Comparison of Intravascular Ultrasound, Optical Coherence Tomography, and Conventional Angiography-guided Percutaneous Coronary Interventions: A Systematic Review, Network Meta-analysis, and Meta-regression

Abstract

Background

Intracoronary imaging modalities, including intravascular ultrasound (IVUS) and optical coherence tomography (OCT), provide valuable supplemental data unavailable on CA and has shown to improve clinical outcomes. We sought to compare the clinical efficacy of IVUS, OCT, and conventional coronary angiography (CA)-guided percutaneous coronary interventions (PCI).

Methods

Frequentist and Bayesian network meta-analyses of randomized clinical trials were performed to compare clinical outcomes of PCI performed with IVUS, OCT or CA alone.

Results

A total of 28 trials comprising 12,895 patients were included. IVUS when compared with CA alone was associated with a significantly reduced risk of major adverse cardiovascular events (MACE) (relative risk [RR] 0.74, 95% confidence interval [CI] 0.63–0.88), cardiac death (RR 0.64, 95% CI 0.43–0.94), target lesion revascularization (RR 0.68, 95% CI 0.57–0.80), and target vessel revascularization (RR 0.64, 95% CI 0.50–0.81). No differences in comparative clinical efficacy were found between IVUS and OCT. Rank probability analysis bestowed the highest probability to IVUS in ranking as the best imaging modality for all studied outcomes except for all-cause mortality.

Conclusion

Compared with CA, the use of IVUS in PCI guidance provides significant benefit in reducing MACE, cardiac death, and revascularization. OCT had similar outcomes to IVUS, but more dedicated studies are needed to confirm the superiority of OCT over CA.

Condensed Abstract

Using network meta-analysis models, we pooled data from 28 randomized clinical trials to compare the clinical efficacy of percutaneous coronary interventions (PCIs) guided by intravascular ultrasound (IVUS), optical coherence tomography (OCT), and conventional coronary angiography (CA). IVUS was associated with a significantly reduced risk of major adverse cardiovascular events, cardiac death, target lesion revascularization, and target vessel revascularization compared with CA. No differences in outcomes were found between IVUS and OCT. In conclusion, IVUS provides significant benefit in guiding PCIs, while more studies are needed to confirm the superiority of OCT over CA.

Introduction

Intracoronary imaging with intravascular ultrasound (IVUS) or optical coherence tomography (OCT) allows for better evaluation of vessel anatomy, plaque characteristics, and stent optimization when compared with conventional angiography (CA) (1). Recent randomized controlled trials (RCT) have reported lower rates of major adverse cardiovascular events (MACE) after intracoronary imaging-guided percutaneous coronary intervention (PCI) compared with CA-guided PCI(2–4).While use of IVUS during PCI currently has a class IIa recommendation, particularly in left main or complex coronary artery disease or in patients with prior stent failure (5,6), results from IVUS trials have not yet translated into a class I recommendation for IVUS. Similarly, data on use of OCT-guidance for PCI remains encouraging albeit limited.

IVUS and OCT are two technologically different intravascular imaging modalities for a common end point of optimizing PCI’s. Four randomized controlled trials have compared IVUS and OCT (7–10) and suggested OCT as an acceptable alternative to IVUS. However, IVUS and OCT are technically distinct with inherent limitations. Therefore, we sought to perform a network meta-analysis comparing IVUS-, OCT-, and CA-guided PCI by combining both direct and indirect evidence to produce robust pooled outcomes distinguishing the major intracoronary imaging modalities.

Methods

All supporting data are available within the article and the supplement. We conducted our systematic review in compliance with a published protocol available on Open Science Framework (10.17605/OSF.IO/TGJRE). Our study adhered to the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines(11). Our study was exempt from the purview of the institutional review board as only data from previously published papers was used.

Selection of Trials

Two authors (D.P. and S.An) systematically searched PubMed, Embase, Cochrane Library, and Google Scholar from inception to March 5, 2023. The search strategies are listed in Table S1. The inclusion criteria were as follows: (1) RCT comparing ≥2 invasive coronary imaging modalities in guiding PCI, (2) coronary imaging modalities limited to CA, IVUS, and OCT, (3) documentation of pre-specified clinical endpoints consisting of MACE, all-cause mortality, cardiac death, myocardial infarction (MI), target lesion revascularization (TLR), target vessel revascularization (TVR), and stent thrombosis. Since fractional flow reserve does not provide anatomical information, it was included in the CA arm. Sensitivity analysis was later performed wherein trials that used fractional flow reserve-guided CA were excluded. All clinical presentations were allowed. Full papers written in languages other than English were excluded. In case of ≥2 publications on the same trial, the one with data on longer follow-up was used. Multi-arm trials that compared all the three imaging modalities in a parallel-group design were divided into three parts: part A (IVUS versus CA), part B (OCT versus CA), and part C (OCT versus IVUS).

Data Extraction and Quality Assessment

Two authors (D.P. and S.An) reviewed titles, abstracts, and full texts to select articles that met the eligibility criteria. Selected articles were perused to identify other eligible studies by citation searching. Conflicts were settled by the adjudication of a third author (A.V.). For each selected RCT, trial name (author name if trial name not present), publication year, stent type (bare-metal stent [BMS], drug-eluting stent [DES], or bioresorbable vascular scaffold), follow-up period, intracoronary imaging modalities, and corresponding sample sizes were arranged into a table (Table 1). Baseline demographics, clinical characteristics, clinical presentation, angiographic characteristics, and procedural characteristics were also organized into a table. The primary outcome of interest was MACE, and secondary outcomes included all-cause mortality, cardiac death, MI, stent thrombosis, TLR, and TVR. A table defining all the prespecified outcomes used in each of the trials was created (Table S2). The Cochrane Collaboration’s tool was used to assess the quality and the risk of bias in the RCTs (Table S3) (12).

Table 1.

Trials included in this study

Trial	Yeara	Recruitmentb	Stent	Follow Up	Total Size	Experimental Group		Control Group
						Imaging	Size	Imaging	Size	FFR
RENOVATE-COMPLEX-PCI Part A	2023	2018–2021	DES	3 years	1347	IVUS	800	CA	547	No
RENOVATE-COMPLEX-PCI Part B	2023	2018–2021	DES	3 years	825	OCT	278	CA	547	No
FLAVOUR	2022	2016–2019	DES	2 years	1682	IVUS	844	CA	838	Yes
iSIGHT Part A	2021	2015–2016	DES	2.4 years	99	IVUS	50	CA	49	No
iSIGHT Part B	2021	2015–2016	DES	2.4 years	100	OCT	51	CA	49	No
iSIGHT Part C	2021	2015–2016	DES	2.5 years	101	OCT	51	IVUS	50	No
ULTIMATE	2021	2014–2017	DES	3 years	1423	IVUS	714	CA	709	No
FORZA	2020	2013	DES	13 months	350	OCT	174	CA	176	Yes
IVUS-XPL	2020	2010–2014	DES	5 years	1400	IVUS	700	CA	700	No
MISTIC-1	2020	2014–2016	DES	3 years	109	OCT	54	IVUS	55	No
OPTICO BVS	2020	2016–2017	BVS	1 year	38	OCT	19	CA	19	No
Liu et al.	2018	2010–2015	DES	1 year	336	IVUS	167	CA	169	No
OCT STEMI	2017	2011–2012	DES	9 months	201	OCT	105	CA	96	No
OPINION	2017	2013–2014	DES	1 year	817	OCT	412	IVUS	405	No
DOCTORS	2016	2013–2015	BMS/DES	6 months	240	OCT	120	CA	120	No
ILLUMIEN III Part A	2016	2015–2016	DES	1 month	283	IVUS	143	CA	140	No
ILLUMIEN III Part B	2016	2015–2016	DES	1 month	298	OCT	158	CA	140	No
ILLUMIEN III Part C	2016	2015–2016	DES	1 month	301	OCT	158	IVUS	143	No
AIR-CTO	2015	2010–2011	DES	2 years	230	IVUS	115	CA	115	No
CTO-IVUS	2015	2012–2013	DES	1 year	402	IVUS	201	CA	201	No
Tan et al.	2015	2009–2012	DES	2 years	123	IVUS	61	CA	62	No
OCTACS	2015	2011–2013	DES	6 months	85	OCT	40	CA	45	No
Wang et al.	2015	2012–2013	NR	1 year	80	IVUS	38	CA	42	No
AVIO	2013	2008–2011	DES	2 years	284	IVUS	142	CA	142	No
RESET	2013	2009–2012	DES	1 year	543	IVUS	269	CA	274	No
HOME DES IVUS	2010	2004–2005	DES	18 months	210	IVUS	105	CA	105	No
AVID	2009	1995–1999	BMS	1 year	744	IVUS	369	CA	375	No
DIPOL	2007	2000–2002	BMS	6 months	163	IVUS	83	CA	80	No
TULIP	2003	1998–2001	BMS	6 months	144	IVUS	73	CA	71	No
Gaster et al.	2003	1996–1998	BMS	2.5 years	114	IVUS	54	CA	54	No
OPTICUS	2001	1996–1998	BMS	1 year	548	IVUS	273	CA	275	No
SIPS	2000	1996	BMS	2 years	269	IVUS	121	CA	148	No
RESIST	1998	1995–1997	BMS	6 months	144	IVUS	71	CA	73	No

^aPublication year

^bYear patients were recruited

Abbreviations: BMS, bare-metal stent; BVS, bioresorbable vascular scaffold; CA, conventional angiography; DES, drug-eluting stent; FFR, fractional flow reserve; IVUS, intravascular ultrasound; NR, not reported; OCT, optical coherence tomography

Statistical Analysis

To ensure consistent homogenous data from all the selected trials, risk ratios (RR) and 95% confidence intervals (CI) were manually calculated and zero-cells were corrected using the modified Haldane-Ascombe method(13). Pooled estimates were calculated using frequentist network meta-analysis. Node-splitting analysis was conducted to examine potential inconsistencies between direct and indirect evidence. Heterogeneities in the network models were assessed using tau-squared and I-squared values. P-scores of each imaging modality were generated for all outcomes of interest but were interpreted only when the network meta-analysis showed significant difference among the imaging modalities. Briefly, P-scores denote the mean extent of certainty that one imaging modality is better than its competitors averaged over all competitors with equal weights(14). P-scores range from 0 to 1 in increasing certainty of superiority. We also calculated pooled estimates using Bayesian network meta-analysis whereby a generalized linear model was fitted under a hierarchic Bayesian random-effects framework. Markov-chain Monte Carlo simulations incorporating 4 chains, 5,000 adaptations, and 100,000 iterations produced final Bayesian models, and convergence was confirmed by visualizing time-series and density plots. Surface Under the Cumulative Ranking (SUCRA) scores, a Bayesian counterpart of P-scores, were generated. Rankograms were produced by performing rank probability analyses in the Bayesian framework. Sensitivity analyses including (1) trials that solely used DES, (2) trials that were predominantly done in stable ischemic heart disease or (3) acute coronary syndrome, and (4) trials that did not use fractional flow reserve in CA-guided PCI were conducted.

In addition, we performed mixed-effects logistic meta-regression of MACE reported in trials that compared IVUS or OCT with CA over the years from which the patients were recruited. From all meta-regression models, we generated beta-coefficients with P values, tau-squared, I-squared, H-squared, and R-squared indexes. Sensitivity analysis of meta-regression was conducted in which 2 trials that used fractional flow reserve-guided CA was excluded to better characterize the trend of IVUS and OCT as compared with CA over time. Frequentist and Bayesian network meta-analyses were performed using meta/netmeta and gemtc/rjags packages, respectively. Meta-regression was performed using the meta package. All analyses were done using R version 4.1.0 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Twenty-eight RCTs were included after a systematic search. Eighteen trials compared IVUS with CA (3,4,15–30), 5 trials compared OCT with CA (2,31–34), and 2 trials compared OCT with IVUS (Figure 1) (9,10). Two trials (7,8) were multi-arm trials that compared IVUS, OCT, and CA in a parallel-group design. The remaining 1 trial combined both IVUS and OCT into 1 arm, so the results from its subgroup analysis comparing IVUS with CA, and OCT with CA were used. (35) A total of 12,895 patients were included, comprising of 5,856, 1,411, and 5,628 patients who underwent IVUS-guided, OCT-guided, and CA-guided PCI, respectively (Figure 2). Eighteen trials used DES while seven trials employed BMS (Table 1). One trial used bioresorbable vascular scaffold (34), and another trial used both DES and BMS (33). One trial did not report the type of stents used (25). The Comparison of Fractional flow Reserve And Intravascular ultrasound-guided Intervention Strategy for Clinical OUtcomes in Patients with InteRmediate Stenosis (FLAVOUR) trial and the FFR or OCT Guidance to Revascularize Intermediate Coronary Stenosis Using Angioplasty (FORZA) trial used fractional flow reserve when performing CA-guided PCIs (2,29,30). The demographics and comorbidities of patients included in the trials varied across trials (Table S4). The clinical presentations also were variable across trials, including stable ischemic heart disease, acute coronary syndrome, chronic total occlusion, and left main disease (Table S4).

Figure 1. PRISMA flow diagram of the search for relevant trials.

The flow diagram shows the process whereby databases were screened and surveyed for relevant trials that met the inclusion criteria.

Figure 2. Network plot of selected trials.

The network plot demonstrates the number of comparator-arms and patients included among trials that compared IVUS-, OCT-, and CA-guided percutaneous coronary intervention. The term “comparator-arms” has been used to represent one set of experimental arm and control arm because 2 multi-arm trials simultaneously compared all the three intracoronary imaging modalities with a parallel-group design. The size of the blue circles and blue lines are proportional to the total sample size and number of comparators, respectively.

Variations were noted in the definition of MACE in the trials (Table S2). Some trials included cardiac death whereas others used all-cause mortality. Similarly, many trials incorporated TLR while others used TVR. In addition, a few trials included stent thrombosis in the composite outcomes constituting MACE. The definitions of other outcomes were largely similar among the trials. The risk of bias ranged from low to high (Table S3). Many older trials failed to mention the use of the randomization sequence when randomizing patients. No inconsistencies were observed in the note-splitting analysis for each outcome (Table S5 and Figure S1). Low to moderate heterogeneity were seen in the network models (Table S6).

Pooled estimates from the frequentist network meta-analysis demonstrated that the use of IVUS was associated with lower risk of MACE (RR 0.74, 95% CI 0.63–0.88), cardiac death (RR 0.64, 95% CI 0.43–0.94), TLR (RR 0.68, 95% CI 0.57–0.80), and TVR (RR 0.64, 95% CI 0.50–0.81) when compared with CA (Central Illustration). No difference was seen between IVUS and CA in the outcomes of all-cause mortality, MI, or stent thrombosis (Table 2). None of the outcomes were significantly different between IVUS and OCT, and between OCT and CA. A sensitivity analysis excluding trials that did not solely use DES demonstrated similar findings, except for the lower risk of stent thrombosis seen in IVUS compared with CA and the lower risk of MACE in OCT compared with CA (Table S7). A sensitivity analysis including trials in which >50% of patients had stable ischemic heart disease produced similar findings of MACE, TLR, and TVR, but not with cardiac death (Table S8). Another sensitivity analysis exclusive to trials in which >50% of patients had acute coronary syndrome generated similar findings in addition to the lower risk of cardiac death in OCT compared with CA. Bayesian network meta-analysis reiterated the findings from its frequentist counterpart, but additionally showed lower risk of MI and stent thrombosis in IVUS compared with CA (Table S11).

Central Illustration. Comparison of outcomes among intracoronary imaging modalities.

The figure illustrates statistically significant difference in outcomes among IVUS-, OCT-, and CA-guided PCI in the network meta-analysis.

Abbreviations: ACS, acute coronary syndrome; CA, conventional angiography; DES, drug-eluting stent; IVUS, intravascular ultrasound; MACE, major adverse cardiovascular events; OCT, optical coherence tomography; TLR, target lesion revascularization; TVR, target vessel revascularization

Table 2.

Pooled estimates of network meta-analysis for each outcome

Major adverse cardiovascular events	IVUS
	0.96 (0.65–1.40)	OCT
	0.74 (0.63–0.88)	0.78 (0.53–1.14)	CA
All-cause mortality	IVUS
	0.78 (0.35–1.74)	OCT
	1.15 (0.85–1.56)	1.47 (0.66–3.26)	CA
Cardiac death	IVUS
	0.82 (0.25–2.72)	OCT
	0.64 (0.43–0.94)	0.77 (0.24–2.49)	CA
Myocardial infarction	IVUS
	0.94 (0.48–1.87)	OCT
	0.82 (0.64–1.04)	0.87 (0.44–1.71)	CA
Stent thrombosis	IVUS
	0.79 (0.27–2.32)	OCT
	0.61 (0.36–1.04)	0.77 (0.27–2.23)	CA
Target lesion revascularization	IVUS
	0.83 (0.44–1.55)	OCT
	0.68 (0.57–0.80)	0.82 (0.43–1.55)	CA
Target vessel revascularization	IVUS
	0.86 (0.53–1.41)	OCT
	0.64 (0.50–0.81)	0.74 (0.44–1.24)	CA

Abbreviations: CA, conventional angiography; IVUS, intravascular ultrasound; OCT, optical coherence tomography

P-scores demonstrated the superiority of IVUS over OCT and CA in the outcomes of MACE, cardiac death, TLR, and TVR (Figure 3). SUCRA scores from the Bayesian network model mirrored these findings (Figure S2). Rank probability analysis showed that IVUS had an 60% probability of being the best intracoronary imaging modality in reducing MACE, followed by OCT (39%) and CA (0%) (Figure S3). IVUS consistently outperformed other imaging modalities for all other outcomes except for all-cause mortality.

Figure 3. Bar graph showing P-scores of each type of medication for every outcome.

The bar graphs show the P-scores of IVUS (blue), OCT (red), and CA (green) from the frequentist network meta-analysis for each outcome. Outcomes with significant differences in the pooled estimates of network meta-analysis (Table 2) are marked with an asterisk (*).

Abbreviations: CA, conventional angiography; IVUS, intravascular ultrasound; OCT, optical coherence tomography

Meta-regression of trials that compared IVUS or OCT with CA demonstrated a marginally decreasing trend of MACE (β = −0.018, P-value = 0.058) as the year of recruitment became more recent (Figure S4). In trials that solely compared IVUS with CA, a similar marginally decreasing trend (β = −0.019, P-value = 0.069) was found, but not in those that solely compared OCT with CA (β = −0.103, P-value = 0.503). Full results of the meta-regression are shown in Table S12. When the FORZA trial and the FLAVOUR trial (2,30), both of which incorporated fractional flow reserve into CA-guided PCI, were excluded from the meta-regression models, a significantly decreasing trend of MACE was observed with IVUS or OCT versus CA and IVUS versus CA (Figure 4).

Figure 4. Meta-regression analysis excluding trials that used fractional flow reserve in CA-guided PCI.

The three graphs show the logarithmic association of the risk of MACE with the recruitment year in each trial that compared either IVUS or OCT with CA (Figure 4A). Additional analyses stratified to IVUS-only versus CA (Figure 4B) and OCT-only versus CA (Figure 4C) have also been conducted. Negative beta-coefficient (β) signifies negative correlation and vice versa. P-value of the trend is shown at the center of each graph.

Abbreviations: CA, conventional angiography; IVUS, intravascular ultrasound; MACE, major adverse cardiovascular events; OCT, optical coherence tomography

Discussion

Principal Findings

Our comprehensive analysis of 28 RCTs on IVUS-, OCT-, and CA-guided PCI demonstrated that IVUS was associated with lower risk of MACE, cardiac death, TLR, and TVR when compared with CA alone. Bayesian analysis also indicated that the risks of MI and stent thrombosis were lower in IVUS versus CA. No difference in any of the outcomes was seen between IVUS and OCT. OCT had lower risk of MACE and cardiac death compared with CA only in sensitivity analyses of trials that solely used DES and those that predominantly consisted of acute coronary syndrome, respectively. However, studies on OCT were fewer and smaller, thus having lack of power to detect a potential difference.

While many previous meta-analyses have shown that IVUS-guided PCI improves outcomes when compared with CA alone (36–39), our study is unique in that our findings are solely based on RCTs, which are less influenced by unmeasured confounders and selection bias associated with observational studies. Sattar et al and Siddqi et al recently published meta-analyses on similar topics, but they only included 7 and 13 studies, respectively, about half of which consisted of observational studies(40,41). In contrast, we incorporated 28 RCTs and produced more robust results by generating network meta-analyses models under both frequentist and Bayesian frameworks, unlike conventional pairwise meta-analyses which solely use direct evidence. Importantly, we performed an updated network meta-analysis incorporating additional 5-years of data from the last network meta-analysis on this subject(42) to compare the three imaging modalities. We subsequently performed a sensitivity analysis by excluding BMS use, and also conducted meta-regression between MACE and recruitment years.

Previously, Elgendy et al reported that IVUS-guided PCI reduced the risk of MACE by lowering the risk of ischemia-driven TLR, and Darmoch et al found that the risk of cardiac death, MI, TLR, and stent thrombosis were all significantly lower with IVUS guidance(36,37). Similar findings were noted by Steinvil et al and Malik et al(38,39). IVUS allows visualization of plaque thickness, lesion length, severity of calcification, and measurement of external elastic lamina for stent sizing and lesion preparation prior to stent implantation(43). After stent deployment, IVUS can also evaluate the minimum stent area, evidence of malapposition or under expansion, and edge dissection, which are associated with increased rates of both early and very late stent thrombosis(1).

Similar to IVUS, OCT is the another commercially available intravascular imaging technology. However, in contrast to IVUS, OCT relies on near-infrared light for tissue penetration and can provide cross-sectional images of tissue structure on a micron scale. OCT has nearly 10 times greater resolution when compared with IVUS but is limited in its depth of penetration(44). This allows for high-resolution images of the vessel wall, with advantages in assessing calcium thickness, lipid, thrombus, fibroatheroma, and plaque rupture, as well as neointimal thickness, stent apposition, and edge dissections(43). Commercially available OCT platform in the United States also provides artificial intelligence assistance that can facilitate procedural decisions (45), a feature currently unavailable in many commercial IVUS platforms. However, OCT has its own limitations, including the necessity of replacing blood column by contrast or saline prior to imaging and having to repeat these steps prior to imaging, which may explain why it is not as widely utilized as IVUS(46).

Kuku et al pooled data from 3 RCTs and 3 observational studies and reported no difference in MACE, cardiac death, MI, TLR, or stent thrombosis between IVUS and OCT(47). In a more recent meta-analysis, Sattar et al combined data from 4 RCTs and 3 observational studies and found no significant differences between OCT and IVUS across outcomes(40). Recent trials like the Optical Coherence Tomography Versus Intravascular Ultrasound and Angiography to Guide Percutaneous Coronary Interventions (iSIGHT) trial, the Optical Coherence Tomography Compared with Intravascular Ultrasound and with Angiography to Guide Coronary Stent Implantation (ILUMIEN-III) trial, and the Optical Frequency Domain Imaging versus Intravascular Ultrasound in Percutaneous Coronary Intervention (OPINION) trial also showed the non-inferiority of OCT to IVUS (1,6–9). This is in line with our findings, as well as the ACC/AHA/SCAI guidelines that have endorsed OCT as a reasonable alternative to IVUS except in ostial left main disease(6).

The largest dedicated trial comparing OCT with CA is the Fractional flow Reserve versus Optical Coherence Tomography to Guide Revascularization of Intermediate Coronary Stenoses (FORZA) trial, which did not show a benefit of OCT over fractional flow reserve-guided CA in the outcomes of all-cause mortality, MACE, non-fatal MI, and TVR(2). A previous network meta-analysis by Buccheri et al reported a significant reduction of MACE and cardiac death with OCT versus CA, but included 14 observational studies, and a sensitivity analysis including only RCTs did not demonstrate any significant difference in all studied outcomes(42). A more recent meta-analysis of 6 RCTs and 5 observational studies also did not identify any differences in clinical outcomes, including all-cause mortality, MACE, MI, TLR, TVR, and stent thrombosis(48). In our analysis, the use of fractional flow reserve in the CA arm may have diluted the comparative efficacy of OCT in the FORZA trial, but sensitivity analysis excluding trials that used fractional flow reserve also produced similar results. However, we were able to find that OCT was associated with lower risk of cardiac death and MACE compared with CA in two of our sensitivity analyses.

Many factors should be considered when evaluating the non-significant difference between OCT and CA. Despite the pooling of data from multiple RCTs, comparison of OCT and CA remains underpowered as demonstrated by the wider confidence intervals in the network meta-analysis models. One of the major OCT trials was the ILLUMIEN III trial, which had a follow-up period of 1 month and thus likely resulted in underestimations of ischemic outcomes that may favor OCT in the long-term (7). The Randomized Comparison of Optical Coherence Tomography Versus Angiography to Guide Bioresorbable Vascular Scaffold Implantation (OPTICO BVS) trial (34) used bioresorbable vascular scaffolds, which have inherently higher rates of stent thrombosis owing to greater strut thickness, higher rates of stent under expansion, lower radial strength, and lower achievement of optimization criteria (34,49). Additionally, criteria for optimal stent expansion were less stringent in OCT trials, which more frequently aimed for minimal lumen area >75–80% of reference lumen area instead of minimal stent area ≥90% of reference lumen area (Table S13). The percentage of procedures that met the optimal stent criteria were also lower across OCT arms (mean 56.6%) than across IVUS arms (mean 65.8%) (Figure S5).

Our meta-regression analysis excluding the 3 trials that used fractional flow reserve demonstrated that while a decreasing trend of MACE was seen with later recruitment years in trials that compared IVUS with CA, no trend was seen in those that compared OCT with CA. The risk of MACE is expected to decrease with the introduction of newer-generation DES, improvements in medical therapy, and technical improvements, as seen with IVUS-guided PCIs, but such a pattern was not seen in trials that compared OCT with CA. Although these findings should be interpreted in the context of fewer trials on OCT distributed into a shorter time span (5 years for OCT versus 21 years for IVUS), they call for standardizing OCT-guided stent optimization criteria and appropriate patient selection. Until then, it is reasonable to follow expert consensus and global guidelines based on higher quality evidence on the similarity of OCT compared with IVUS as both modalities can detect mechanisms of stent failure that CA cannot evaluate. In conclusion, either modality of intravascular imaging can be useful to improve patient outcomes, and outcomes may depend on operator expertise.

Limitations

We acknowledge several limitations of our study. The number of trials in the network model was unbalanced with many more arms on IVUS and CA compared with OCT, yielding lower power to comparisons with OCT. The definition of MACE varied among the trials as summarized in Table S2. Significant differences were seen in comorbidities, clinical presentations, target vessels, and complexities in coronary anatomy, adding to the heterogeneity seen in the network models. The CA arm was heterogenous in that some PCIs were guided by fractional flow reserve, but sensitivity analysis excluding trials that used fractional flow reserve yielded similar results. Follow-up periods ranged from 1 month to 5 years. Included trials spanned more than 2 decades, with earlier trials using BMS or first-generation DES; we sought to reduce this heterogeneity by performing a sensitivity analysis of trials that solely used DES, but results were similar. We also performed a meta-regression to account for technical improvements, newer medications, and technological advancements especially seen in IVUS-guided PCIs. However, specific medication utilization across arms could not be accounted for in our analyses. The criteria used to assess optimal stent expansion differed among the trials as summarized in Table S11. The percentage of procedures that met the optimal stent criteria also widely differed from trial to trial as illustrated in Figure S5. Finally, we did not have access to individual patient-level data, thus limiting our analyses to those possible with publicly available study-level data alone.

Conclusion

IVUS-guided PCI was associated with lower risk of MACE, cardiac death, TLR, and TVR compared with CA-guided PCI. Our findings including the results from newer trials continually support the current guidelines that have endorsed the use of IVUS for PCI compared with a CA-alone approach. OCT was similar to IVUS in all outcomes, but additional dedicated trials will be helpful to confirm the superiority of OCT over CA.

Clinical Perspectives.

What is Known?

• Intravascular imaging modalities, such as IVUS and OCT, provide valuable information that can aid stent optimization. Previous studies have suggested the superiority of IVUS over conventional angiography in guiding PCI, but comparisons with OCT have been lacking.

What is New?

• We used both frequentist and Bayesian network meta-analyses including data from 28 randomized controlled trials to compare IVUS, OCT, and conventional angiography. IVUS was superior to conventional angiography in MACE, cardiac death, TLR, and TVR, but no differences were seen between IVUS and OCT, and between OCT and conventional angiography. Meta-regression analyses showed a decreasing trend of MACE with later recruitment years in trials comparing IVUS with conventional angiography, but no significant trend was seen in trials that compared OCT with conventional angiography.

What is Next?

• More randomized controlled trials with larger sample sizes are needed to determine the superiority of OCT over conventional angiography. Establishment of stent optimization criteria and more stringent adherence to the criteria may also be needed to establish the benefits of OCT. The results of our network meta-analysis provide further up-to-date evidence favoring the use of IVUS in PCI and support use of OCT as an alternative to IVUS.

Abbreviations

BMS

Bare-metal stent

CA

Conventional angiography

CI

Confidence interval

DES

Drug-eluting stent

IVUS

Intravascular ultrasound

MACE

Major adverse cardiovascular events

MI

Myocardial infarction

OCT

Optical coherence tomography

PCI

Percutaneous coronary intervention

RCT

Randomized controlled trial

RR

Risk ratio

SUCRA

Surface under the cumulative ranking

TLR

Target lesion revascularization

TVR

Target vessel revascularization

Data Availability Statement

The datasets supporting the conclusions of this article are included within the article and its supplementary file.