Roadmap Fusion Imaging in Percutaneous Coronary Intervention Reduces Contrast Medium Exposure Irrespective of Investigator’s Experience Level

Objectives. Dynamic Coronary Roadmap (DCR) is a software tool that creates a real-time dynamic coronary artery overlay on fluoroscopic images. The efficacy of DCR in significantly reducing contrast medium use during percutaneous coronary interventions (PCI) has previously been shown. In this study, we aimed to determine if DCR is equally effective irrespective of the performing investigator’s experience level.

Methods. In this sub-analysis of a monocentric, open-label, randomized trial, 130 patients with hemodynamic relevant coronary type A and B lesions were randomized and contrast medium use was conducted with (+) or without (-) DCR software. PCI was randomly allocated and performed by an investigator with high (A) or medium (B) experience level.

Results. Overall, contrast medium use was significantly reduced by both investigators in the +DCR group, and Investigator B used significantly less contrast medium with the software than Investigator A. The DCR software was not accompanied by increased radiation exposure for the patients or the teams. On the contrary, dose area product was reduced by both investigators, but was significantly reduced by the highly experienced investigator when using DCR. Fluoroscopy time was not different between investigators. Procedural success was 100%. Serious in-hospital adverse events were not observed. One of Investigator A's patients suffered from post-procedural acute kidney injury in the –DCR group.

Conclusions. DCR significantly reduces contrast medium use during PCI irrespective of investigator’s experience level.

Percutaneous coronary interventions (PCI) are considered a suitable, safe, and effective therapeutic option in daily practice.^1-3 Although the complexity of interventions is constantly increasing, technical improvements and medical therapy contribute to reduced complication rates.

Dynamic Coronary Roadmap (DCR; Philips Healthcare, Best, the Netherlands) is a commercially available software tool for percutaneous coronary interventions. The software creates a digital, motion-compensated overlay from a library based on an existing coronary angiogram that can be superimposed on the fluoroscopic image. This enables the visualization of coronary arteries and the navigation of wires and devices during the procedure without renewed administration of contrast medium (Figure 1). We have demonstrated the first-in-man bifurcation intervention,⁴ investigated the technical aspects of DCR in a safety and efficacy study,⁵ and performed coronary roadmaps in a monocentric, open-label, prospective randomized trial.⁶ However, it is not known and was never investigated whether experience level is a mandatory prerequisite for the effective implementation of DCR in daily clinical practice. While it is known that radiation doses in CT-guided interventions are reduced⁷ and that procedure time, fluoroscopy time, and radiation exposure in mechanical thrombectomy in the anterior circulation are significantly influenced by the interventionalist’s experience,⁸ the impact of coronary roadmaps depending on the interventionalist’s preexisting experience level is unknown. Therefore, we are the first to analyze whether there are different results if DCR is used by either medium- or high-experience level investigators.

A subgroup analysis comparing 2 investigators at 2 different experience levels was performed from the data of a monocentric, open-label, prospective randomized trial.⁶ This clinical trial was registered at ClinicalTrials.gov (NCT03074721). The study was approved by the local ethics committee (Heinrich-Heine University Düsseldorf, No. MPG-MO-36) and performed in accordance with the Declaration of Helsinki; informed consent to participate in the study was obtained from all participants.

DCR software was obtained from Philips Healthcare, Best, the Netherlands. In brief, patients with hemodynamic relevant coronary type A or B lesions⁹ were scheduled for elective PCI. Eligible participants, all 18 years old or older, gave written informed consent before the PCI and were block-randomized (1:1 ratio) to either the Dynamic Coronary Roadmap (+DCR) group or the conventional PCI (-DCR) group. Baseline in-hospital data, procedural and follow-up characteristics were collected. Patients were then allocated to interventionalists with either a high (2800 coronary interventions, Investigator A) or a medium (1300 coronary interventions, Investigator B) experience level. The assignment to Investigator A or Investigator B was random and according to availability in everyday clinical practice.

According to the EAPCI Core Curriculum for Percutaneous Cardiovascular Interventions,¹⁰ both investigators correspond to level V concerning their interventional skills, with total numbers of percutaneous coronary interventions as indicated above determining their classification as high- or medium-experience level.

The primary endpoint of this study was the amount of contrast medium use in each investigation. Secondary endpoints of interest included procedural characteristics (dose area product, fluoroscopy time), the occurrence of major adverse cardiovascular and cerebral events (MACCE) according to the standard definition (cardiovascular death, stroke, myocardial ischemia, and repeated revascularization), procedural success (defined as an open target vessel with a maximum of 20% residual stenosis and thrombolysis in myocardial infarction [TIMI] III flow), and post-procedural acute kidney injury (AKI).

Patients with acute coronary syndrome, who were unable to give written informed consent, and/or had hemodynamic relevant type C lesion were excluded. Two patients from PCI group with DCR were excluded in analysis due to assignment to a third examiner with only 2 examinations within the trial and therefore not sufficient for meaningful analysis. Electronic case report forms were used for documentation.

Statistical analysis. Characteristics are summarized by descriptive statistics. Categorial data are presented as numbers or percentage of total, respectively. Continuous variables are expressed as mean with standard deviation as indicated. The data set was evaluated for normal distribution with the Kolmogorov-Smirnov-Lilliefors test. Parameters were compared between groups using student’s t test or 2-way analysis of variance (ANOVA) as appropriate. P-value less than .05 was considered statistically significant. The effect size of DCR for each endpoint and investigator was determined according to Cohen’s d¹¹ with pooled and n-weighted standard deviation. The sizes were reported as follows: d of 0 to 0.2 represented no to very low effect; d of 0.2 to 0.5 represented low effect; d of 0.5 to 0.8 represented medium effect; and d greater than 0.8 represented strong effect. SPSS Version 20 (IBM) and Prism 8 (GraphPad Software) were used for analysis and graphing.

A patient flowchart (Figure 2) illustrates the patient selection process and allocation to Investigators A and B. Three screening failures occurred, one due to a non-ST-elevation myocardial infarction (NSTEMI), and 2 patients were ruled out because they presented with hemodynamic relevant type C lesions. Procedures were performed with monoplane fluoroscopy imaging by 2 interventional cardiologists (Investigators A and B) in the period between April 2017 and November 2020. Two patients from the PCI group with DCR were excluded in analysis due to assignment to a third examiner with only 2 examinations within the trial.

Baseline characteristics. Patients were between 46 and 91 years old, and predominantly male. Body mass index and body surface area were comparable between the groups examined by Investigators A and B (Table 1).

Procedural characteristics and results. One hundred-thirty patients were randomized to either the –DCR or +DCR group and subsequently 128 were treated by either Investigator A or Investigator B. All patients in the study cohort were characterized by a type A or type B coronary lesion. Hemodynamic relevance of the coronary artery stenosis was verified most frequently by invasive hemodynamic assessment (Investigator A: 86.7%; Investigator B: 90.6%, P=.974), followed by pharmacological cardiac magnetic resonance imaging stress testing (Investigator A: 12%; Investigator B: 9.4%, P=.974). All patients underwent PCI with drug-eluting stents. A 1-stent strategy was most often utilized by both investigators (Investigator A: 85.3%; Investigator B: 98.1%, P=.628), followed by the deployment of 2 stents (Investigator A: 13.3%; Investigator B: 1.9%). Procedural success was 100%. Serious in-hospital adverse events were not observed (Tables 1, 2).

Primary and secondary endpoints. Contrast medium use was not different between investigators when working without software (Table 2). The primary endpoint amount of contrast medium was significantly reduced (Figure 3A-C) by both investigators in the +DCR group compared to –DCR group. Interestingly, Investigator B with medium experience was able to reduce the contrast usage even further than Investigator A.

Secondary endpoints are shown in Table 2. Dose area product without software was significantly higher with investigator A compared to investigator B (Investigator A: –DCR 2415 ± 1430 cGycm²; Investigator B: –DCR 1414 ±1145 cGycm², P=.0023; Investigator A: +DCR 1727 ± 1148cGycm²; Investigator B +DCR 1176 ±766 cGycm², P=.1482). Investigator A was able to significantly reduce DAP with DCR compared to without DCR (P=.0284). Investigator B reduced DAP by tendency, but less than Investigator A (Figure 4A-C).

Fluoroscopy time was not significantly different between groups (Investigator A: –DCR 4.9 ± 3.5 min vs +DCR 4.3 ± 1.9 min, P =.9999; Investigator B: –DCR 6.8 ± 7.3 min vs +DCR 5 ± 2.9 min, P =.1846) and was not different between investigators without software (P =.0745) nor with coronary roadmap (P =.9413) (Figure 4D-F).

When looking at the effect size of DCR in each investigator, Investigator B benefitted most concerning contrast agent reduction. Concerning reduction in dose area product and fluoroscopy time, DCR shows a similar effect size in both investigators.

Procedural success was 100% in both investigators. Serious in-hospital adverse events were not observed. In the -DCR group, one of Investigator A’s patients suffered from post-procedural AKI.

Follow-up. Median follow-up time was 348 days (range 195.5-378 days). No procedure-related readmission was observed. MACCE was driven only by acute coronary syndrome with NSTEMI constellation and subsequent PCI in 1 (1.3%) patient of Investigator A and 0 (0%) patients of Investigator B (P =.484). Heart failure was responsible for 1 (1.3%) patient of Investigator A and 0 (0%) patients of Investigator B (P =.484) to be readmitted during follow-up. Further follow-up data are presented in Table 3.

The Dynamic Coronary Roadmap is a new software tool designed to improve procedural efficacy of percutaneous coronary interventions and facilitate ultra-low contrast percutaneous coronary interventions.¹² In a prospective randomized trial, we showed that the use of contrast medium was significantly reduced by DCR with unchanged fluoroscopy time and complication rate.⁶ However, it was not known whether this result is achievable only by highly experienced interventionalists. Therefore, we undertook a subgroup analysis of safety and efficacy data comparing investigators with medium vs high interventional experience.

In our sub-analysis of the prospective randomized controlled trial, we show that DCR use demonstrates equivalent significant contrast medium reduction by both medium- and high-experience level investigators, DCR is not associated with increased fluoroscopy time for either investigator level, and both investigators performed coronary interventions with optimal success and safety results.

Through this sub-analysis of the prospective study, we were able to show, for the first time, that DCR use is equally applicable to medium- and high-experience level investigators. Therefore, this tool is attractive for clinical routine, and investigators at either experience level can significantly reduce contrast agent exposure by using DCR (Figures 1, 3). Experience level is not only important for the patient’s overall outcome, but has influence on secondary endpoints. For example, during mechanical thrombectomy in the anterior circulation procedure time, fluoroscopy time and radiation exposure were significantly influenced by the interventionalist’s experience.⁸We also know from computed tomography-guided interventions that radiation doses are reduced depending on the investigator’s experience.⁷In our study, we observed a relevant reduction of contrast medium use by DCR at both experience levels, with pronounced savings in contrast medium use by Investigator B, which might lead to the assumption that investigators with medium-experienced spatial vision skills benefit highly from the software overlay. 

Dose area product use was significantly reduced by Investigator A (Figure 4A, B). This could suggest that coronary technical and imaging tools might therefore increase the motivation even of high-experience level investigators to reduce dose area product or foster sensitivity for responsible use of radiation. On the other hand, high dose area product use by Investigator A without the aid of software could be due to the fact that Investigator A was assigned to a higher proportion of right coronary artery (RCA) procedures associated with angulations with high radiation exposure than Investigator B.

Fluoroscopy time was reduced by tendency by both investigators without reaching significance (Figure 4D, E). Long fluoroscopy time from Investigator B without the aid of software resulted from a few complex interventions with long fluoroscopy times, as indicated by a higher standard deviation. This was due to high calcification grade and therefore difficult introduction of interventional equipment with repeated exchange for conversion of material. Due to low radiation angulations in these cases, higher fluoroscopy time did not result in significantly higher dose area product use from Investigator B.

New interventional technology often comes along with training, learning curves, and increased complexity. For example, coronary hemodynamic measurement or intracoronary imaging require time and expertise to obtain useful results. However, DCR works in a different way: activated by 1 click on a touch screen, coronary angiographies are stored in an image library. As soon as images are needed for PCI, they are shown on the screen. The software is robust and minimal technical challenges have been described elsewhere.⁵ This description is supported by our results concerning procedural characteristics. Neither the high- nor the medium-experience level investigator prolonged procedure duration when using the DCR. Dose area product use was significantly reduced by Investigator A. Effect size of DCR was very mild with the selected endpoints (d between 0 and 0.2) except for a higher effect size for DCR (d= 0.26) concerning contrast medium use by Investigator A.

Contrast-induced acute kidney injury (CI-AKI) remains one of the most serious and challenging complications of PCI and necessitates process optimization to reduce contrast agent volume. Since treatment of acute kidney injury is limited to supportive measures, only prevention is highly desired. Different scores help to identify patients at risk for kidney injury.^13,14 Contrast agent-induced AKI after PCI has a nonlinear correlation to the contrast medium volume applied. Furthermore, incidence also varies heterogeneously along different baseline risk factors.¹⁵ Consequently, contrast medium applied during PCI needs to be reduced and should be kept to a minimum.¹⁶ DCR is able to achieve this goal irrespective of whether the investigator is operating at a medium or high experience level.

Efficacy and safety must be rigorously investigated as soon as new technology is integrated in our workflows. This applies to DCR, as the roadmap is used for coronary interventions. Looking at safety and efficacy results, we were able to show very encouraging data for both the medium- and the high-experience level investigator. This is in part related to the highly selected patient cohort with type A or B stenosis. However, the evolution of the DCR technology has reached a precision point, which supports safe and effective PCIs. Whether these excellent results can also be translated into more complex lesion settings such as bifurcation or CTO is yet to be evaluated in future trials.

Limitations. This is a highly selected, single-center patient population with low-risk coronary lesions and subsequent low event rates. The results of the present study are a sub-group analysis of the randomized prospective trial.⁶ Furthermore, there were only 2 operators, and neither was a low-volume operator, thereby limiting the generalizability of the results. Further research will be needed to evaluate the benefit of roadmap fusion imaging when used by low-volume investigators, or even trainee investigators under supervision, with potential impact on the learning curve.

CI-AKI remains one of the most challenging complications of PCI. Since treatment of AKI is limited to supportive measures only, process optimization to reduce contrast agent volume is highly needed. This study is the first to evaluate the impact of investigators’ experience level on the safety and efficacy of DCR during PCI. Designed to improve the procedural efficacy of PCI, DCR allows significantly lower use of contrast medium irrespective of experience level. With both medium and high-experience level investigators benefitting from the DCR imaging tool with excellent interventional results, this tool is advantageous to clinical routine.

From ¹Heinrich-Heine University, Medical Faculty, Department of Cardiology, Pulmonology and Vascular Medicine, Düsseldorf, Germany; ²CARID (Cardiovascular Research Institute Düsseldorf), Düsseldorf, Germany

Dr Quast and Dr Phinicarides contributed equally to the manuscript.

Disclosures: Drs Afzal, Polzin, and Zeus received lecture honoraria from Philips. The remaining authors report no financial relationships or conflicts of interest regarding the content herein.

Funding: Philips Healthcare (Philips Healthcare, Eindhoven, the Netherlands) and the University Hospital Düsseldorf have a master research agreement. This study was supported by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – (TRR259, Grant No. 397484323 to C.Q.), Forschungskommission of the Medical Faculty of the Heinrich Heine University (18-2019 to A.P. and 2021-30 to C.Q.), funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – (PE-2704/3-1 to A.P.; LE940/7-1 to A.P.); SFB 1116 - Grant No. 236177352- to A.P. (subproject B11), M.K. (subproject B12) and C.J. (subproject B06).

Address for correspondence: Tobias Zeus, MD, University Hospital Düsseldorf, Medical Faculty, Heinrich-Heine University Düsseldorf, Division of Cardiology, Pulmonology and Vascular Medicine, Moorenstr. 5, 40225 Düsseldorf, Germany. Email: zeus@med.uni-duesseldorf.de