Cardiovascular Revascularization Medicine, 2018-07-01, Volume 19, Issue 5, Pages 526-531, Copyright © 2017 Elsevier Inc. Abstract Background Robotically-assisted percutaneous coronary intervention (R-PCI) is feasible for simple coronary lesions. Objectives To determine the frequency and reasons for partial manual assistance or manual conversion during R-PCI in clinical practice. Methods The CorPath 200 System (Corindus, Waltham, MA) enables the operator to sit in a radiation-shielded cockpit and remotely control intracoronary devices including guidewires, balloons, and stents. Consecutive R-PCI procedures performed over 18 months were analyzed to identify reasons for planned or unplanned manual assistance or manual conversion, and categorized as due to 1) adverse event; 2) technical limitation of the robotic platform; or 3) limited guide catheter/wire support. Results During the study period, 108 R-PCI procedures (68.1 ± 11.0 years, 77.8% men, 69.4% elective PCI, 78.3% type B2/C lesions, and 50.3% left anterior descending/left main target lesion segment) were performed. High robotic technical success (91.7%) and clinical procedural success (99.1%) were achieved. Twenty procedures (18.5%) required either planned partial manual assistance (3.7%), unplanned partial manual assistance (7.4%), or manual conversion (7.4%). Among these procedures, manual assistance/conversion was required in 3 procedures for an adverse event (15%), 8 for technical limitation of the robotic platform (40%), and 9 for guide catheter/wire support issues (45%). Conclusions High clinical success with R-PCI for a complex lesion cohort is possible with only occasional partial manual assistance or manual conversion. The majority of procedures requiring manual assistance/conversion were due to limited guide catheter/wire support or robotic platform limitations, rather than occurrence of adverse events. Highlights High clinical success with R-PCI for a complex lesion cohort is possible. However, occasional partial manual assistance or manual conversion is required. Partial manual assistance/conversion was due to guide catheter/wire support. Partial manual assistance/conversion was due to robotic platform limitation. 1 Introduction Tremendous advances in adjunctive pharmacotherapy and device technology have been made since the initial description of percutaneous coronary intervention (PCI), but the fundamental technique of operators manually advancing intracoronary devices at the patient's tableside while wearing heavy lead aprons in close proximity to an X-ray radiation source remains largely unchanged. The heavy lead apron worn by cardiovascular interventionalists is associated with orthopedic complications while significant radiation exposure is an additional occupational hazard. Robotically-assisted PCI (R-PCI), allowing operators to remotely manipulate intracoronary devices without lead-apron protection, can potentially limit the orthopedic and radiation-associated risks. The PRECISE (Percutaneous Robotically Enhanced Coronary Intervention) trial, in a large multicenter study consisting of 164 patients, demonstrated the safety and feasibility of robotic-assisted PCI in simple coronary lesions with approximately 1% of the cases requiring any manual assistance. PCI in this study was primarily performed for simple lesions (12.3% type C), and 99% of the procedures were completed entirely robotically. The current study sought to determine the frequency of partial manual assistance or manual conversion during R-PCI, and identify the associated reasons for assistance or conversion, among subjects treated in clinical practice and with more complex coronary anatomy. 2 Methods This study was approved by the Institutional Review Board of the University of California, San Diego. Data for all patients enrolled in the PRECISION registry ( ClinicalTrials.gov Identifier: NCT01917682 ) at UC San Diego were queried. The PRECISION registry is a post-market, prospective, single-arm, multi-center registry collecting data on the use, safety and effectiveness of the CorPath 200 Vascular Robotic System for patients undergoing PCI procedures. All patients participating in this study were ≥ 18 years of age, underwent R-PCI using the CorPath200 system, and voluntarily agreed to participate in the study after providing informed consent. The occurrence and reason for either partial manual assistance (planned or unplanned) or manual conversion were recorded and categorized as being due to 1) an adverse event, 2) technical limitation of the robotic platform (e.g. requirement for two balloons or stents to be inflated together, need for any over-the-wire equipment), or 3) limited guide catheter/wire support. Baseline subject characteristics, procedural details, and laboratory data were routinely measured with post-PCI creatinine phosphokinase myocardial band (CK-MB) and total creatine phosphokinase (CPK) collected every 8 h in all subjects. 2.1 CorPath 200 vascular robotic system The CorPath 200 system has been previously described . Briefly, it consists of an interventional cockpit and a robotic arm mounted on the catheterization bedside rail ( Fig. 1 ). This robotic arm contains a drive housing a single-use sterile cassette, which is connected to the guiding catheter after manual engagement of the target coronary vessel. The interventional cockpit is located within the cardiac catheterization laboratory and is connected via cables to the bedside drive. Monitors displaying the live fluoroscopic images and hemodynamic data are mounted within the cockpit. Controls allow the operator to remotely advance, retract, and rotate a 0.014-inch guidewire. Additionally, rapid-exchange balloons and stents can be remotely advanced and retracted. Fig. 1 Robotic PCI platform. (A) Robotic console and tableside drive (CorPath 200, Corindus, Waltham, MA) with robotic cassette (inset). 2.2 Key definitions Clinical success was defined as R-PCI completion (final flow TIMI 3 and residual stenosis < 30%) without an in-hospital major adverse cardiovascular event (MACE: myocardial infarction, urgent target vessel revascularization, emergent coronary artery bypass grafting, all-cause death). Clinically-relevant post-PCI myocardial infarction was defined as rise in CK-MB > 5 × upper limit of normal (ULN) with evidence of myocardial injury or an asymptomatic CK-MB > 10 × ULN . Data for the universal definition of myocardial infarction (CK-MB > 3 × ULN) were also collected. Robotic technical success was defined as clinical success and the completion of the PCI procedure entirely robotically or with partial manual assistance in the absence of MACE. Manual assistance was defined as temporary disengagement of the robotic drive in order to utilize bedside manipulation of either the guide catheter or wire, with ultimate completion of the procedure utilizing the re-engaged robotic drive. Planned manual assistance was defined as anticipated temporary disengagement of the robotic drive in order to utilize bedside manipulation of either the guide catheter or wire, with ultimate completion of the procedure utilizing the re-engaged robotic drive (e.g. deployment and retrieval of a distal embolic protection device during vein graft PCI). Unplanned manual assistance was defined as unanticipated temporary disengagement of the guide catheter or wire, with ultimate completion of the procedure utilizing the re-engaged robotic drive. Manual conversion was defined as the disengagement of the robotic drive in order to utilize bedside manipulation of either the guide catheter, wire, or stent, which was required until the end of the procedure. An example of such a case is presented ( Figs. 2 and 3 ). Fig. 2 Baseline coronary angiography of a high complexity robotic PCI procedure. Procedure performed robotically showing pre PCI baseline angiogram of an 83 year old woman presenting with unstable angina with (A–C) ostial left circumflex stenosis (yellow arrow) and ostial left anterior descending stenosis (blue arrow); (D) ostial/proximal right coronary artery stenosis (double yellow arrows) and with severe left ventricular dysfunction (ejection fraction 27%). Fig. 3 Coronary angiography after robotic PCI with Impella hemodynamic support. Final angiogram after (A, B) proximal right coronary artery stenting; (C–E) sequential left circumflex and left anterior descending artery stenting and final kissing balloon angioplasty (yellow arrows) in the presence of percutaneous hemodynamic support with the Impella (Abiomed, Danvers, MA) device (red arrow). Given that the patient needed bifurcation stenting with simultaneous kissing stent deployment, the terminal portion of the procedure was completed manually. The current robotic platform (CorPath 200) does not allow the advancement or maintained control of two stents at the same time with the robotic system. 3 Results During the 18-month study period, a total of 108 R-PCI procedures (157 lesions) were performed. Subjects undergoing R-PCI had a high prevalence of diabetes mellitus, hypertension, dyslipidemia, prior myocardial infarction, and prior coronary artery revascularization ( Table 1 ). A high proportion of complex coronary lesions (78.3% type B2/C lesions) were treated within this cohort ( Table 2 ) predominantly via femoral arterial access (88.0%). Overall subject characteristics were similar between procedures completed entirely robotically ( n = 88, 81.5%) versus those utilizing any manual aspect. Table 1 Baseline subject characteristics for robotic PCI cohort. VariableOverall ( N = 108)Complete R-PCI ( n = 88)Any manual assistance ( N = 20) P -value Age 68.1 ± 11.0 years 68.1 ± 11.3 years 68.2 ± 10.1 years 0.983 Male gender 77.8% 73.9% 95.0% 0.041 Diabetes mellitus 55.6% 54.5% 60.0% 0.658 Hypertension 95.4% 95.5% 95.0% 1.00 Hyperlipidemia 97.2% 96.6% 100% 1.00 Chronic kidney disease 20.4% 22.7% 10.0% 0.355 Prior myocardial infarction 52.8% 52.3% 55.0% 0.825 Prior PCI 45.4% 46.6% 40.0% 0.593 Prior CABG 11.1% 10.2% 15.0% 0.692 Ejection fraction 57.5 ± 11.1% 58.4 ± 10.5% 53.4 ± 13.7% 0.077 PCI Indication 0.467 ACS 20.4% 21.6% 15.0% Unstable angina 10.2% 11.4% 5.0% Elective 69.4% 67.0% 80.0% Table 2 Angiographic and procedural characteristics of the robotic PCI cohort. VariableOverall ( N = 108)Complete R-PCI ( n = 88)Any manual assistance ( n = 20) P -value Radial artery access 12.0% 12.5% 10.0% 1.00 Primary target vessel 0.591 Left main segment 4.6% 5.7% 0% Left anterior descending 46.3% 46.6% 45.0% Left circumflex 20.4% 21.6% 15.0% Right coronary artery 26.9% 25.0% 35.0% Other 1.9% 1.1% 5.0% Overall lesion classification ( N = 157) 0.470 Type A 1.9% 1.7% 2.6% Type B1 19.7% 19.3% 21.1% Type B2 8.9% 10.9% 2.6% Type C 69.4% 68.1% 73.7% Lesion complexity characteristics Chronic total occlusion 5.6% 3.4% 15.0% 0.076 Ostial location 7.4% 8.0% 5.0% 1.00 Bifurcation lesion 7.4% 4.5% 20.0% 0.037 Severe calcification 10.5% 8.3% 11.5% 0.376 Severe tortuosity 14.7% 9.3% 17.3% 0.053 Syntax score 19.6 ± 13.0 19.2 ± 12.8 21.7 ± 13.9 0.487 Primary lesion stenosis 84.9 ± 9.2 mm 84.1 ± 9.4 88.5 ± 7.3 0.027 Primary lesion length 22.2 ± 10.6 mm 21.9 ± 9.8 24.0 ± 13.6 0.427 Total lesions treated 1.47 ± 0.69 1.35 ± 0.55 2.00 ± 0.97 0.009 Total stents deployed 1.59 ± 0.79 1.43 ± 0.66 2.30 ± 0.98 < 0.001 Contrast volume used (ml) 183.4 ± 74.0 170.0 ± 63.6 242.8 ± 88.1 < 0.001 Fluoroscopy time (min) 18.2 ± 10.4 15.2 ± 7.2 31.6 ± 11.7 < 0.001 Dose-area product (cGy-cm 2 ) 12,519 ± 15,971 10,440 ± 13,726 21,667 ± 21,576 0.036 Among the 20 procedures requiring any manual aspect, planned partial manual assistance was performed in 4 procedures (3.7%), unplanned partial manual assistance in 8 procedures (7.4%), and manual conversion in 8 procedures (7.4%) ( Fig. 4 ). Procedures utilizing any manual aspect were associated with primary lesions with higher pre-PCI stenosis (88.5 ± 7.3 vs. 84.1 ± 9.4%, P = 0.027) and greater number of lesions treated (2.00 ± 0.97 vs 1.35 ± 0.55, P = 0.009). Such procedures were also associated with more stents deployed (2.30 ± 0.98 vs 1.43 ± 0.66, P < 0.001), more contrast (242.8 ± 88.1 vs 170.0 ± 63.6 ml, P < 0.001), longer fluoroscopy time (31.6 ± 11.7 vs 15.2 ± 7.2 min, P < 0.001), and higher X-radiation dose-area product (21,667 ± 21,576 vs. 10,440 ± 13,726 cGy-cm 2 , P = 0.036) ( Table 2 ). Fig. 4 Reasons for manual assistance or conversion during robotic PCI. The predominant reasons for manual assistance or conversion during robotic PCI were either guide catheter/wire support or a limitation of the robotic platform. Planned partial manual assistance ( n = 4) was utilized in all cases due to a limitation of the studied CorPath 200 robotic platform. Planned manual assistance was used in one procedure for the deployment and retrieval of a FilterWire EZ embolic protection device (Boston Scientific, Marlborough, MA) during saphenous vein graft PCI, one for the use of an OptiCross intravascular ultrasound (IVUS) catheter (Boston Scientific), and two for planned hybrid robotic/manual PCI (manual in one case for bifurcation stenting of the LAD/diagonal branch; manual in the second case for stenting of an obtuse marginal branch with kissing balloon inflation into the LCx; both after treating another lesion robotically) ( Tables 3 and 4 ). Table 3 Case characteristics for any manual assistance or conversion utilized during robotic PCI. Caseconversion typeReason for conversionRadial accessSyntax scoreTotal lesionsR-PCI target lesionsR-PCI target vesselR-PCI lesion typeProcedure duration (h:min)Contrast volume (ml)Fluoroscopy time (min)Dose-area product (cGy-cm 2 )MACE (SCAI) ⁎ CK-MB > 3 × ULN 1 Manual conversion Dissection No 5 2 1 RCA B1 0:58:00 110 21 6343.7 No No 2 Planned manual assistance FilterWire No 1 1 SVG-OM C 1:53:14 300 48.1 25,924.7 Yes Yes 3 Manual conversion Bifurcation lesion (side branch PTCA without KBI) No 12 3 1 LCx c 1:29:36 380 33.4 24,695.1 No No 4 Unplanned manual assistance Dissection No 34 4 1 LAD c 2:06:17 450 45.9 38,693.2 No No 5 Unplanned manual assistance Poor guide catheter support No 25 2 1 RCA c 1:09:00 200 29 87,822.2 No No 6 Unplanned manual assistance Removing jailed side branch wire with manual guide catheter No 30 1 1 LAD c 0:50:07 120 18.7 10,361 No No 7 Unplanned manual assistance Manual guide reengagement No 17 1 1 RCA c 1:01:39 220 28.3 15,846.4 No No 8 Planned manual assistance Hybrid approach (manual bifurcation stenting of LAD/diagonal) No 23 4 1 LCx c 0:27:18 350 37 13,048.8 No No 9 Manual conversion Poor guide catheter support No 20 3 3 LAD, RCA, PLV B1.C 1:22:00 280 41 23,834.7 No No 10 Planned manual assistance Hybrid approach (manual stenting of OM and KBI into LCX) Yes 21 2 1 LAD c 1:05:00 220 28 11,111.3 No No 11 Unplanned manual assistance Manual guide reengagement No 13 2 2 mid-distal LAD c,c 1:51:00 270 23.2 70,274 No No 12 Unplanned manual assistance GuideLiner No 7 3 3 LAD, mid-distal A, C, C 1:27:00 260 32.6 13,359.7 No No 13 Unplanned manual assistance GuideLiner No 62 2 2 prox-mid LAD C,C 1:43:00 340 46.7 21,519.9 No No 14 Manual conversion Bifurcation stenting and KBI No 23 2 2 LAD, diagonal C, B1 1:13:00 205 31 8304.8 No No 15 Planned manual assistance OptiCross IVUS No 1 1 RCA C 0:52:00 170 18 6047.6 No No 16 Unplanned manual assistance OptiCross IVUS No 6 1 1 RCA c 0:45:00 210 19 6520.6 No No 17 Manual conversion GuideLiner No 1 1 LAD c 0:33:00 150 20 5991.9 No No 18 Manual conversion Bifurcation stenting and KBI No 28 3 3 RCA, LAD. LCx c, c,c 1:11:00 250 27 11,331.9 No No 19 Manual conversion Acute vessel closure during wiring No 10 1 1 RCA c 1:50:14 200 60.5 21,953.4 No No 20 Manual conversion GuideLiner Yes 32 2 2 LCx, OM c, c 1:07:00 170 23 10,351.5 No No Table 4 Reasons for manual assistance or conversion during robotic PCI. Type of manual assistanceReasons for manual intervention Planned partial manual assistance ( n = 4) Robotic platform limitation (4) Planned hybrid approach (2) Manual for bifurcation stenting of LAD/diagonal Manual for stenting of OM and kissing balloon intervention into LCx OptiCross IVUS FilterWire Unplanned partial manual assistance ( n = 8) Adverse event (1) Non-flow-limiting dissection (without MACE) Robotic platform limitation (2) OptiCross IVUS Jailed side branch wire removal with manual guide catheter manipulation Limited support (5) Poor guide catheter support Manual guide catheter reengagement (2) GuideLiner (2) Manual conversion ( n = 8) Adverse event (2) Non-flow-limiting dissection (without MACE) Acute vessel closure during wiring (without MACE) Robotic platform limitation (3) Provisional bifurcation lesion intervention (side branch PTCA without kissing balloon intervention) Provisional bifurcation stenting and kissing balloon intervention (2) Limited support (3) Poor guide catheter support GuideLiner (2) Unplanned partial manual assistance ( n = 8) was utilized in one procedure due to a non-flow-limiting dissection that was rapidly stented without MACE, two procedures due to limitation of the robotic platform (one for use of an OptiCross IVUS catheter, another for manual removal of a jailed side-branch wire and manual guide catheter manipulation), and five procedures due to limited guide catheter/wire support (one for manual guidewire/catheter advancement during poor guide catheter support, two for manual guide catheter reengagement, and two for use of a GuideLiner support catheter (Vascular Solutions, Inc., Minneapolis, MN)) ( Tables 3 and 4 ). Manual conversion ( n = 8) was utilized in two procedures for adverse events (one for a non-flow-limiting dissection without subsequent MACE, another for acute vessel closure during guidewire advancement without subsequent MACE), in three procedures due to limitation of the robotic platform (two for provisional bifurcation stenting and final kissing balloon inflation, one for bifurcation side branch balloon angioplasty only), and in three cases due to limited guide catheter support (one for manual guidewire/catheter advancement during poor guide support, and two for use of a GuideLiner support catheter) ( Tables 3 and 4 ). A single MACE was observed in the robotic PCI cohort following revascularization of a diseased saphenous vein graft. Planned partial manual assistance was used for deployment and retrieval of a distal embolic protection device during this procedure, followed by asymptomatic elevation of post-PCI CKMB > 10 × ULN during observation. Two non-flow-limiting dissections were observed in different procedures, resulting in unplanned partial manual assistance and manual conversion, respectively, without subsequent signs/symptoms of myocardial injury. In the overall cohort, CKMB > 3 × ULN was observed in 6 (5.6%) of the robotic PCI procedures. 4 Discussion This is the first study evaluating the reasons for partial manual assistance or manual conversion during R-PCI with the CorPath 200 robotic system in clinical practice. In over one-hundred consecutive patients with a high proportion of complex coronary lesions (78.3% type B2/C) undergoing R-PCI, 81.5% of all procedures were completed entirely robotically. Planned or unplanned manual assistance was required in 11.1%, while 7.4% of procedures required manual conversion for procedure completion. In this complex patient series with significant clinical co-morbidities and anatomical lesion complexity, including mild-moderate calcification, bifurcation lesions, chronic total occlusions and unprotected left main disease, high clinical and robotic technical success are reported. The predominant reasons for manual assistance were limitation of the robotic platform and limited guide catheter support. The prospective multicenter PRECISE study demonstrated safety and feasibility of R-PCI using the CorPath 200 robotic system; however, the lesions treated in the trial were primarily short, single-vessel lesions of low-to-moderate complexity (average lesion length 12 mm; 87% non-C). Increasingly, high-risk PCI is being performed for complex CAD in patients who are poor candidates for surgical revascularization. Patients with extensive or complex CAD anatomy have prolonged procedure times, resulting in higher radiation exposure to both the patient and physician as well as the potential for operator fatigue. Longer fluoroscopy time is associated with higher procedural complexity and periprocedural complications . A recent study using data from the Cath-PCI registry demonstrates that non-modifiable factors such as ACC/AHA type C lesions, bifurcations, or use of IABP may significantly increase fluoroscopy time during PCI . Although efforts to decrease radiation dose and contrast volume are made, these factors often limit the degree of revascularization that can be safely achieved during a single procedure. R-PCI for complex coronary anatomy has the potential to address this issue and there have been reports describing the feasibility of robotic PCI for complex coronary anatomy . In the currently reported series, the proportion of procedures requiring manual assistance/conversion (18.5%) was higher compared to the PRECISE study (1.2%) due to the more complex coronary anatomy as described above (78.3% compared to 31.7% type B2/C lesions). When comparing procedures requiring any manual aspect against those completed entirely robotically, higher Syntax score, longer procedure time, higher contrast use, increased fluoroscopy time and patient cumulative radiation exposure were observed. Nevertheless, even partial robotic assistance in these more complex procedures significantly mitigates primary operator radiation exposure and its associated occupational hazards. This study reports the initial experience with the robotic platform at this center and incorporates the learning curve of a new technology. Importantly, no clinically relevant adverse event associated with the use of the robotic platform was reported in this complex cohort of patients treated with R-PCI. The selection of individual procedures for robotic or a manual approach was not randomized and could have been subject to a bias. As the majority of procedures in this study requiring manual assistance or conversion was due to the limited ability to manipulate the guide catheter or simultaneously advance two balloons/stents with a robotic drive, important iterative steps are required for further advancing this robotic technology. The CorPath GRX is the next generation robotic platform and recently received FDA approval. This newer version of the CorPath robotic platform has a third control knob that enables robotic manipulation of the guide catheter. It will address the issue of almost half of the procedures requiring manual assistance and increase the number of complex procedures that can be completed entirely robotically. 5 Conclusion This is the first report demonstrating the frequency and reasons for manual assistance or conversion during R-PCI in clinical practice with a high proportion of complex PCI lesions. A high clinical success with the R-PCI platform for complex CAD occasionally requires partial manual assistance or manual conversion. The majority of procedures requiring manual assistance/conversion were due to limitations of the current robotic platform including the restricted ability to manipulate the guide catheter and the inability to simultaneously advance two balloons/stents. With the recent improvements in the robotic platform, the next generation system should enable a greater proportion of patients with complex coronary anatomy to be treated with R-PCI.