Rationale and design of the precise percutaneous coronary intervention plan (P3) study: Prospective evaluation of a virtual computed tomography‐based percutaneous intervention planner

Abstract Introduction Fractional flow reserve (FFR) measured after percutaneous coronary intervention (PCI) has been identified as a surrogate marker for vessel related adverse events. FFR can be derived from standard coronary computed tomography angiography (CTA). Moreover, the FFR derived from coronary CTA (FFRCT) Planner is a tool that simulates PCI providing modeled FFRCT values after stenosis opening. Aim To validate the accuracy of the FFRCT Planner in predicting FFR after PCI with invasive FFR as a reference standard. Methods Prospective, international and multicenter study of patients with chronic coronary syndromes undergoing PCI. Patients will undergo coronary CTA with FFRCT prior to PCI. Combined morphological and functional evaluations with motorized FFR hyperemic pullbacks, and optical coherence tomography (OCT) will be performed before and after PCI. The FFRCT Planner will be applied by an independent core laboratory blinded to invasive data, replicating the invasive procedure. The primary objective is to assess the agreement between the predicted FFRCT post‐PCI derived from the Planner and invasive FFR. A total of 127 patients will be included in the study. Results Patient enrollment started in February 2019. Until December 2020, 100 patients have been included. Mean age was 64.1 ± 9.03, 76% were males and 24% diabetics. The target vessels for PCI were LAD 83%, LCX 6%, and RCA 11%. The final results are expected in 2021. Conclusion This study will determine the accuracy and precision of the FFRCT Planner to predict post‐PCI FFR in patients with chronic coronary syndromes undergoing percutaneous revascularization.

hyperemic pullbacks, and optical coherence tomography (OCT) will be performed before and after PCI. The FFR CT Planner will be applied by an independent core laboratory blinded to invasive data, replicating the invasive procedure. The primary objective is to assess the agreement between the predicted FFR CT post-PCI derived from the Planner and invasive FFR. A total of 127 patients will be included in the study.
Results: Patient enrollment started in February 2019. Until December 2020, 100 patients have been included. Mean age was 64.1 ± 9.03, 76% were males and 24% diabetics. The target vessels for PCI were LAD 83%, LCX 6%, and RCA 11%. The final results are expected in 2021.
Conclusion: This study will determine the accuracy and precision of the FFR CT Planner to predict post-PCI FFR in patients with chronic coronary syndromes undergoing percutaneous revascularization.

| INTRODUCTION
Fractional flow reserve (FFR), an invasive measurement of epicardial conductance during hyperemia, quantifies the amount of flow reduction due to epicardial narrowing and correates with myocardial ischemia. 1 Coronary flow can be restored by revascularization.
Percutaneous coronary intervention (PCI) is an effective method to improve myocardial perfusion and relieve patients from angina. [2][3][4] The International Study of Comparative Health Effectiveness With Medical and Invasive Approaches (ISCHEMIA) study confirmed the clinical benefit of revascularization in terms of relieve from angina compared with a conservative strategy; nonetheless, the invasive strategy showed no benefit concerning the occurrence of adverse cardiovascular events in stable patients at 3 years follow-up. 4 After a successful PCI, approximately one fourth of patients remain with impaired coronary flow. 5 The degree of functional revascularization can be assessed invasively by measuring FFR immediately after stent implantation. Complete functional revascularization (i.e. high post-PCI FFR) has been associated with improved clinical outcomes after PCI 6 whereas low post-PCI FFR has been identified as an independent predictor of vessel-related adverse events. 7 Therefore, a tool that predicts improvement in epicardial conductance would be of benefit for clinical decision making about revascularization and procedural planning.
Coronary computed tomography angiography (CTA) allows for the evaluation of coronary artery disease (CAD) in the non-invasive setting. 8 Coronary geometries derived from CTA can be utilized to perform blood flow simulation and estimate FFR. FFR derived from CT (FFR CT ) has been shown to be accurate compared with invasive FFR. 9 The FFR CT Planner (HeartFlow, Inc., Redwood City, CA) is a novel tool able to recompute FFR CT values after opening coronary stenoses. The Planner leverages the results of multiple simulations and reduced order modeling to instantly calculate a FFR CT value in the desired lumen configuration. This provides the benefit of anticipating the effect of PCI influencing treatment planning prior to the catheterization laboratory. 10 The hypothesis of the present study is that the FFR CT Planner would be accurate and precise in predicting the results of PCI in terms of coronary physiology. Thus, the present study aims to validate the performance of the FFR CT Planner to predict FFR post PCI with invasive FFR as a reference standard.

| Study design
The PRECISE PCI PLAN (P3) study is an investigator-initiated, prospective, multi-center study evaluating the diagnostic accuracy of FFR CT Planner. Patients with chronic coronary syndromes with invasive FFR≤0.80 in at least one vessel and guideline-directed indication for PCI will be eligible for inclusion. Table 1 shows inclusion and exclusion criteria. Prior to PCI, all patients will undergo coronary CTA with calculation of FFR CT . Invasive FFR will be followed by a motorized hyperaemic pullback evaluation; this will be performed before and repeated after PCI using a dedicated acquisition protocol. 11 In addition, optical coherence tomography (OCT) will be used to guide PCI and optimize stent implantation. PCI optimization either based on FFR pullbacks and/or OCT will be allowed at the discretion of the operators. Each participating center will undergo peer-to-peer review of coronary CTA, angiography, OCT, FFR and motorized FFR standards before initiation of study enrollment. The study protocol has been approved at each participating center by the local Ethics Committee. All study subjects provide written informed consent prior to undergoing any study-specific procedures. This study is registered as NCT03782688.
The study leadership is composed by a principal investigator, a co-principal investigator, a chairman and steering committee. Clinical events will be adjudicated by an independent clinical events committee. Imaging (i.e. coronary CTA, invasive coronary angiography and OCT) and physiology data (i.e. hyperaemic pressure tracings) will be analyzed by an independent core-laboratory. In addition, the quality of the coronary CTA images will be assessed by an independent CT quality committee. Details of each of these committees and their members are shown in the Supplemental Appendix.

| Primary and secondary endpoint
The primary endpoint is to assess the agreement between FFR CT Planner and invasively measured

| Study logistics
Coronary CTA and FFR CT will be performed as part of standard of care.
Once eligibility has been confirmed, patients will be invited to participate in the study. PCI will be performed following a dedicated protocol including combined invasive FFR and OCT evaluation for procedural guidance and stent optimization. Morphological and functional data will be centrally collected by the core laboratory (CoreAalst BV, Aalst, Belgium) for analysis. The FFR CT diagnostic model with the position of the stent, derived from the invasive coronary angiography, will be sent to the FFR CT core laboratory (HeartFlow Inc, Redwood city, California, US) to apply the FFR CT Planner blinded to the invasive data. All data will be centrally processed, analyzed and co-registered by the central core laboratory (Supplemental figure).

| Coronary CTA analysis
The coronary CTA will be performed using contemporary single-and dualsource CT scanners with a minimum of 128 detector rows and gantry rotation times <330 milliseconds. Laboratories will follow local CT acquisition protocols being in accordance with quality standards defined by the Society of Cardiovascular Computed Tomography. 12 Oral or intravenous betablockers will be administered to achieve heart rate ≤ 65 bpm. Before the scan, patients will receive nitrates to ensure coronary vasodilatation. The coronary CTA images will be transferred to an independent core laboratory for analysis. Non-invasive quantitative coronary analysis (NI-QCA) will be performed using the luminal dimensions from the FFR CT model. 13 Lesion length will be defined as the length between normal ( without plaque) proximal and distal reference segments. Minimal lumen area, proximal and distal reference lumen areas, and area stenosis will be assessed.
The quality of coronary CTA will be adjudicated using a four-points Likert score at the vessel level.

| FFR derived from CT
Coronary CTA datasets will be processed for FFR CT using a validated method (HeartFlow, Inc, Redwood City, California, USA). Briefly, models will be constructed from automated algorithms and trained analysts. Blood flow simulations will be performed on patient-specific coronary geometries to compute FFR CT values. 14 The FFR CT Planner will be applied blinded to the invasive functional data to remodel the lumen and provide a FFR CT value after stenosis removal ( Figure 1). For the primary endpoint, the FFR CT value matching the invasive pressure wire sensor position will be used. FFR CT values will be extracted at every 0.1 mm to create FFR CT pullback curves for co-registration with invasive motorized FFR pullbacks.

| Coronary angiography
Invasive coronary angiography will be performed following a dedicated protocol. Briefly, intracoronary nitroglycerin injection (100-200 μg) will be administered before angiography. At least two projections separated by at least 30 will be obtained before and after PCI. Stent position before implantation will be recorded in two projections facilitating co-registration with the FFR CT model. Coronary angiography will be analyzed with 3D-QCA software (CAAS 8.2 Software, Pie Medical Imaging, Maastricht, The Netherlands). In case of lesions involving coronary bifurcations dedicated QCA bifurcation software packages will be used. 15 Minimum lumen area, percent area stenosis, reference vessel areas and lesion length will be reported. QCA analysis will be performed by an independent core laboratory. Invasive FFR measurements will be performed pre-and post-PCI at the same anatomical location. In addition, FFR pullbacks will be performed using a motorized device (R 100, Volcano, San Diego, Ca, USA) at a speed of 1 mm per second. The standardization of the pressure and vessel length relationship will allow for co-registration of the pressure along the coronary vessel. Pressure tracings and pullback curve quality will be assessed by an independent core laboratory.

| Fractional flow reserve
Functional gain will be defined as the FFR post-PCI minus the FFRpre-PCI. Lesion gradients will be defined as the FFR values at the proximal edge of the stent minus the FFR at the distal edge of the stent. To quantify the pattern of coronary artery disease (e.g. focal or diffuse), the pressure pullback gradient (PPG) will be calculated. Analysis of invasive functional data will be performed using Coroflow software (Coroventis research, Uppsala, Sweden).

| Optical coherence tomography
Optical coherence tomography (Abbott Vascular, St. Paul, Minnesota) will be acquired pre-and post-PCI. OCT images will be used online to guide PCI and optimize stent implantation at operator discretion.
Lesion length will be based on normal-to-normal landing zones in the pre-PCI OCT. Minimum lumen area, reference lumen areas and lesion length will be analyzed. 16 Post PCI, minimum stent area (MSA) and stent expansion will be reported. Stent expansion will be defined as MSA in both the proximal and distal halves of the stent relative to the closest reference segment. OCT images will be analyzed by an independent core laboratory.. Figure 2 shows angiographic, FFR and OCT acquisition protocol.

| Co-registration
Coronary imaging data from coronary CTA, invasive angiography and OCT will be matched using fiduciary points. In addition, physiological data from invasive FFR and FFR CT pullbacks will be co-localized. Furthermore, morphological (i.e., coronary CTA, angiography and OCT) and physiologic (FFR CT and invasive FFR) will be co-registered. For co-registration, vessels will be divided in three segments (i.e., proximal, lesion and distal). The proximal segment will be defined as from the ostium to the proximal lesion edge. The lesion will be defined as the stented segment, and the distal segment will be defined from the distal stent edge to the position of the pressure sensor. Co-registration of invasive and non-invasive FFR pullbacks will be performed as previously described. 17 Co-registration between coronary CTA and OCT will be performed using a proprietary automated algorithm based on side branches location. . Figure 3 shows an example of the co-registration process. Case examples are shown in Figure 4.

| Clinical outcomes
Clinical follow-up will be performed in-hospital, at 6 ± 1 month and 1 year after the procedure. Peri-procedural myocardial infarction will be defined according to the Fourth Universal definition. 18 The rate of TVF and its components (cardiac death, target vessel myocardial infarction, ischemia driven target vessel revascularization) and stent thrombosis will be assessed at 6 months, 1 year and yearly until 5 years follow-up. Seattle angina questionnaire-7 items (SAQ-7) will be administrated at 6 to 12 months after PCI. Adverse events will be adjudicated by an independent clinical events committee.
F I G U R E 4 Legend on next page.

| Statistical analysis and sample size calculation
The P3 study will assess the agreement between the FFR CT Planner and invasively measured FFR after stent implantation. The agreement will be assessed using the Bland-Altman method. Mean difference or bias will be considered a metric of accuracy, and standard deviation as a metric of precision. Based on prior data, the mean difference between FFR CT Planner and invasive FFR post PCI is assumed to be 0.04 FFR units with a standard deviation of 0.07. 19 With these assumptions and confidence levels of 0.95 and power of 80% a sample size of 123 vessels would be required. Assuming an attrition rate of 2.5%, 127 vessels will be included. The P3 study is not statistically powered to assess its secondary endpoints.

| Planned sub-studies
In addition to the primary and secondary objectives, several sub-studies are planned. Briefly, we will (1) assess the accuracy of the FFR CT Planner stratified by coronary CTA image quality; (2) assess the relationship between the pre-PCI pattern of functional coronary artery disease (e.g. focal or diffuse) quantified by the PPG and post-PCI FFR using both invasive and non-invasive FFR pullbacks; (3) assess the relationship between luminal dimensions derived from coronary CTA and OCT, and (4) we planned to assess the relationship between plaque type and trans-lesional pressure gradients using coronary CTA and iOCT.

| RESULTS
Patient enrollment started in February 2019, until December 2020, 100 patients have been included. Mean age was 64.1 ± 9.03, 76% were males and 24% diabetics. The target vessels for PCI were LAD 83%, LCX 6% and RCA 11%. Table 2 shows the baseline characteristics of the population included. Recruitment is ongoing and it is anticipated that the primary results will be presented in Summer 2021.

| DISCUSSION
The present study will assess the accuracy and precision of the The combination of coronary CTA with its ability to assess atherosclerotic plaque extent, and FFR CT allowing to assess pressure losses along the coronary vessel provides a unique opportunity to evaluate the anatomical and functional CAD patterns. Two predominant phenotypes of coronary artery (i.e. focal or diffuse) have been described. 23,24 In cases of focal functional CAD with lesion-specific ischemia, PCI often results in complete functional revascularization. In contrary, in cases of diffuse CAD, no focal pressure gradients are present despite the presence of one or more angiographic stenoses. In the latter, PCI results are frequently sub-optimal in terms of post-PCI FFR whereas in the former PCI restore epicardial conductance.
FFR CTP lanner will facilitate the integration of a comprenhensive functional assessement for PCI planning.
At variance with diagnostic FFR CT , the FFR CT Planner is designed to be used in patients with significant CAD. The FFR CT Planner could be used at two phases of the evaluation of patients with CAD. First, to determine the suitability for percutaneous revascularization.
Patients with diffuse disease, for example, showing negligible functional improvement with PCI could be informed of the anticipated results, the likelihood of angina relief and other more suited therapeutics options. Second, in the catheterization laboratory, the FFR CT Planner may help in assessing the location of pressure losses, length of coronary segments to be treated to achieve optimal post-PCI FFR values and tailor your PCI strategy in cases of serial lesions. By this, the FFR CT Planner has the potential to increase the degree of complete functional revascularization.

| Limitations
The present study has several limitations. First, the evaluation of the performance of the FFR CT Planner is based on its accuracy compared to invasively measured FFR. Post-PCI FFR is a surrogate of clinical outcomes after percutaneous revascularization. The study, however, is not powered to assess clinical outcomes. Second, the information of the FFR CT planner will not be used to guide PCI; thus, the clinical impact of the prospective application of this technology remains to be determined. Third, the sample size is relatively small; nevertheless, powered to assess the accuracy and precision of the FFR CT planner.
Fourth, stent optimization will be guided by FFR pullbacks and OCT which may not represent routine clinical practice.

| CONCLUSION
This prospective and multicenter study will determine the accuracy and precision of the FFR CT Planner to predict post-PCI FFR. Prediction of post-PCI FFR may improve patient selection for percutaneous revascularization, anticipate the clinical benefit of the intervention and refine the revascularization strategy. A larger clinical trial will be required to assess the impact of the FFR CT Planner guided strategy on clinical outcomes.