Intracoronary imaging techniques (referred to as imaging hereafter), such as intravascular ultrasound (IVUS) and optical coherence tomography (OCT), can provide superior visualization than angiography alone in the assessment of lesion characteristics and poststenting results in percutaneous coronary intervention (PCI) (1–3). Randomized controlled trials (RCTs) have shown that imaging can reduce the rate of target vessel revascularization and myocardial infarction in selected patients, but the mortality outcomes are consistently neutral (4–8). In the contrary, many observational studies showed a mortality benefit with IVUS-guided PCI after confounder adjustment (9–11). In a meta-analysis of 31 RCTs and adjusted observational studies, IVUS was associated with lower mortality, but such association was neutralized if the analysis was restricted to RCTs (12). Although this disparity can be because of the unmeasured bias inherent to observational studies, it can also be contributed by operators' differential competency in performing angiography- and imaging-guided PCI. In practices with a high proportion of imaging-guided PCI, operators may become reliant on imaging and become less familiar with performing PCI with angiography alone. It is possible that the mortality benefit seen in observation studies is a reflection of worse survivals with angiography-guided PCI in operators who heavily rely on imaging guidance, and therefore not reproduced in RCTs which operators, by design, perform half of the interventions with angiography guidance.

The hypothesis that imaging may improve outcomes at an individual level, but paradoxically worsen outcomes for the patients receiving angiography-guided PCI with a high proportion of imaging use at a population level can be alarming. Thus, we aimed to determine the association between the utilization rate of imaging at a population level (referred to as “background imaging rate” hereafter) and mortality in patients receiving angiography-guided PCI.

In this study, imaging-guided PCI was associated with a similar mortality rate at 1 year compared with angiography-guided PCI at an individual level. However, higher background imaging rate was associated with a paradoxically increase in all-cause mortality and cardiovascular mortality rates at 1 year for patients undergoing angiography-guided PCI. To our best knowledge, this relationship has not been previously described and we refer to it as the imaging paradox.

Many RCTs have shown that imaging could reduce major adverse cardiovascular events (MACEs), driven by a lower rate of target vessel revascularization but not mortality (4, 5, 7, 8). In the IVUS-XPL trial (n = 1,400), IVUS-guided PCI resulted in a lower risk of target lesion revascularization than angiography-guided PCI (4, 17). In the ULTIMATE trial (n = 1,448), IVUS reduced target lesion failure based on lesion-level analysis but not patient-level analysis (5). These findings are consistently discrepant with observational studies, where IVUS-guided PCI was associated with a reduction in MACEs including mortality (9–11). In a meta-analysis of 31 RCTs and adjusted observational studies, IVUS was associated with lower mortality, but such association was neutralized if the analysis was restricted to RCTs (12). This disparity may be contributed by operators' differential competency of performing angiography- and imaging-guided PCI. Since RCT are ethically bounded to be performed by operators competent in both angiography- and imaging-guided PCI, and by the process of randomization half of the PCI are done under angiography guidance, they may not reflect the real-world scenario in which operators are at different proficiency levels in imaging vs. angiography-guided PCI. The existence of the imaging paradox can potentially exaggerate the mortality benefits of imaging guidance seen in the observational studies.

The imaging paradox may originate from certain operators becoming over reliant on imaging guidance and less familiar with angiography-guided PCI. It is known that PCI outcomes are volume-dependent (18, 19). Operators who utilize imaging guidance routinely, will consequently have a lower volume of angiography-guided PCI. These operators may lose the ability to detect and interpret subtle details from angiograms. For example, it was shown that a careful read on angiography is adequate to guide stent optimization in calcified lesions (20, 21), nevertheless IVUS can provide superior sensitivity for detection of coronary calcification (22). Similarly, many other elements of PCI optimization, including landing zone, stent sizing, stent expansion, and apposition can be readily detectable by imaging (23–26). Operators who become overly reliant on imaging may become less sensitive to important angiographic manifestations, and resulting in suboptimal outcomes when performing angiography-guided PCI.

The imaging paradox has a resemblance to the radial paradox (27). It was noted in an RCT that clinical benefits of radial access were entirely confined to centers where the proportion of radial PCI was high (i.e., 80–98%) (28, 29), raising suspicion that results were merely a reflection of suboptimal outcomes with femoral access performed by operators who used radial access almost exclusively in pretrial procedures (30). We suspect that the imaging paradox is even more relevant than the radial paradox because (1) the visual superiority of imaging is far too superior to angiography-making operators easily become reliant (31), (2) there is no quick or easy way to improve angiography reading (32, 33), (3) the difference is in mortality.

Currently, imaging for PCI is considered adjunctive, and major American or European guidelines do not recommend for routine use (13, 14). Imaging is sometimes unsuitable for certain patients, and also requires additional procedural time, equipment, and cost (31, 34). Therefore, the ability to proficiently perform PCI under angiography guidance is still the backbone. While we recognize that the uptake of imaging is highly variable, with utilization rate of 30 to 80% in certain regions in Asia (35, 36). but only 5% or less in Europe and the USA (11, 37, 38), we believe that training programs should include solid and comprehensive training in angiography-guided PCI. Operators should perform a certain minimum volume of angiography-guided PCI to maintain proficiency and minimize complications. From our threshold analysis, we suspect that hospitals should maintain ~50% of the PCI performed under angiography guidance.

The imaging paradox has important implications in future design of RCT for imaging. The design should ensure that any benefit detected is truly attributable to imaging guidance. One method is to select centers and operators who have a good balance of angiography- and imaging-guided PCI during the pretrial period. Future studies should also not only focus on the effects of imaging on the individual level, but also on a population level. Our data can help generate hypothesis and aid sample size calculation for future RCTs.

This study had some limitations. First, the observational nature of the study conferred risks of unmeasured confounding and bias. It was possible that, for instance, general frailty or perceivable limited life expectancy could possibly incline operators to forgo imaging. However, we used multivariate analysis to minimize the effects of confounders, and the results were consistent in multiple sensitivity analysis including the exclusion of early non-survivors. More importantly, the existence of the imaging paradox cannot be studied in RCTs by nature. Second, analysis was not performed at the operator level, as the information available to us was incomplete, and PCI are commonly performed by multiple operators in this locality. However, institutional experience correlates more strongly with survivals than operator experience (39–41). Third, this study is not designed to detect a deskilling process. It is uncertain that whether operators highly skilled in angiography-guided PCI who transitioned to imaging-guided PCI might have been able to skilfully revert back to angiography-guided PCI when necessary, or, the imaging paradox is contributed by newly trained operators who predominantly perform imaging-guided PCI and struggle with angiography-guided PCI because of suboptimal training.

In conclusion, we observed an increased mortality risk in the patients undergoing angiography-guided PCI in practices with a higher proportion of PCI done under intracoronary imaging guidance at a population level. The existence of the imaging paradox should call for appropriate training and maintenance of competency to improve outcomes for patients receiving angiography-guided PCI.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

The studies involving human participants were reviewed and approved by Institutional Review Board of the University of Hong Kong/Hospital Authority (Reference Number UW 20-176). Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

AN and C-WS were responsible for the conception and design of the study. AN analyzed the data collected by AI. AN interpreted the data. AN, PN, and L-TL drafted the manuscript. All authors revised and approved the final manuscript, and are accountable for the accuracy and integrity of the work.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.