of OCT guidance was non-inferior to that of IVUS guidance at 12 months. These results are consistent with what we would have predicted from the similar minimum stent area achieved with the two imaging methods in our study. Another 240 patient multicentre randomised controlled trial, DOCTORS, showed that OCT guidance slightly improved postprocedural fractional flow reserve in patients with non-ST-segment elevation myocardial infarction compared with angiography guidance, without affecting clinical outcomes. Our study is the first to compare all three methods and substantiates the prevailing view that imaging guidance confers advantages compared with angiography guidance alone and, furthermore, that OCT guidance is non-inferior to IVUS guidance for achieving acute procedural success. OCT guidance led to fewer untreated major dissections and malappositions than did IVUS guidance. Larger randomised controlled trials than this one are required to show whether or not OCT guidance results in superior clinical outcomes than does IVUS guidance or angiography guidance.

Articles

www.thelancet.com Publushed online October 30, 2016 http://dx.doi.org/10.1016/S0140-6736(16)31922-5 3

See Online for appendix

stent implantation and sizing on the basis of reference vessel diameter measurements and did a three­arm randomised trial to establish whether or not this novel OCT­guided approach is non­inferior or superior to IVUS guidance and superior to angiography guidance during PCI.

MethodsStudy design and participantsThe ILUMIEN III: OPTIMIZE PCI study was a randomised controlled trial done at 29 hospitals in eight countries. We considered for enrolment patients aged 18 years or older undergoing PCI with a metallic drug­eluting stent for angina (stable or unstable), silent ischaemia, non­ST­segment elevation myocardial infarct­ion (MI), or recent ST­segment elevation MI (>24 h from initial presentation). Eligible patients had one or more target lesions located in a native coronary artery with a visually estimated reference vessel diameter (by angiography) of 2·25–3·50 mm and a length of less than 40 mm. We excluded patients with the following lesions from randomisation: left main or ostial right coronary artery stenoses, bypass graft stenoses, chronic total occlusions, planned two­stent bifurcations, and in­stent restenosis. Detailed inclusion and exclusion criteria appear in the appendix. All participants provided written informed consent.

The trial was designed by the steering committee and funder. Trial monitors ensured data accuracy and collected source documents for review. Angiography, IVUS, and OCT core laboratory analysis, clinical event adjudication, data management, and biostatistics were done by the Cardiovascular Research Foundation. The institutional review board or ethics committee at each participating centre approved the study protocol.

Randomisation and maskingStudy participants who met the enrolment criteria were randomly assigned by study coordinators or investigators at the site in a 1:1:1 ratio to undergo either OCT­guided, IVUS­guided, or angiography­guided stent implantation. Randomisation was done with use of an interactive web­based system in block sizes of three, stratified by site. Operators in the IVUS and angiography groups were masked to the final OCT images. The independent core laboratory that analysed the primary outcome and the independent clinical events committee that adjud­icated clinical and safety events were also masked to treatment assignment.

ProceduresWe treated patients randomly allocated to angiography­guided PCI according to local standard practice. We also did IVUS guidance as per standard practice to optimise lumen dimensions, with pre­PCI and post­PCI imaging mandated. Patients randomly allocated to OCT were treated according to a specific stent optimisation algorithm based

on measurement of the external elastic lamina in the proximal and distal reference segments, designed to achieve larger stent dimensions and more complete lesion coverage than would occur with sizing to the proximal and distal reference lumens. The detailed OCT­guided stent optimisation algorithm with examples are shown in the appendix. We measured proximal and distal reference mean external elastic lamina diameters and used the smaller of these diameters rounded down to the nearest 0·25 mm to determine stent diameter. If the external elastic lamina could not be visualised, we used the proximal and distal reference lumen diameters. We determined stent length as the distance from distal to proximal reference site using the OCT automated lumen detection feature. After stent deployment, we did OCT imaging and, if necessary, iterative high­pressure or larger non­compliant balloon inflations to achieve at least acceptable stent expansion (a minimum stent area of at least 90% in both the proximal and distal halves of the stent relative to the closest reference segment). To avoid perforation, in all cases, the non­compliant balloon diameter could be no larger than the nearest reference vessel external elastic lamina, or up to 0·5 mm larger than the post­PCI mean reference lumen diameter (if the external elastic lamina was not visible). To standardise final intravascular imaging dimensional measurements and qualitative assessment for the primary and secondary endpoints, all randomly allocated patients in the three groups underwent a final OCT run, the images of

Figure: Trial profileIVUS=intravascular ultrasound. OCT=optical coherence tomography. PCI=percutaneous coronary intervention. *These non-randomly allocated patients were used to show investigators’ ability to follow the prescribed OCT guidance procedure.

1759 patients consented

450 randomised

1309 excluded1230 ineligible

1 withdrawal of consent3 withdrawal by investigator1 software malfunction

74 roll-in patients*

158 assigned OCT-guided PCI

146 assigned IVUS-guided PCI

146 assigned angiography-guided PCI

18 excluded5 unable to pass

catheter7 uninterpretable image6 no final OCT done

11 excluded4 unable to pass

catheter4 uninterpretable image3 no final OCT done

6 excluded1 bioabsorbable scaffold4 uninterpretable image1 image received after

data lock

140 included in primary endpoint analysis

135 included in primary endpoint analysis

140 included in primary endpoint analysis

158 30 day follow-up completed

143 30 day follow-up completed

140 30 day follow-up completed

3 lost to follow-up 6 lost to follow-up

Articles

4 www.thelancet.com Publushed online October 30, 2016 http://dx.doi.org/10.1016/S0140-6736(16)31922-5

which were masked to the operators in the IVUS and angiography groups.

We did PCI via femoral or radial access, using unfractionated heparin or bivalirudin anticoagulation per operator preference. We administered dual antiplatelet therapy and other medications per local standard of care, but did not vary this care according to group. We did clinical follow­up by telephone or clinic visit at 30 days and 1 year (currently complete in all patients at 30 days).

OutcomesThe primary efficacy endpoint was post­PCI minimum stent area assessed by OCT in each group, analysed at an independent core laboratory at completion of enrolment. The primary safety endpoint was procedural MACE, defined as procedural complications (angiographic dissection, perforation, thrombus, or acute closure) requiring active intervention (prolonged balloon inflations, additional stent implantation, or peri­

cardiocentesis). OCT imaging measures included acute procedural success defined as the percentage of patients achieving optimal (≥95%) or acceptable (90% to <95%) stent expansion. We also assessed minimum stent expansion, mean stent expansion, tissue or thrombus protrusion, and untreated reference segment disease. We classified edge dissections as major (≥60° of the circumference of the vessel at the site of dissection or ≥3 mm in length) or minor (any visible edge dissection <60° of the circumference of the vessel and <3 mm in length). We also subcategorised stent malapposition, defined as struts clearly separated from the vessel wall by ≥0·2 mm, as major (associated with unacceptable stent expansion [<90%]) or otherwise minor. We also assessed external elastic lamina visibility, intrastent flow area, total flow area, and OCT­guided altered clinical decision making. We also analysed IVUS versus OCT detection of malapposition, dissection, and tissue or thrombus pro­trusion. Angiographic endpoints were minimum lumen diameter, percent diameter stenosis, acute luminal gain, maximum stent or reference vessel diameter, and dissection of at least National Heart, Lung, and Blood Institute type B. A full list of secondary endpoints is provided in the appendix.

We defined MI according to the consensus definition of the Society of Cardio vascular Angiography and Interventions13 and stent thrombosis according to the definite or probable criteria of the Academic Research Consortium.14 We defined MACE as the composite of death, MI, stent thrombosis, or repeat revascularisation. We defined target lesion failure as the composite of cardiovascular death, target vessel MI, or ischaemia­driven target lesion re vascularis ation. An independent clinical events committee adjudicated all clinical and safety events.

Statistical analysisWe planned the trial to randomly allocate about 420 patients (about 140 patients to each of the three groups). We chose the sample size to examine whether or not OCT­guided stent implantation was non­inferior to IVUS­guided stent implantation for post­PCI minimum stent area, as assessed by OCT. With a one­sided α of 0·025, a non­inferiority margin (δ) of 1·0 mm², and an SD of 2·3 mm², and anticipating 5% of cases with immeasurable final minimum stent area, 140 patients per group provided a 94% power to show non­inferiority. If we established that OCT­guided stenting was non­inferior to IVUS­guided stenting, then superiority of OCT­guided stenting to angiography­guided stenting was to be tested. For this test, with a two­sided α of 0·05 and otherwise similar assumptions, an 80% power would be present to detect a minimum difference in minimum stent area of 0·8 mm². Finally, in the event of this superiority being established, we specified a superiority test of OCT­guided stenting to IVUS­guided stenting, which would also have an 80% power to show a minimum stent area difference of 0·8 mm².

OCT (n=158) IVUS (n=146) Angiography (n=146)

Age 66 (59–72) 66 (61–72) 67 (56–74)

Male sex 109 (69%) 107 (73%) 107 (73%)

Hypertension 124 (78%) 113 (77%) 109 (75%)

Dyslipidaemia 115 (73%) 107 (73%) 112 (77%)

Diabetes 52 (33%) 55 (38%) 42 (29%)

Insulin-treated 18 (11%) 16 (11%) 17 (12%)

Current smoker 28 (18%) 19 (13%) 35/145 (24%)

BMI (kg/m²) 28·1 (25·0–31·5) 28·0 (24·9–30·5) 27·4 (25·0–31·2)

Previous MI 35 (22%) 29 (20%) 32 (22%)

Previous PCI in target vessel 11 (7%) 8 (5%) 15 (10%)

Previous CABG 3 (2%) 11 (8%) 8 (5%)

Serum creatinine concentration (mg/dL)

0·90 (0·80–1·03) 0·91 (0·80–1·10) 0·90 (0·79–1·10)

Renal insufficiency* 22 (14%) 16 (11%) 24 (16%)

Dialysis 13 (8%) 6 (4%) 11 (8%)

Clinical presentation

Silent ischaemia 53 (34%) 39 (27%) 41 (28%)

Stable angina 54 (34%) 49 (34%) 50 (34%)

Unstable angina 25 (16%) 33 (23%) 27 (18%)

NSTEMI 20 (13%) 19 (13%) 24 (16%)

Recent STEMI 6 (4%) 6 (4%) 4 (3%)

Pharmacological therapy

Aspirin 126 (80%) 100 (68%) 110 (75%)

ADP antagonist 62 (39%) 55 (38%) 51 (35%)

β adrenergic antagonist 86 (54%) 76 (52%) 83 (57%)

Calcium channel antagonist 45 (28%) 43 (29%) 41 (28%)

ACEI or ARB 72 (46%) 79 (54%) 81 (55%)

Statins 101 (64%) 94 (64%) 97 (66%)

Data are median (IQR) or n (%). OCT=optical coherence tomography. IVUS=intravascular ultrasound. BMI=body-mass index. MI=myocardial infarction. PCI=percutaneous coronary intervention. CABG=coronary artery bypass graft. NSTEMI=non-ST-segment elevation myocardial infarction. STEMI=ST-segment elevation myocardial infarction. ACEI=angiotensin-converting enzyme inhibitor. ARB=angiotensin receptor blocker. *Creatinine clearance of more than 60 mL/min as calculated by the Modification of Diet in Renal Disease formula.

Table 1: Baseline characteristics

Articles

www.thelancet.com Publushed online October 30, 2016 http://dx.doi.org/10.1016/S0140-6736(16)31922-5 5

We did primary analyses in all randomly allocated participants who had primary outcome data, regardless of imaging tests or treatments received. We did secondary analyses in the per­protocol population, consisting of patients without major protocol violations in whom the assigned imaging method was used. According to the Shapiro­Wilk test, the distribution of the minimum stent area primary endpoint was not normally distributed. Therefore, continuous variables are presented as medians with IQRs and we tested groups for non­inferiority and superiority using the Mann­Whitney U test. We tested superiority for all other secondary endpoints using analysis of variance with Dunnett’s adjustment for multiple testing. Models were unadjusted for baseline covariates. We compared categorical data with a χ² or Fisher’s exact test. We considered a two­tailed p value of less than 0·05 statistically significant for all superiority tests. We did all analyses with SAS version 9.4. An independent data and safety monitoring board oversaw the study. This trial is registered at ClinicalTrials.gov, number NCT02471586.

Role of the funding sourceThe funder of the study contributed to study design and data collection, but had no role in data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

ResultsBetween May 13, 2015, and April 5, 2016, we randomly assigned 450 patients at 29 hospitals in eight countries to receive either OCT­guided PCI (158 [35%]), IVUS­guided PCI (146 [32%]), or angiography­guided PCI (146 [32%]). Patient assignment and follow­up are shown in the figure. Final OCT imaging was analysable in 140 (89%) patients in the OCT­guided group, 135 (92%) in the IVUS­guided group, and 140 (96%) in the angiography­guided group. We used imaging to guide stent implantation in 154 (97%) patients assigned to OCT­guided stenting, 139 (95%) assigned to IVUS­guided stenting, and none assigned to angiography­guided stenting. Baseline clinical characteristics are presented in table 1, procedural characteristics are presented in table 2, and angiographic characteristics are presented in table 3.

As measured with OCT, the final median post­PCI minimum stent area (primary outcome) was 5·79 mm² (IQR 4·54–7·34) with OCT guidance, 5·89 mm² (4·67–7·80) with IVUS guidance, and 5·49 mm² (4·39–6·59) with angiography guidance (table 4). Minimum stent area with OCT guidance was non­inferior to IVUS guidance (one­sided 97·5% lower CI –0·70 mm²; p=0·001), but not superior (p=0·42). Minimum stent area with OCT guidance was also not superior to angiography guidance (p=0·12). These results were similar when analysed in the per­protocol population (appendix). No significant interactions were present between site and the primary outcome measure

OCT (n=158) IVUS (n=146) Angiography (n=146) OCT vs IVUS p value OCT vs angiography p value

Arterial access

Radial 104 (66%) 87 (60%) 90/145 (62%) 0·26 0·50

Femoral 54 (34%) 59 (40%) 55/145 (38%) 0·26 0·50

Stents per lesion 1 (1–1) 1 (1–1) 1 (1–1) 0·58 0·93

Stent length (mm) 23·5 (15·0–32·0) 24·0 (16·0–32·0) 20·0 (16·0–30·0) 1 0·27

Maximum stent diameter (mm) 3·00 (2·75–3·50) 3·00 (2·75–3·50) 3·00 (2·75–3·50) 0·36 0·39

Postdilatation balloons used 2 (1–3) 2 (1–3) 1 (1–2) 0·80 0·0005

Maximum balloon size (mm) 3·5 (3·0–4·0) 3·5 (3·0–4·0) 3·0 (3·0–3·5) 0·94 0·0007

Maximum inflation pressure (atm) 18 (16–20) 20 (16–20) 18 (16–20) 0·48 0·02

Stent type

Everolimus-eluting 102 (65%) 84 (58%) 88 (60%) 0·21 0·44

Zotarolimus-eluting 43 (27%) 52 (36%) 51 (35%) 0·11 0·15

Sirolimus-eluting 11 (7%) 9 (6%) 5 (3%) 0·78 0·17

Biolimus-eluting 2 (1%) 1 (1%) 1 (1%) 1 1

Other 0 0 1 (1%) ·· 0·48

Procedure duration (min) 71·0 (57·0–101·0) 73·0 (54·0–97·0) 57·5 (39·0–78·0) 0·99 <0·0001

Fluoroscopy duration (min) 16·0 (10·0–24·0) 16·0 (11·0–25·9) 13·0 (9·0–20·2) 0·74 0·11

Radiation dose (Gy) 1·34 (0·85–2·04) 1·20 (0·74–2·28) 1·16 (0·70–2·02) 0·87 0·39

Contrast volume (mL) 222 (164–285) 190 (140–250) 183 (140–250) 0·004 0·001

Data are n (%), n/N (%), or median (IQR). OCT=optical coherence tomography. IVUS=intravascular ultrasound.

Table 2: Procedural characteristics

Articles

6 www.thelancet.com Publushed online October 30, 2016 http://dx.doi.org/10.1016/S0140-6736(16)31922-5

for any of the two­group comparisons (all interaction p values >0·15).

As shown in table 4, minimum and mean stent expansion were significantly greater with OCT­guided PCI than with angiography­guided PCI, and were similar to that with IVUS­guided PCI. The proportion of procedural success was also higher with OCT­guided PCI than with angiography­guided PCI. Untreated edge dissections, as detected by OCT, were more frequently present after IVUS­guided and angiography­guided PCI than after OCT­guided PCI. Notably, untreated major dissections were more common after IVUS­guided PCI than after OCT­guided PCI. Similarly, untreated major stent

malapposition after PCI was more frequent with both IVUS guidance and angiography guidance than with OCT guidance. OCT­detected frequency of plaque or thrombus protrusion was not significantly different between groups, nor was the prevalence of untreated reference segment disease. Finally, with quantitative coronary angiography, we noted no significant differences between groups in postprocedural stent dimensions or other angiographic secondary outcomes (table 3).

By local site assessment, a more than 180° external elastic lamina visibility at the reference segments (the threshold required to make external elastic lamina­based stent sizing decisions) was similar between OCT (131 [85%] of 155) and

OCT (n=158) IVUS (n=146) Angiography (n=146) OCT vs IVUS p value OCT vs angiography p value

Baseline

Target vessels

Left anterior descending 80 (51%) 68 (47%) 83 (57%) 0·48 0·28

Circumflex 43 (27%) 42 (29%) 31 (21%) 0·76 0·22

Right coronary artery 35 (22%) 36 (25%) 32 (22%) 0·61 0·96

Thrombus 2 (1%) 2 (1%) 2 (1%) 1 1

Calcification (moderate to severe) 32 (20%) 24 (16%) 38 (26%) 0·39 0·23

Reference vessel diameter (mm) 2·78 (2·42–3·12) 2·87 (2·56–3·17) 2·76 (2·50–3·15) 0·34 0·97

Minimum lumen diameter (mm) 0·99 (0·70–1·24) 1·03 ( 0·73–1·26) 0·95 (0·68–1·21) 0·83 0·49

Diameter stenosis (%) 64·1% (55·8–73·1) 63·7% (56·0–73·0) 66·0% (57·8–74·9) 0·99 0·35

Lesion length (mm) 15·5 (11·0–23·2) 15·3 (11·0–23·0) 14·8 (10·6–20·4) 0·99 0·40

TIMI III flow 144 (91%) 131 (90%) 128/145 (88%) 0·68 0·41

Intraprocedural complications

Dissection of at least type B 11 (7%) 11 (8%) 10 (7%) 0·85 0·97

Slow flow or no reflow 1 (1%) 2 (1%) 1 (1%) 0·61 1

Thrombus 1 (1%) 1 (1%) 0 1 0·48

Abrupt closure 0 1 (1%) 0 0·48 ··

Perforation 0 1 (1%) 1 (1%) 0·48 0·48

Final post-PCI

Reference vessel diameter (mm) 2·82 (2·47–3·13) 2·84 (2·58–3·17) 2·72 (2·45–3·04) 0·57 0·56

Acute lumen gain (mm)

In-stent 1·76 (1·45–2·05) 1·75 (1·52–2·10) 1·75 (1·48–2·04) 0·34 0·93

In-segment 1·32 (1·02–1·65) 1·38 (1·15–1·71) 1·36 (1·16–1·69) 0·17 0·24

Minimum lumen diameter (mm)

In-stent 2·71 (2·40–3·05) 2·78 (2·54–3·09) 2·71 (2·44–2·96) 0·15 0·78

In-segment 2·30 (2·03–2·59) 2·41 (2·12–2·74) 2·37 (2·11–2·60) 0·06 0·51

Diameter stenosis (%)

In-stent 5·0% (2·2–10·6) 5·1% (2·6–8·6) 6·2% (2·5–9·9) 0·82 0·91

In-segment 14·7% (9·6–22·9) 13·4% (9·0–20·3) 13·0% (8·7–18·2) 0·28 0·09

TIMI III flow 152/157 (97%) 139/145 (96%) 142 (97%) 0·66 0·82

Complications

Dissection of at least type B 1 (1%) 1 (1%) 1 (1%) 1 1

Slow flow or no reflow 0 1 (1%) 1 (1%) 0·48 0·48

Thrombus 0 0 0 ·· ··

Abrupt closure 0/157 0/144 1 (1%) ·· 0·48

Perforation 0/157 0/144 0 ·· ··

Data are n (%), median (IQR), or n/N (%). OCT=optical coherence tomography. IVUS=intravascular ultrasound. TIMI=thrombolysis in myocardial infarction.

Table 3: Angiographic characteristics and results

Articles

www.thelancet.com Publushed online October 30, 2016 http://dx.doi.org/10.1016/S0140-6736(16)31922-5 7

IVUS (115 [83%]) imaging (p=0·78), although by core laboratory assessment, more than 180° external elastic lamina visibility at the reference segments was slightly less with OCT (137 [95%] of 144 vs 126 [100%] of 126; p=0·02; appendix). The sites were able to use the reference segment external elastic lamina for stent sizing (rather than the lumen) in a similar proportion of cases with OCT and IVUS guidance (appendix). Baseline imaging guidance with both OCT and IVUS substantially changed PCI strategy compared with angiography alone, resulting in an increase in the number of stents, maximum stent diameter, and total stent length used (appendix). Finally, by core laboratory assessment in the IVUS­guided group (in which both IVUS and OCT were done after PCI), OCT was significantly more sensitive at detecting major and minor postprocedural dissections, malapposition, and tissue protrusion (appendix).

OCT imaging guidance resulted in more frequent postdilatation, larger maximum balloon size, and higher balloon pressure than did angiography guidance alone (table 2). Contrast media use was greatest in the OCT­guided group. Procedure duration was longer in the two imaging groups, although we noted no differences in fluoroscopy times or radiation dose. Intraprocedural complications were uncommon and occurred with sim­ilar frequency in the three groups (table 3). Major protocol deviations are presented in the appendix.

Procedural MACE (primary safety outcome) was infrequent and not significantly different between the three groups (table 5). No patient developed acute renal failure. Similarly, the proportions of patients with 30 day MACE and target lesion failure were low and similar between the three groups (table 6).

DiscussionThis study, in which we used a novel OCT­based protocol to determine stent length and diameter according to external elastic lamina measurements in the proximal and distal reference segments to optimise lumen dimensions and lesion coverage, is the first randomised controlled trial comparing OCT­guided, IVUS­guided, and angiography­guided PCI. Post­PCI minimum stent area achieved after OCT­guided PCI was non­inferior to that achieved with IVUS­guided PCI, meeting the primary endpoint of the trial. OCT­guided PCI resulted in significantly greater minimum and mean stent expansion than did angiography­guided PCI, with greater procedural success. We found no significant differences between OCT­guided PCI and IVUS­guided PCI in terms of minimum stent area. However, OCT guidance resulted in fewer untreated major dissections than did IVUS guidance and fewer areas of major stent malapposition than did both IVUS guidance and angiography guidance. Finally, the reference segment external elastic lamina­based OCT stent sizing strategy

OCT (n=140) IVUS (n=135) Angiography (n=140) OCT vs IVUS p value OCT vs angiography p value

Minimum stent area (mm²) 5·79 (4·54–7·34) 5·89 (4·67–7·80) 5·49 (4·39–6·59) 0·42 0·12

Minimum stent expansion (%) 87·6% (16·6) 86·5% (15·9) 82·9% (12·9) 0·77 0·02

Mean stent expansion (%) 105·8% (97·8–119·8) 106·3% (96·7–116·6) 101·4% (91·9–110·2) 0·63 0·001

Acute procedural success

Optimal (≥95%) 36 (26%) 32/130 (25%) 23/136 (17%) 0·84 0·07

Acceptable (90 to <95%) 22 (16%) 16/130 (12%) 5/136 (4%) 0·42 0·0008

Unacceptable (<90%) 82 (59%) 82/130 (63%) 108/136 (79%) 0·45 0·0002

Intrastent flow area (mm²) 5·54 (4·34–7·05) 5·71 (4·59–7·58) 5·42 (4·25–6·36) 0·56 0·32

Total flow area (mm²) 5·68 (4·59–7·30) 5·87 (4·76–7·59) 5·52 (4·42–6·63) 0·72 0·27

Any dissection 39 (28%) 53/134 (40%) 61 (44%) 0·04 0·006

Major 19 (14%) 35/134 (26%) 26 (19%) 0·009 0·25

Minor 20 (14%) 18/134 (13%) 35 (25%) 0·84 0·02

Intimal 16 (11%) 11/134 (8%) 21 (15%) 0·37 0·38

Medial 27 (19%) 45/134 (34%) 40 (29%) 0·007 0·07

Adventitial 1 (1%) 0/134 0 1 1

Any malapposition 58 (41%) 52 (39%) 83 (59%) 0·62 0·003

Major 15 (11%) 28 (21%) 44 (31%) 0·02 <0·0001

Minor 43 (31%) 24 (18%) 39 (28%) 0·01 0·60

Any plaque or thrombus protrusion 94 (67%) 100 (74%) 95 (68%) 0·21 0·90

Major 27 (19%) 27 (20%) 25 (18%) 0·88 0·76

Minor 67 (48%) 73 (54%) 70 (50%) 0·30 0·72

Reference segment disease 44 (31%) 45 (33%) 39 (28%) 0·74 0·51

Data are median (IQR), mean (SD), n (%), or n/N (%). OCT=optical tomography. IVUS=intravascular ultrasound.

Table 4: Postprocedure OCT measures

Articles

8 www.thelancet.com Publushed online October 30, 2016 http://dx.doi.org/10.1016/S0140-6736(16)31922-5

was safe, with few procedural and 30 day MACE, similar between groups.

The higher resolution of OCT than of IVUS imaging confers greater sensitivity for detection of postprocedural dissections, malapposition, and thrombus and tissue protrusion.15,16 These advantages were again seen in this study. However, the low­depth penetration of light through lipid­rich plaque results in an inability of OCT to visualise the true vessel size (delineated by the external elastic lamina) at the lesion site in some cases.10,12 Most previous OCT studies11,12,17 have thus used luminal dimensions for selection of stent size, not external elastic lamina. By contrast, the greater depth penetration of IVUS than of OCT allows visualisation of the external elastic lamina at the lesion site in almost all cases without circumferential calcium or attenuated plaque.12,18 Another factor that might lead operators to use smaller devices (and thus achieve smaller lumen dimensions) with OCT guidance than with IVUS guidance is the fact that IVUS overestimates linear dimensions by about 10%, whereas OCT more accurately represents true measures of distance than does IVUS.16

This factor led to operators in the ILUMIEN I study11 to choose a smaller stent size in 31% of patients than they otherwise would have by angiography guidance alone. In a small randomised trial,12 larger poststent dimensions were achieved with IVUS guidance than with OCT guidance, driven by more frequent postdilatation and higher implantation and postdilatation pressures in the IVUS group. Since the post­PCI minimum stent area achieved is the most important determinant of freedom from early and late MACE after stenting,2–8,19 achievement of smaller luminal dimensions with OCT guidance than with IVUS guidance would represent a major limitation of OCT, offsetting the benefits of its better resolution. However, investigators of non­randomised studies20 using propensity score­matched analysis have reported similar stent expansion with OCT guidance and IVUS guidance; therefore, whether or not OCT guidance can achieve similar luminal dimensions to those with IVUS guidance is unknown.

To overcome the low­depth penetration of OCT in lipid­rich lesions, we developed a novel OCT protocol for stent optimisation in which stent diameter and length

OCT (n=158) IVUS (n=146) Angiography (n=146) OCT vs IVUS p value OCT vs angiography p value

Procedural MACE 4 (3%) 1 (1%) 1 (1%) 0·37 0·37

Procedural complications

Dissection 2 (1%) 0 1 (1%) 0·50 1

Perforation 0 1 (1%) 0 0·48 ··

Thrombus 2 (1%) 0 0 0·50 0·50

Acute closure 1 (1%) 0 0 1 1

Active interventions required

Prolonged balloon inflation 0 0 0 ·· ··

Additional stent required 4 (3%) 1 (1%) 1 (1%) 0·37 0·37

Pericardiocentesis 0 0 0 ·· ··

Thrombus aspiration 0 0 0 ·· ··

Data are n (%). OCT=optical coherence tomography. IVUS=intravascular ultrasound. MACE=major adverse cardiac events.

Table 5: Procedural safety outcomes

OCT (n=158) IVUS (n=143) Angiography (n=140) OCT vs IVUS p value OCT vs angiography p value

MACE 4 (3%) 2 (1%) 1 (1%) 0·69 0·38

Target lesion failure 2 (1%) 1 (1%) 1 (1%) 1 1

Death 0 0 0 ·· ··

Myocardial infarction 2 (1%) 1 (1%) 0 1 0·50

Target vessel myocardial infarction 1 (1%) 1 (1%) 0 1 1

Periprocedural 1 (1%) 1 (1%) 0 1 1

Spontaneous 0 0 0 ·· ··

Non-target vessel myocardial infarction 1 (1%) 0 0 1 1

Any revascularisation 3 (2%) 1 (1%) 1 (1%) 0·62 0·63

Ischaemia-driven target lesion revascularisation 1 (1%) 0 1 (1%) 1 1

Stent thrombosis 1 (1%) 0 0 1 1

OCT=optical coherence tomography. IVUS=intravascular ultrasound. MACE=major adverse cardiac events.

Table 6: Clinical outcomes at 30 days

Articles

www.thelancet.com Publushed online October 30, 2016 http://dx.doi.org/10.1016/S0140-6736(16)31922-5 9

are selected on the basis of external elastic lamina dimensions in the proximal and distal reference segments, which are nearly always visible by OCT in disease­free segments.10 Moreover, definition of the reference segments as regions with at least 180° of external elastic lamina visible should ensure that stent landing zones have minimal residual plaque and that edge dissections will be uncommon (both of which are strongly related to MACE after stent implantation).19,21 In this study, this novel OCT­based stent optimisation protocol resulted in a similar degree of stent expansion and minimum stent area as with IVUS­guided stent implantation, and resulted in fewer untreated edge dissections (especially major dissections and medial dissections, which are most likely to be associated with subsequent adverse events). Overall, OCT detected nearly 25% of total dissections and 15% of major dissections that were missed by IVUS, substantiating the ability of OCT to detect dissections not visible by IVUS (as previously shown),15,16 thus facilitating their treatment with additional balloon inflations or stents if necessary. Additionally, the better resolution of OCT in detecting stent malapposition resulted in fewer un­treated major stent malappositions with OCT gui dance than with either IVUS guidance or angiography guidance and fewer untreated total malappositions than with angiography guidance. As investigators of some studies22–25 have found a relationship between late stent malapposition and subsequent adverse outcomes (par­ticularly late stent thrombosis), this advantage of OCT guidance might confer a prognostic benefit. Finally, the reference segment external elastic lamina­based strategy tested in this study was safe, with no perforations and few procedural and 30 day MACE, similar to those with IVUS­guided and angiography­guided PCI.

The minimum stent area attained with this novel external elastic lamina­based OCT sizing strategy was non­inferior to that attained with IVUS guided­PCI, achieving the primary endpoint of the study. Indeed, the median OCT­guided minimum stent area in ILUMIEN III was about 0·8 mm² larger than that achieved with standard OCT guidance in ILUMIEN II20 (5·8 mm² [95% CI 4·5–7·3] vs 5·0 mm² [3·9–6·4]), despite similar pre­PCI reference vessel diameters (2·8 mm [2·4–3·1] vs 2·7 [2·3–3·0]). OCT guidance and IVUS guidance both led to use of more postdilatation balloon inflations, with greater maximal balloon diameters and pressure than with angiography guidance alone. However, median minimum stent area was not significantly different between OCT guidance and angiography guidance (although OCT guidance did result in better stent expansion and procedural success than did angiography guidance). Although this difference might have become significant in a large study (our trial was powered to detect a 0·8 mm² difference), the close outcomes might also be partly explained by interventions in the angiography group

being done by operators experienced in intravascular imaging. Furthermore, during OCT­guided stent optimisation, operators might have forgone further post­PCI balloon dilatations because of concerns about safety when the achieved minimum stent area was perceived to be large enough, despite not achieving optimal expansion. Detailed analysis is ongoing to establish whether or not OCT­guided minimum stent area would have been even greater had the OCT optimisation protocol been followed to the letter. In this regard, at least 180° of the external elastic lamina could be identified in 95% of reference segments by the core laboratory, suggesting that nearly all patients could undergo OCT external elastic lamina­guided stenting according to our criteria. However, site­reported reference segment external elastic lamina detection of at least 180° was only about 85%, and the external elastic lamina was used for OCT­based sizing in only about 75% of cases, showing that the results of OCT guidance might have been even better with enhanced training than was observed in this study.

This study provides complementary information to two other randomised trials. The Optical Frequency Domain Imaging vs Intravascular Ultrasound in Percutaneous Coronary Intervention study26 was a multicentre random­ised controlled non­inferiority trial comparing 1 year clinical outcomes of lumen sizing with OCT to vessel sizing with IVUS in 800 patients. The primary endpoint, target vessel failure at 12 months, was not significantly different between the two imaging methods. Of note, this study was done in Japan where intravascular imaging is used in most cases and thus did not have an angiographic control arm. The randomised Does Optical Coherence Tomography Optimize Results of Stenting study27 invest­igated the ability of OCT guidance to improve post­PCI fractional flow reserve in 240 patients with non­ST­segment elevation MI compared with angiography. Post­PCI OCT identified cases of underexpansion, malapposition, incomplete lesion coverage, and edge dissection, which led to more frequent use of postdilatation and higher final fractional flow reserve than did angiography alone (0·94 [SD 0·04] vs 0·92 [0·05]; p=0·005). Minimum stent area was the best predictor of post­PCI fractional flow reserve of more than 0·90 with receiver operator characteristic analysis. However, the clinical implications of a slightly greater post­PCI fractional flow reserve achieved with OCT guidance than with angiography guidance are unknown,28 and no significant differences in the prop­ortions of 6 month events occurred between the two groups. Larger randomised trials than have been done are required to establish whether or not differences exist between clinical outcomes after PCI guided by OCT, IVUS, and angiography.

Our study has various limitations. First, we did not mandate a specific IVUS­guided stent strategy. As no standardised consensus exists on how to use IVUS for stent selection and implantation, requiring a specific

Articles

10 www.thelancet.com Publushed online October 30, 2016 http://dx.doi.org/10.1016/S0140-6736(16)31922-5

technique in the IVUS group might have introduced bias and reduced the generalisability of the results. Importantly, external elastic lamina­based sizing was used in more than 70% of patients in the IVUS group. Nonetheless, additional studies are required to establish whether or not the IVUS­guided results might be further improved by application of a specific IVUS stent optimisation protocol. A similar consideration holds for angiography guidance (eg, whether or not online quantitative coronary angiography might improve results). Second, bioresorbable scaffolds were not allowed; whether or not imaging­based optimisation might improve luminal dimensions and outcomes with these novel devices is unknown and should not be extrapolated from this study. Third, masking of investigators and patients was not feasible. However, we used post­PCI OCT assessment in each group (masked to the operators in the IVUS guidance and angiographic guidance arms) to eliminate dis crepancies in measurement between groups attrib utable to the imaging device. We further minimised bias by masking the core laboratories and clinical events adjudication committee to the imaging guidance method. Fourth, the trial was powered for a non­inferiority margin of 1·0 mm²; a substantially larger trial than this one would be required to exclude small differences between OCT guidance and IVUS guidance. Finally, demonstration of improved clinical outcomes in an adequately powered randomised trial is required before routine use of OCT guidance during stent implantation can be recommended. The observations in this study that external elastic lamina­based OCT sizing is safe and effective in achieving non­inferior stent dimensions to IVUS guidance and superior stent expansion to angiography guidance provides the foundation for the planned ILUMIEN IV clinical outcomes trial.ContributorsZAA, AM, PG, GSM, OB­Y, and GWS designed the study, analysed data, and prepared the manuscript. RAS, FF, TMN, PMM, FA, HS, EBC, and MAL enrolled patients into the study and revised the manuscript. GG, HS, and TA enrolled patients into the study, were members of the steering committee, provided crucial input into the conduct of the trial, and revised the manuscript. MM analysed data. MOO was the primary biostatistician.

Declaration of interestsZAA reports grants from St Jude Medical and personal fees from St Jude Medical, Acist Medical, and Cardiovascular Systems, outside the submitted work. AM reports payment for independent clinical events committees, core laboratory, and statistical analysis for work at the Cardiovascular Research Foundation from St Jude Medical during the conduct of the study and grants and personal fees from St Jude Medical and Boston Scientific outside the submitted work. PG reports payment for independent clinical events committees, core laboratory, and statistical analysis for work at the Cardiovascular Research Foundation from St Jude Medical during the conduct of the study and personal fees from Abbott Vascular, grants from Boston Scientific, and grants and personal fees from Cardiovascular Systems, outside the submitted work. RAS reports personal fees from Cardiovascular Systems outside the submitted work and is a proctor for optical coherence tomography training on behalf of St Jude Medical. However, no financial relationship exists—it is for educational purposes only. GG was on the advisory board

through the hospital for St Jude Medical during the conduct of the study and reports personal fees from St Jude Medical and Boston Scientific and grants from Abbott Vascular and Boston Scientific, outside the submitted work. PMM reports grants, personal fees, and non­financial support from Medtronic, Abiomed, and Avinger outside the submitted work. HS reports grants from St Jude Medical during the conduct of the study and outside the submitted work. TA reports grants from St Jude Medical Japan during the conduct of the study and grants and personal fees from St Jude Medical Japan, Terumo, and Goodman, and personal fees from Daiichi Sankyo, outside the submitted work. EBC reports personal fees from Boston Scientific outside the submitted work. MM reports payment for independent clinical events committees, core laboratory, and statistical analysis for work at the Cardiovascular Research Foundation from St Jude Medical during the conduct of the study. MOO reports payment for statistical analysis to the Cardiovascular Research Foundation from St Jude Medical during the conduct of the study. GSM reports grants from St Jude during the conduct of the study and grants from St Jude, grants and personal fees from Boston Scientific and Volcano, and personal fees from Acist, outside the submitted work. OB­Y reports grants from St Jude Medical during the conduct of the study. GWS reports grants from the Cardiovascular Research Foundation during the conduct of the study and personal fees from Velomedix, Toray, Matrizyme, Miracor, TherOx, Reva, V­wave, Vascular Dynamics, Ablative Solutions, Neovasc, and Medical Development Technologies, outside the submitted work, has equity or options in the MedFocus family of funds, Guided Delivery Systems, Micardia, Vascular Nonotransfer Technologies, Cagent, Qool Therapeutics, Caliber, Aria, and the Biostar family of funds outside the submitted work, and he is a consultant on prasugrel patent litigation paid for by Lupin Pharmaceuticals. All other authors declare no competing interests.

AcknowledgmentsThis study was funded by St Jude Medical.

References1 Mintz GS, Popma JJ, Pichard AD, et al. Limitations of angiography

in the assessment of plaque distribution in coronary artery disease: a systematic study of target lesion eccentricity in 1446 lesions. Circulation 1996; 93: 924–31.

2 Zhang YJ, Pang S, Chen XY, et al. Comparison of intravascular ultrasound guided versus angiography guided drug eluting stent implantation: a systematic review and meta­analysis. BMC Cardiovasc Disord 2015; 15: 153.

3 Jang JS, Song YJ, Kang W, et al. Intravascular ultrasound­guided implantation of drug­eluting stents to improve outcome: a meta­analysis. JACC Cardiovasc Interv 2014; 7: 233–43.

4 Witzenbichler B, Maehara A, Weisz G, et al. Relationship between intravascular ultrasound guidance and clinical outcomes after drug­eluting stents: the assessment of dual antiplatelet therapy with drug­eluting stents (ADAPT­DES) study. Circulation 2014; 129: 463–70.

5 Elgendy IY, Mahmoud AN, Elgendy AY, Bavry AA. Outcomes with intravascular ultrasound­guided stent implantation: a meta­analysis of randomized trials in the era of drug­eluting stents. Circ Cardiovasc Interv 2016; 9: e003700.

6 Ahn JM, Kang SJ, Yoon SH, et al. Meta­analysis of outcomes after intravascular ultrasound­guided versus angiography­guided drug­eluting stent implantation in 26,503 patients enrolled in three randomized trials and 14 observational studies. Am J Cardiol 2014; 113: 1338–47.

7 Hong SJ, Kim BK, Shin DH, et al. Effect of intravascular ultrasound­guided vs angiography­guided everolimus­eluting stent implantation: the IVUS­XPL randomized clinical trial. JAMA 2015; 314: 2155–63.

8 Song HG, Kang SJ, Ahn JM, et al. Intravascular ultrasound assessment of optimal stent area to prevent in­stent restenosis after zotarolimus­, everolimus­, and sirolimus­eluting stent implantation. Catheter Cardiovasc Interv 2014; 83: 873–78.

9 Lotfi A, Jeremias A, Fearon WF, et al. Expert consensus statement on the use of fractional flow reserve, intravascular ultrasound, and optical coherence tomography: a consensus statement of the Society of Cardiovascular Angiography and Interventions. Catheter Cardiovasc Interv 2014; 83: 509–18.

Articles

www.thelancet.com Publushed online October 30, 2016 http://dx.doi.org/10.1016/S0140-6736(16)31922-5 11

10 Tearney GJ, Regar E, Akasaka T, et al. Consensus standards for acquisition, measurement, and reporting of intravascular optical coherence tomography studies: a report from the International Working Group for Intravascular Optical Coherence Tomography Standardization and Validation. J Am Coll Cardiol 2012; 59: 1058–72.

11 Wijns W, Shite J, Jones MR, et al. Optical coherence tomography imaging during percutaneous coronary intervention impacts physician decision­making: ILUMIEN I study. Eur Heart J 2015; 36: 3346–55.

12 Habara M, Nasu K, Terashima M, et al. Impact of frequency­domain optical coherence tomography guidance for optimal coronary stent implantation in comparison with intravascular ultrasound guidance. Circ Cardiovasc Interv 2012; 5: 193–201.

13 Moussa ID, Klein LW, Shah B, et al. Consideration of a new definition of clinically relevant myocardial infarction after coronary revascularization: an expert consensus document from the Society for Cardiovascular Angiography and Interventions (SCAI). J Am Coll Cardiol 2013; 62: 1563–70.

14 Cutlip DE, Windecker S, Mehran R, et al. Clinical end points in coronary stent trials: a case for standardized definitions. Circulation 2007; 115: 2344–51.

15 Bezerra HG, Attizzani GF, Sirbu V, et al. Optical coherence tomography versus intravascular ultrasound to evaluate coronary artery disease and percutaneous coronary intervention. JACC Cardiovasc Interv 2013; 6: 228–36.

16 Kubo T, Akasaka T, Shite J, et al. OCT compared with IVUS in a coronary lesion assessment: the OPUS­CLASS study. JACC Cardiovasc Imaging 2013; 6: 1095–104.

17 Kubo T, Shinke T, Okamura T, et al. Optical frequency domain imaging vs. intravascular ultrasound in percutaneous coronary intervention (OPINION trial): study protocol for a randomized controlled trial. J Cardiol 2016; published online Jan 4. DOI:10.1016/j.jjcc.2015.11.007.

18 Mintz GS, Nissen SE, Anderson WD, et al. American College of Cardiology Clinical Expert Consensus Document on Standards for Acquisition, Measurement and Reporting of Intravascular Ultrasound Studies (IVUS). A report of the American College of Cardiology Task Force on Clinical Expert Consensus Documents. J Am Coll Cardiol 2001; 37: 1478–92.

19 Prati F, Romagnoli E, Burzotta F, et al. Clinical impact of OCT findings during PCI: the CLI­OPCI II study. JACC Cardiovasc Imaging 2015; 8: 1297–305.

20 Maehara A, Ben­Yehuda O, Ali Z, et al. Comparison of stent expansion guided by optical coherence tomography versus intravascular ultrasound: the ILUMIEN II study (observational study of optical coherence tomography [OCT] in patients undergoing fractional flow reserve [FFR] and percutaneous coronary intervention). JACC Cardiovasc Interv 2015; 8: 1704–14.

21 Kobayashi N, Mintz GS, Witzenbichler B, et al. Prevalence, features, and prognostic importance of edge dissection after drug­eluting stent implantation: an ADAPT­DES intravascular ultrasound substudy. Circ Cardiovasc Interv 2016; 9: e003553.

22 Guagliumi G, Sirbu V, Musumeci G, et al. Examination of the in vivo mechanisms of late drug­eluting stent thrombosis: findings from optical coherence tomography and intravascular ultrasound imaging. JACC Cardiovasc Interv 2012; 5: 12–20.

23 Souteyrand G, Amabile N, Mangin L, et al. Mechanisms of stent thrombosis analysed by optical coherence tomography: insights from the national PESTO French registry. Eur Heart J 2016; 37: 1208–16.

24 Nakano M, Yahagi K, Otsuka F, et al. Causes of early stent thrombosis in patients presenting with acute coronary syndrome: an ex vivo human autopsy study. J Am Coll Cardiol 2014; 63: 2510–20.

25 Taniwaki M, Radu MD, Zaugg S, et al. Mechanisms of very late drug­eluting stent thrombosis assessed by optical coherence tomography. Circulation 2016; 133: 650–60.

26 Cox CE. Positive signs for OCT guidance in PCI. May 20, 2015. http://www.tctmd.com/show.aspx?id=128936 (accessed Sept 13, 2016).

27 Meneveau N, Souteyrand G, Motreff P, et al. Optical coherence tomography to optimize results of percutaneous coronary intervention in patients with non­ST­elevation acute coronary syndrome: results of the multicenter, randomized DOCTORS study (Does Optical Coherence Tomography Optimize Results of Stenting). Circulation 2016; 134: 906–17.

28 Wolfrum M, Fahrni G, de Maria GL, et al. Impact of impaired fractional flow reserve after coronary interventions on outcomes: a systematic review and meta­analysis. BMC Cardiovasc Disord 2016; 16: 177.

Comment

www.thelancet.com Published online October 30, 2016 http://dx.doi.org/10.1016/S0140-6736(16)32062-1 1

Optical coherence tomography: not quite ready Intravascular imaging has revolutionised the way coronary intervention has been viewed. Both intravascular ultra­sound (IVUS) and, to a greater degree, optical coherence tomography (OCT), allow exquisite assessment of the luminal wall and structure to a near­histological tissue assessment level. OCT also allows detailed assessment of stent characteristics and, in particular, of adequacy of deployment and early identification of procedural complications, such as edge dissections. Although this level of imaging provides invaluable information for the interventionalist, data are scarce for its use in clinical practice to improve outcomes.

Previously, findings from retrospective observational studies1,2 have shown benefits with use of OCT compared with angiography alone; however, OCT has not been robustly challenged in the setting of a randomised trial until now. Ziad Ali and colleagues3 are to be congratulated for the ILUMIEN III: OPTIMIZE PCI trial, published in The Lancet, which was a considered and well conducted prospective, multicentre, randomised, controlled trial. It aimed to show non­inferiority of OCT to IVUS in achievement of a minimal stent area (MSA) and superiority to angiography alone. The rationale for the choice of MSA as the primary endpoint was based on findings from previous studies showing lower IVUS­derived MSA as being associated with greater in­stent stenosis (late lumen loss) in drug­eluting and bare­metal stents.4,5

The study achieved its prespecified non­inferiority result with regards to MSA compared with IVUS (one­sided 97·5% lower confidence interval –0·70 mm²; p=0·001); however, it did not show superiority to angiography alone (p=0·42). Ali and colleagues also found that OCT guidance was associated with fewer untreated major dissections than with IVUS (19 [14%] of 140 vs 35 [26%] of 135; p=0·009) and fewer major stent malappositions (15 [11%] vs 28 [21%]; p=0·02). Although this finding certainly appears to be important, the clinical significance of these imaging­detected complications remains to be further elucidated. Indeed, use of potential surrogate measures of clinical outcomes (the primary measure and focus of this study) should be viewed with caution. The notion of changing management by a strategy such as OCT because of its ability to detect suboptimal stent deployment is based on perceived

issues or problems rather than real validated prospective clinical outcomes. This putative advantage has been speculated as a benefit for years, yet has not translated into any meaningful clinical result yet.

This issue is coupled with fundamental design comparator issues, which form the basis of the non­inferiority design—ie, no accepted standard to size vessels with IVUS (luminal or media­to­media) or angiography (so­called eyeballing vs quantitative coronary angiography). The basis of sizing of a vessel with use of the proposed external elastic lamina­based OCT technique was achieved by operators in 85% of cases and used only in 75% of cases. Furthermore, operators using this technique had considerable expertise in intravascular imaging, meaning its broad applicability is unknown.

Although OCT represents a visually powerful tool, which can provide insights during situations of uncertainty or complex percutaneous coronary intervention, its safety in routine clinical practice remains paramount. Largely, this safety was substantiated in this trial, although OCT was associated with increased procedural time and contrast use than with angiography. Although the numbers of adverse events are too small for analysis, procedural major adverse cardiac events were increased in the OCT group (four [3%] compared with one [1%] each in the IVUS and angiography­alone groups; p=0·37 for OCT vs angiography). Furthermore, at 30 days, only one (1%) stent thrombosis occurred, in the OCT group, showing

Published Online October 30, 2016 http://dx.doi.org/10.1016/S0140­6736(16)32062­1

See Online/Articles http://dx.doi.org/10.1016/S0140­6736(16)31922­5

Astie

r/BS

IP/S

cienc

e Ph

oto

Libr

ary

Comment

2 www.thelancet.com Published online October 30, 2016 http://dx.doi.org/10.1016/S0140-6736(16)32062-1

that stent failure in the era of contemporary second­generation drug­eluting stents and potent antiplatelet agents is now an uncommon event, but can occur despite optimisation using precise imaging techniques. Clearly, the issue of stent failure is multifactorial, and the possibility that a small increase in MSA could improve outcomes has to be placed in this context and viewed against the potential negatives, including disruption of plaques in vessels during imaging, patients having longer procedures with more contrast than with angiography alone, and the obvious extra cost to the patient, insurer, and health­care system.

Although we applaud and welcome the findings of the ILUMIEN III: OPTIMIZE PCI trial as an important step in the right direction, the interventional cardiology community have previously fallen into the trap of using surrogate measures to argue for the benefit of a strategy or approach which, when tested systematically, has shown no difference or change in clinical outcomes. Although this study strengthens the evidence for intracoronary OCT use, whether or not it should be used routinely in all cases remains unclear and in this regard,

is therefore still not ready for widespread use and further clinical outcome­based studies are necessary.

*Ravinay Bhindi, Usaid K Allahwala Department of Cardiology, Royal North Shore Hospital, Sydney, NSW 2065, Australia ravinay.bhindi@sydney.edu.au

We declare no competing interests.

1 Prati F, Di Vito L, Biondi­Zoccai G, et al. Angiography alone versus angiography plus optical coherence tomography to guide decision­making during percutaneous coronary intervention: the Centro per la Lotta contro l’Infarto­Optimisation of Percutaneous Coronary Intervention (CLI­OPCI) study. EuroIntervention 2012; 8: 823–29.

2 Prati F, Romagnoli E, Burzotta F, et al. Clinical impact of OCT findings during PCI: the CLI­OPCI II study. JACC Cardiovasc Imaging 2015; 8: 1297–305.

3 Ali ZA, Maehara A, Généreux P, et al, for the for the ILUMIEN III: OPTIMIZE PCI Investigators. Optical coherence tomography compared with intravascular ultrasound and with angiography to guide coronary stent implantation (ILUMIEN III: OPTIMIZE PCI): a randomised controlled trial. Lancet 2016; published online Oct 30. http://dx.doi.org/10.1016/S0140­6736(16)31922­5.

4 Hong MK, Mintz GS, Lee CW, et al. Intravascular ultrasound predictors of angiographic restenosis after sirolimus­eluting stent implantation. Eur Heart J 2006; 27: 1305–10.

5 Doi H, Maehara A, Mintz GS, et al. Impact of post­intervention minimal stent area on 9­month follow­up patency of paclitaxel­eluting stents: an integrated intravascular ultrasound analysis from the TAXUS IV, V, and VI and TAXUS ATLAS workhorse, long lesion, and direct stent trials. J Am Coll Cardiol Interv 2009; 2: 1269–75.