Author + information
- Received December 15, 2016
- Revision received March 19, 2017
- Accepted March 23, 2017
- Published online July 3, 2017.
- Ehtisham Mahmud, MD∗ (, )
- Jesse Naghi, MD,
- Lawrence Ang, MD,
- Jonathan Harrison, MD,
- Omid Behnamfar, MD,
- Ali Pourdjabbar, MD,
- Ryan Reeves, MD and
- Mitul Patel, MD
- Division of Cardiovascular Medicine, University of California, San Diego, Sulpizio Cardiovascular Center, La Jolla, California
- ↵∗Address for correspondence:
Dr. Ehtisham Mahmud, UC San Diego Sulpizio Cardiovascular Center, 9434 Medical Center Drive, La Jolla, California 92037.
Objectives The aims of this study were to evaluate the feasibility and technical success of robotically assisted percutaneous coronary intervention (R-PCI) for the treatment of coronary artery disease (CAD) in clinical practice, especially in complex lesions, and to determine the safety and clinical success of R-PCI compared with manual percutaneous coronary intervention (M-PCI).
Background R-PCI is safe and feasible for simple coronary lesions. The utility of R-PCI for complex coronary lesions is unknown.
Methods All consecutive PCI procedures performed robotically (study group) or manually (control group) over 18 months were included. R-PCI technical success, defined as the completion of the procedure robotically or with partial manual assistance and without a major adverse cardiovascular event, was determined. Procedures ineligible for R-PCI (i.e., atherectomy, planned 2-stent strategy for bifurcation lesion, chronic total occlusion requiring hybrid approach) were excluded for analysis from the M-PCI group. Clinical success, defined as completion of the PCI procedure without a major adverse cardiovascular event, procedure time, stent use, and fluoroscopy time were compared between groups.
Results A total of 315 patients (mean age 67.7 ± 11.8 years; 78% men) underwent 334 PCI procedures (108 R-PCIs, 157 lesions, 78.3% type B2/C; 226 M-PCIs, 336 lesions, 68.8% type B2/C). Technical success with R-PCI was 91.7% (rate of manual assistance 11.1%, rate of manual conversion 7.4%, rate of major adverse cardiovascular events 0.93%). Clinical success (99.1% with R-PCI vs. 99.1% with M-PCI; p = 1.00), stent use (stents per procedure 1.59 ± 0.79 with R-PCI vs. 1.54 ± 0.75 with M-PCI; p = 0.73), and fluoroscopy time (18.2 ± 10.4 min with R-PCI vs. 19.2 ± 11.4 min with M-PCI; p = 0.39) were similar between the groups, although procedure time was longer in the R-PCI group (44:30 ± 26:04 min:s vs. 36:34 ± 23:03 min:s; p = 0.002). Propensity-matched analysis confirmed that procedure time was longer in the robotic group (42:59 ± 26:14 min:s with R-PCI vs. 34:01 ± 17:14 min:s with M-PCI; p = 0.007), although clinical success remained similar (98.8% with R-PCI vs. 100% with M-PCI; p = 1.00).
Conclusions This study demonstrates the feasibility, safety, and high technical success of R-PCI for the treatment of complex coronary disease. Furthermore, comparable clinical outcomes, without an adverse effect on stent use or fluoroscopy time, were observed with R-PCI and M-PCI.
Percutaneous coronary intervention (PCI) has evolved tremendously since its inception, with refinement in pharmacotherapy and improvement in interventional devices. However, the fundamental technique of manually advancing intracoronary guidewires, balloons, and stents at the patient’s tableside in relative close proximity to the x-ray radiation source, while wearing heavy lead aprons, remains largely unchanged (1,2). As a result, occupational hazards inherent to performing the PCI procedure expose the operator to both orthopedic and radiation-related adverse effects (3–5).
Robotic PCI (R-PCI) (CorPath 200, Corindus, Waltham, Massachusetts) can potentially mitigate both the orthopedic and radiation-related occupational hazards associated with the practice of interventional cardiology. The ability to perform remote-controlled R-PCI was initially described by Beyar et al. (6), followed by small in-human feasibility studies (7). Subsequently, the PRECISE (Percutaneous Robotically Enhanced Coronary Intervention) trial demonstrated the safety and feasibility of R-PCI in a multicenter registry study of 164 patients (8) without an increase in patient radiation or contrast use (9,10). However, in that trial, the majority of lesions treated were simple (mean lesion length 12.2 ± 4.8 mm, only 12.8% American College of Cardiology/American Heart Association type C lesions) and not reflective of clinical practice. Although there have been reports of complex cases treated with R-PCI (11,12), no systematic evaluation of R-PCI for complex coronary anatomy has been performed. Hence, we designed this study to 1) evaluate the feasibility and technical success of R-PCI for the treatment of coronary artery disease in clinical practice, especially in complex lesions; and 2) determine the safety and clinical success of R-PCI compared with manual PCI (M-PCI).
The study protocol was approved by the University of California, San Diego, Human Subjects Protection Program and designed by the investigators (E.M., J.N., J.H., R.R., M.P.), who verify the authenticity of the data, and all authors participated in drafting the manuscript. Consecutive PCI procedures performed robotically or manually over 18 months by a single operator were included.
The robotic system consists of an interventional cockpit and a robotic arm mounted on the catheterization bedside rail (8). This robotic arm contains a drive housing a single-use sterile cassette, which is connected to the guiding catheter after manually engaging the coronary artery. The interventional cockpit is located within the cardiac catheterization laboratory and is connected via cables to the bedside drive (Figure 1). It contains monitors that display the live fluoroscopic image and hemodynamic data. The robotic system enables the operator to remotely advance and retract rapid exchange balloons and stents. Additionally, the operator can rotate and advance the guidewire, transmitting torque and permitting guidewire manipulation. Passive control of the guiding catheter is possible with guidewire and balloon manipulation. The fluoroscopy and cine pedal is controlled by the seated primary operator, and contrast injection is performed by the tableside team. Once streamlined, the robotic setup takes approximately 5 min and is performed simultaneously as preparation for ad hoc PCI is undertaken or as a part of the room setup for planned interventions. The robotic system has a capital cost and a disposable cassette cost of approximately $900 per procedure (capital cost depreciated over 6 years with an estimated R-PCI volume of 250 procedures per year).
Data for R-PCI (study group) were prospectively collected as part of the ongoing PRECISION (Post-Market CorPath Registry on the CorPath 200 System in Percutaneous Coronary Interventions) registry (NCT01917682), and all R-PCI procedures at this center since the introduction of the technology constituted the study group. The PRECISION registry is a post-market, prospective, single-arm, multicenter registry collecting data on the use, safety, and effectiveness of the CorPath 200 system for PCI procedures. All patients participating in this study must be ≥18 years of age, have coronary artery disease revascularized with robotically assisted PCI using the CorPath 200 system, and voluntarily agree to participate in the study after providing informed consent.
Data for M-PCI (control group) were simultaneously collected as a part of the National Cardiovascular Data Registry CathPCI registry. The CathPCI registry is an initiative of the American College of Cardiology Foundation and the Society for Cardiovascular Angiography and Interventions. Details of the CathPCI registry have been previously described (13). The registry collects demographic, clinical, and procedural data elements for consecutive PCI procedures at each participating center.
Inclusion and exclusion criteria
This was an all-comers study, and patients presenting to the catheterization laboratory for elective or urgent angiography found to have 1 or more lesions amenable to percutaneous revascularization were included. Patients with complex coronary lesions, including mild to moderate calcification, chronic total occlusion, bifurcation disease, severe tortuosity, unprotected left main coronary artery stenosis, and multivessel disease or left ventricular dysfunction with or without hemodynamic support, were enrolled. Of note, at the time of this study, only 1 of the 4 catheterization laboratories at our institution was equipped with the Corindus CorPath 200 robotic system. There were no selection criteria for allocating specific cases for diagnostic angiography in the robotically equipped room. Procedures deemed ineligible for R-PCI because of mechanical limitations of the current-generation device (any over-the-wire device requirement or planned bifurcation stenting) or ST-segment elevation myocardial infarction (MI) presentation were excluded for analysis from the M-PCI group.
Quantitative coronary angiography
Blinded quantitative coronary angiography was performed before and after PCI (AGFA, Mortsel, Belgium) for each treated lesion to define lesion length, minimal luminal diameter, reference vessel diameter, and percentage stenosis. American College of Cardiology/American Heart Association lesion classification was determined for each lesion. A lesion complexity score to determine the extent and degree of disease was calculated as sum of all discrete lesions treated (1 = type A, 2 = type B1, 3 = type B2, and 4 = type C) during each procedure. The cumulative score for treated lesions was categorized as low (≤4), intermediate (5 or 6), or high (>6) complexity.
Endpoints and definitions
Clinical success was defined as angiographic success (residual stenosis after stenting of <30% with final TIMI [Thrombolysis In Myocardial Infarction] flow grade 3) without an in-hospital major adverse cardiovascular event (MACE) (death, MI, stroke, or urgent repeat revascularization). MI was defined as either creatine kinase (CK)-MB >5 times the upper limit of normal (ULN) with evidence of a myocardial injury or asymptomatic CK-MB >10 times ULN per the Society for Cardiovascular Angiography and Interventions definition of periprocedural MI (14). In patients who presented with acute coronary syndromes and elevated biomarkers of ischemic injury before cardiac catheterization, a new absolute rise of 5 times (symptomatic) or 10 times (asymptomatic) ULN from the nadir in markers of ischemic injury was required to be adjudicated as a procedure-related MI. An analysis using the universal definition of MI after PCI (CK-MB >3 times ULN) is also reported. Cardiac markers, including CK and CK-MB, were uniformly measured post-PCI in all subjects.
Robotic technical success was defined as clinical success and the completion of the PCI procedure entirely robotically or with partial manual assistance (MA). MA was defined as temporary disengagement of the robotic drive to use bedside manipulation of either the guide catheter, guidewire, or delivery system, with ultimate completion of the procedure using the re-engaged robotic drive. Manual conversion (MC) was defined as the disengagement of the robotic drive to use bedside manipulation of either the guide catheter, guidewire, or delivery system, which was required until the end of the procedure.
Comparisons of clinical success, stent use, procedure time (PT), contrast use, and patient radiation exposure between R-PCI and M-PCI groups are also reported. PT was defined as time recorded from initial guidewire insertion until guide catheter disengagement.
Data were analyzed using SPSS version 14.0 (IBM, Chicago, Illinois) or SAS version 9.4 (SAS Institute, Cary, North Carolina). Continuous variable results are reported as mean ± SD and were compared between groups using the Wilcoxon rank sum test. Pearson chi-square or Fisher exact tests were used to compare categorical measurements as appropriate. A p value <0.05 was considered to indicate statistical significance. Comparisons were made in the full group and the propensity score–matched samples. Logistic regression was used to estimate the propensity score. Factors entered in the propensity score analysis included age, sex, diabetes mellitus, hypertension, presentation with an acute coronary syndrome, history of angina, previous coronary artery bypass grafting, history of congestive heart failure, history of peripheral artery disease, primary lesion stenosis and length, proportion of type B2/C lesions, lesion complexity score, SYNTAX (Synergy Between PCI With Taxus and Cardiac Surgery) score, and the number of lesions treated. A nearest-neighbor greedy match technique with caliper size one-quarter the SD (caliper = 0.0452) was used to match each R-PCI patient to the nearest M-PCI patient as long as the scores differed by no more than the caliper (15). Matching was conducted without replacement.
During the study period, a total of 413 individual PCI procedures (108 R-PCIs, 305 M-PCIs) were performed. Among the M-PCI group, 79 cases (24%) were excluded because of the presence of pre-determined exclusion criteria (Figure 2). Among the R-PCI cohort, the procedure was completed entirely robotically in 81.5% of the patients. In this cohort, MA and MC rates of 11.1% and 7.4%, respectively, were observed, with a single procedure-related MI, resulting in a technical success rate of 91.7% (Figure 3). Reasons for MA or MC were categorized as follows: 1) adverse event (n = 3); 2) technical limitation of the robotic platform (n = 8); or 3) limited guidewire or catheter support (n = 9).
For the 334 PCI procedures constituting the study cohort, similar baseline demographics were observed in the robotic and manually treated groups (108 R-PCIs, 157 lesions; 226 M-PCIs, 336 lesions) (mean age 68 ± 11 years vs. 67 ± 12 years; p = 0.91; male sex 78% vs. 78%; p = 0.98; acute coronary syndrome presentation 20.4% vs. 17.3%; p = 0.54 [R-PCI vs. M-PCI, respectively]) (Table 1). Angiographic analysis revealed that the 2 groups were similar with regard to target vessel treated and primary lesion stenosis, but the R-PCI group had longer primary lesion length, more type B2/C lesions, and higher SYNTAX score (Table 2). Overall treated lesion complexity score was similar between the 2 groups (5.03 ± 2.26 with R-PCI vs. 4.91 ± 2.67 with M-PCI; p = 0.40).
The rate of in-hospital MACE was low in both groups, with 2 clinically significant periprocedural MIs in the manual group and only 1 in the robotic group, resulting in comparable clinical success (99.1% with R-PCI vs. 99.1% with M-PCI; p = 1.00) (Table 3). The efficacy endpoint of freedom from MACE using the universal definition of periprocedural MI also demonstrated comparable success with both approaches (94.4% with R-PCI vs. 92.5% with M-PCI; p = 0.51) (Table 3).
No significant procedural differences were observed in the 2 groups, including the number of lesions treated, stents per case, and fluoroscopy time. Patient radiation exposure (dose-area product) and total contrast volume used were lower in the R-PCI group (Table 3). A significantly higher PT was observed in the R-PCI group (44:30 ± 26:04 min:s vs. 36:34 ± 23:03 min:s; p = 0.005), which remained significant after multivariate analysis controlling for SYNTAX score, primary lesion length, and lesion complexity (p = 0.026). However, the increase in PT in the R-PCI group was observed only for the procedures with low lesion complexity (low complexity 39:45 ± 26:48 min:s vs. 28:04 ± 15:36 min:s; p < 0.001; intermediate complexity 43:47 ± 18:09 min:s vs. 40:06 ± 15:09 min:s; p = 0.53; high complexity 56:27 ± 23:35 min:s vs. 57:37 ± 28:44 min:s; p = 0.83 [R-PCI vs. M-PCI, respectively]) (Figure 4).
Propensity-matched analysis showed similar baseline demographic, clinical, and angiographic characteristics in the 2 groups. Comparisons of procedural endpoints in the propensity-matched subgroup confirmed the primary findings of the study. PT remained longer in the robotic group (42:59 ± 26:14 min:s vs. 34:01 ± 17:14 min:s; p = 0.007), whereas fluoroscopy time, patient radiation exposure, MACE rate, and contrast volume were not different in the propensity-matched sample (Table 4).
This study demonstrates the safety and feasibility of performing R-PCI in a cohort of patients with complex coronary anatomy and significant comorbidities. Technical success with R-PCI was 91.7%, with only 7.4% of the procedures requiring MC. Although the access site used in both groups was predominantly femoral, no difference in technical success was observed in the radial access group. Hence, in this series of complex patients presenting for revascularization, the feasibility of R-PCI as a primary strategy is shown with a concomitant high rate of clinical success. The clinical outcomes achievable with R-PCI were comparable with M-PCI. R-PCI required slight prolongation of PT that was seen only with relatively simple coronary lesions, although no adverse effect on stent use, contrast use, or patient radiation exposure was observed.
Prior studies have demonstrated the utility of R-PCI in subjects with relatively simple lesions (8–10,16). The pivotal PRECISE study was limited to short non–type C lesions treatable with a single stent. In the present study, a substantial number of patients presenting with acute coronary syndromes, and with comorbidities including chronic kidney disease, prior coronary bypass surgery, diabetes mellitus, and left ventricular dysfunction, representative of typical patients treated at a tertiary interventional cardiology center, were treated with R-PCI. Almost 80% of the robotically treated lesions were type B2/C, including chronic total occlusions, bifurcation lesions, severely tortuous lesions, and unprotected left main coronary artery disease.
Among the small subset of patients requiring MA or MC, 3 procedures were categorized as being due to adverse events and were during the early learning curve. They were either coronary dissection after pre-dilation (n = 2) or abrupt vessel closure during advancement of a guidewire (n = 1) and were successfully treated with manual stent placement. The 2 major reasons for MA or MC were inadequate guide catheter support and limited ability to advance 2 balloons or stents simultaneously with the robotic drive in cases of provisional bifurcation stenting or final kissing balloon angioplasty for bifurcation lesion treatment. Future iterations of the technology will require addressing these limitations, including developing the potential for over-the-wire device delivery, which would enable atherectomy and antegrade wire escalation for chronic total occlusions. The second-generation CorPath GRX system has recently received U.S. Food and Drug Administration approval and has a third joystick with the ability to robotically control the guide catheter. This system should help improve robotic guide catheter support and decrease the number of procedures requiring MA or MC, but it has not been studied clinically. The successful use of R-PCI for patients with ST-segment elevation MI has been described (11), but we intentionally excluded those patients from this study for the potential deleterious effect of the robotic setup on door-to-balloon time.
Clinical success with R-PCI was high and comparable with M-PCI, with no significant differences in the MACE rate. Additionally, given the potential clinical significance of asymptomatic periprocedural biomarkers of ischemic injury elevation, an analysis using the universal definition of a periprocedural MI (i.e., CK-MB >3 times ULN) as an endpoint revealed no significant difference between MACE in the R-PCI and M-PCI groups. When comparing the 2 strategies for coronary revascularization, R-PCI proved to be safe, without evidence of any subclinical harm compared with M-PCI.
In addition to determining the feasibility and safety of R-PCI in more complex anatomy, the objective of this study was to evaluate the efficiency of R-PCI. This study of a large consecutive series of patients undergoing PCI by a single operator allowed direct measurement of the time efficiency of R-PCI compared with M-PCI. As a surrogate for overall procedural efficiency, PT was measured in each case and directly compared between groups. In the overall cohort, measurement of PT yielded a result that favored the M-PCI group. This was not altogether surprising given that aside from the time invested in robotic drive setup, loading the robotic drive with guidewires and balloon catheters throughout the case takes more time. The experience with R-PCI was new and included early procedures in which limitations of R-PCI were still not appreciated, which could have led to longer robotic PT (16). Despite the learning curve of R-PCI, when PT was stratified on the basis of a lesion complexity score, only the low-complexity group had significantly longer PT with R-PCI. We believe that this finding represents the initial nonadjustable investment of time in all R-PCI that is not seen in comparable M-PCI procedures. In a low-complexity procedure, this invested time would account for a substantial portion of the overall PT, whereas this would be diluted by a lengthier overall PT in the intermediate- and high-complexity groups. This finding requires further investigation in future studies.
Using R-PCI, we observed that the number of stents used per case, fluoroscopy time, and patient radiation dose were not adversely affected compared with M-PCI. No prolongation of fluoroscopy or overall patient radiation exposure was observed compared with M-PCI. Although contrast use and patient radiation dose were lower in the R-PCI group, this finding was not confirmed after multivariate or propensity-matched analyses. A cost-effectiveness analysis is beyond the scope of this study, but the benefits of R-PCI (reduction of operator radiation exposure and potentially limiting orthopedic injuries) need to be considered in context of the incremental cost of this new technology.
The study was performed at a single center, with robotic outcomes reported for a single high-volume operator, and therefore the results require replication in a multicenter analysis of R-PCI for complex coronary lesions. Furthermore, the study included procedures performed very early during the experience with R-PCI, and therefore the learning curve associated with this new technique could have negatively affected the results for R-PCI. This was not a randomized comparison of R-PCI versus M-PCI, and the study findings need to be interpreted in the context of a potential bias in favor of the robotic cohort; however, after excluding a pre-specified group of patients from the M-PCI group, the 2 groups were well matched from a baseline demographic and angiographic standpoint. The R-PCI cohort was similar if not more complex in terms of SYNTAX score, presentation for acute coronary syndrome, and proportion of type B2/C lesions treated, which could lead to favorable outcomes in the M-PCI group. Nevertheless, as this was an all-comers inclusive study, the findings are more widely applicable and not limited to carefully selected patients, as in the PRECISE trial or in a randomized controlled trial. Because almost a quarter of the patients treated manually were excluded from this study, the presented findings cannot be extrapolated to calcified lesions requiring atherectomy, chronic total occlusions treated with the hybrid approach, planned bifurcation stenting, or patients with ST-segment elevation MI. Future iterative improvements in robotic technology are required to enable the treatment of these patient subsets.
This study demonstrates the feasibility, safety, and high technical success of R-PCI for the treatment of complex coronary disease. Furthermore, comparable clinical outcomes, without an adverse effect on stent use or patient radiation exposure, were observed with R-PCI and M-PCI.
WHAT IS KNOWN? The feasibility of R-PCI has been demonstrated for simple coronary lesions. However, applicability of robotic technology for PCI in clinical practice and for complex coronary lesions is unknown.
WHAT IS NEW? This study demonstrates the feasibility, safety, and high technical success of R-PCI for the treatment of complex coronary disease, with clinical outcomes comparable with M-PCI.
WHAT IS NEXT? Future studies need to focus on evaluating potential economic and patient-level benefits with this new technology.
Dr. Mahmud has received consulting fees and research support from Corindus. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- creatine kinase
- manual assistance
- major adverse cardiovascular event(s)
- manual conversion
- myocardial infarction
- manual percutaneous coronary intervention
- percutaneous coronary intervention
- procedure time
- robotic percutaneous coronary intervention
- upper limit of normal
- Received December 15, 2016.
- Revision received March 19, 2017.
- Accepted March 23, 2017.
- 2017 American College of Cardiology Foundation
- Goldstein J.A.,
- Balter S.,
- Cowley M.,
- Hodgson J.,
- Klein L.W.,
- for the Interventional Committee of the Society of Cardiovascular Interventions
- Beyar R.,
- Gruberg L.,
- Deleanu D.,
- et al.
- Weisz G.,
- Metzger D.C.,
- Caputo R.P.,
- et al.
- Mahmud E.,
- Dominguez A.,
- Bahadorani J.
- Moussa I.,
- Hermann A.,
- Messenger J.C.,
- et al.
- Moussa I.D.,
- Klein L.W.,
- Shah B.,
- et al.
- Weisz G.,
- Smilowitz N.R.,
- Metzger D.C.,
- et al.