Inaccurate Lesion Length Measurement is Still a Problem and is Easily Solved
Disclosure: Dr. Paul Campbell reports he is a member of the Medical Advisory Board and Consultant for Corindus Vascular Robotics.
Paul Campbell MD, FSCAI, FACC can be contacted at Sanger Heart & Vascular Institute, 100 Medical Park Drive, Suite 210, Concord, North Carolina 28025. Email: Paul.Campbell@carolinashealthcare.org.
Could you describe your study?
Our study1 looked at the accuracy of interventional cardiologists in regards to assessing coronary plaque length and selecting the appropriate stent length based on visual estimation, which is the technique used in the majority of cath labs in the United States and around the world. It was an internet-based study where 40 interventionalists viewed 25 images, of which 5 were repeat images (20 lesions total), presented at random throughout the overall presentation. We looked at both inter and intra operative variability in terms of assessment of plaque length and the resulting stent selection compared to quantitative coronary angiography as the control. What we found was that “eyeballing” or visual estimation is not very accurate. In over half the cases, the lesions were actually assessed too short and in about 20%, the visual assessment was too long. Similarly, when we looked at the length of the stent that was selected, interventionalists selected stents that were too short in more than half of the lesions, and too long in about a quarter. When we looked at the 5 images that were interspersed at random and shown a second time, more than one-third of the images assessed a second time had a difference of 3mm or more between the first and second assessments. Not only was there inaccuracy in assessing lesions and selecting stents, but when you put the same image in front of the operator for a second interpretation, there was variability between the first and second assessments. There was both inter and intra operator inaccuracy in assessing lesions visually.
Can you describe the angiographic images used in the study?
The lesions were A, B1, B2, and C, so a wide variety. We provided two orthogonal views. There was a 6 French guide catheter in the image that could be used as a reference in terms of diameter in millimeters. A big problem in the cath lab is that we are looking at a 3-dimensional object, the heart, in 2 dimensions. When we look at how to improve physician-dependent factors, we have to take into account foreshortening, angulation, and vessel overlap. For example, if we say a lesion is 9mm in length, and don’t take into account that there is actually a bend in the artery that we are not appreciating, and actually the lesion is 13.7mm, when we place a 12mm stent, the edge of the lesion is missed by 1.7mm. That amount doesn’t sound like a lot, but even that small distance of missed lesion coverage can increase the chances of restenosis and target vessel revascularization.
Stenting too short or too long can cause problems.
Neither is good. Too short is called longitudinal geographic miss, and that is where the stent misses the edge of a lesion. Longitudinal geographic miss was first described when radiation was used for in-stent restenosis. More recently, the STLLR trial using sirolimus-eluting stents included 47.6% of cases with longitudinal geographic miss. Patients with longitudinal geographic miss had 2.3 times the rate of target vessel revascularization as compared to patients who did not.2 On the other end of the spectrum is the use of stents that are longer than necessary. Over-stenting or stent length in excess of adequate lesion coverage has also been associated with adverse outcomes and increased costs. Mauri et al, in an evaluation of 6 bare metal stent clinical trials, showed that each 10mm of excess stent length resulted in a 4% increase in late percent diameter stenosis (P<0.0001) independent of stented lesion length.3 A similar analysis of sirolimus-eluting stents demonstrated a 2.1% increase in percent diameter stenosis per 10mm of excess stent length (P=0.04).4 In addition, the use of greater numbers of stents has been shown to contribute to higher cardiovascular-specific initial hospitalization costs.5 Excessive stent length increases the chance of late loss down the road, and stenting too short means you miss the edge of the lesion and increase the patient’s chances of repeat revascularization. Optimally, you want an accurate measurement to put the right-sized stent in the right place and reduce the patient’s chances of coming back. Also, if you miss the edge of a lesion with longitudinal geographic miss and realize it during the case, you are placing a second stent to cover the missed region. Accurate lesion length measurement offers the potential for up-front savings on stent cost, and for lowering the number of patients returning with recurrent symptoms and requiring repeat procedures.
Can you tell us about the interventionalists who examined the images in your study?
All are interventionalists in the United States. The majority were from large hospitals, either academic or community teaching hospitals, and they were fairly experienced, having between 11 and 20 years of experience. Most were between the age of 40 and 60.
It sounds like the inaccuracy arises from human and angiographic frailty.
Yes. A solution to this problem can be found with the use of computers and robotics. There are numerous examples throughout the world of how robots and computers are more accurate than humans in a number of aspects. In fact, if you look at urologic surgeries, such as prostatectomies, if you look at some hip operations with orthopedic surgery, and if you look at hair follicular implants for hair replacement, these are all examples where robots, with robotic precision, can be more accurate and precise than humans. My understanding is that 98% of prostatectomies are now done by urologists via robotic surgery and that only 1-2% of urologists can actually do a perfect job in comparison, so the majority of the prostate procedures are being done robotically. For hip surgeries, there is more accurate measurement using computers to set up angles and alignments with the hip, and so this is another example of where a robot and a computer can improve the abilities of a human. I think the most value in what I do as a physician is in the emotional intelligence and judgment aspect of things. The black-and-white of medicine is straightforward. I don’t need to think about whether someone needs a tetanus shot. Either they need a tetanus shot or they don’t. But when we need to decide whether a patient should have an intervention or not, or decide which procedure is best for a particular patient, it is more of a grey situation. I believe that what we should be concentrating on as physicians is the judgment aspect of decisions, and that the things that are black and white or arithmetic in nature are better done by computers and machines, and we should utilize them when applicable. Our study is an example of exactly this scenario. As interventionalists, we are viewing images on a screen 6 feet away, eyeballing and visually estimating coronary lesions. We are looking at a moving 3-dimensional object in 2 dimensions — there can be foreshortening, vessel angulation, vessel overlap — yet we are making assumptions when in all actuality, a robot or computer could give us a better measurement and assessment.
Clearly there are technologies out there that will measure lesion length.
Yes, there are several, including quantitative coronary angiography (QCA), which has been around for years, intravascular ultrasound (IVUS), and coronary CT angiography. Most of my patients referred for intervention have been diagnosed with angina, have had a stress test showing ischemia, and are on at least one or two antianginal medications. They are coming to see me for a cath because they are having symptomatic angina. Most agree that if we find the lesion that is consistent with their ischemia on stress testing and it is amenable to intervention, that they would like to have it treated (an ad hoc procedure), such that they don’t need a second procedure or miss more days of work. QCA is an accurate method of lesion assessment; however, in real-time, it requires the technologist or nurse at the control monitor to trace the diseased vessel segment, and the reference vessel and the results can be influenced by their precision and accuracy. I use IVUS, but in order to do length assessment, a pullback is required, which needs a motorized unit to pull the catheter back at a set rate. If you are going to do it accurately, you need to hook it up to a motorized drive. This process adds more time to the procedure and so most interventionalists have stopped doing the pullback. We do use IVUS to look at lesion characteristics, plaque composition, and vessel diameter, which IVUS does quickly. I haven’t seen an IVUS pullback done in several years in most labs. Computed tomography (CT) coronary angiography is another option and is an excellent technology. The issue is that it is a whole separate procedure and the patient’s heart rate must be slowed down. The set up and procedure takes an hour or two of the patient’s time. All of these technologies are available for lesion measurement, but I personally don’t use them because of the limitations I described. What we have been using instead is real-time robotic measurement with the CorPath System (Corindus). During the procedure, we deliver a wire prior to stenting that allows us to measure the lesion precisely. It is a real-time measurement and allows us to do ad hoc interventions without bringing the patient back for a second procedure after reassessing with another technology. Most of my patients prefer that their symptomatic angina be treated as efficiently and cost effectively as possible, and we try to do that as best we can, in a timely fashion.
Can you describe how the robotic measurement of coronary plaque length works?
With the CorPath System, we can measure the lesion either with a wire or a balloon. Briefly, the system consists of two major components: an interventional cockpit and a bedside unit. The interventional cockpit is comprised of a radiation-shielded workstation where the interventional cardiologist sits at the cockpit and performs percutaneous coronary intervention (PCI) by manipulating joysticks or touchscreen controls. Monitors inside the cockpit provide fluoroscopy, electrocardiogram (ECG), and hemodynamic data. The bedside unit consists of an articulated arm mounted to the procedure table to support a robotic drive housing a single-use sterile cassette. The guide catheter is connected to the cassette, through which the guide wire and interventional devices (balloons and stents) are introduced. The interventional cardiologist can manipulate the linear and rotational motions of the guide wire and interventional devices to facilitate the lesion treatment. The physician controls the measurement feature by advancing the balloon or guide wire markers to the distal and proximal edges of the lesion of interest. The distal edge is marked as “0” on the touch screen display. When the marker is at the proximal edge, a measurement of the distance traveled by the marker is displayed to provide a lesion length. The CorPath System can make sub-millimeter measurements. An example would be a case where I estimated a lesion was 8mm long and was going to put in a 12mm stent, but the lesion actually measured 13.9mm robotically. If I had placed a 12mm stent, I would have missed the edge of the lesion by 1.9 mm. Instead, I am able to place a 15mm stent because the robotic system provided me with an accurate measurement.
How much time does it take to measure with the robotic system?
It takes less than 30 seconds. Once you cross the lesion with the wire or balloon, you pull it back to the edge of the lesion, then reset the counter and start tapping the touchscreen to bring it back. That’s the nice thing about this technology. It is being used for delivery, precision, and control, all key aspects for putting the right stent in the right place. With robotic measurement, I’m actually doing the measurement myself. If I am not confident in the initial measurement, it can be quickly rechecked. It’s real-time, easy, efficient, and accurate. The other important aspect is that use of a robotic system reduces your radiation. You are in a radiation-shielded cockpit, sitting down, and the screen is right in front of you. The CorPath Precise Trial showed a 95% reduction in radiation exposure to the physician when compared to historical control.6
Are you doing this on every case prior to stenting?
Every case I can, yes. I use it in hemodynamically stable patients who are not acutely ill. This includes Type A, B1, B2 and C lesions. There are some cases with very tortuous, calcified vessels, or chronic total occlusions that I defer to manual mode, but I do try to do robotic measurement in every case possible. It offers improved precision, accuracy, and control of deployment. The other advantages are the radiation safety and the ergonomics: I can sit down and I can take off my lead. It reduces my chances of cataracts, brain tumors, and orthopedic problems involving my lower back. It improves the patient’s care because I am getting the right-sized stent in the right place with better visual assessment of the lesion and precise movements.
How long have you been using the robotic system?
Our hospital has been using the robotic system for a little over one year. We have one system placed in our primary interventional lab where it is used by all the interventionalists.
What caused you to do this study?
Despite advances in PCI technology, restenosis leading to target lesion revascularization still occurs in 3% to 9% of the cases.7-11 The ultimate success of PCI procedures is multi-factorial and dependent on both non-modifiable patient-related factors and modifiable procedure-related factors. Longitudinal geographic miss, as discussed previously, can lead to an increase in the number of patients requiring target vessel revascularization.2
A potential cause of longitudinal geographic miss is the traditional manual measurement of lesion length. Numerous studies have shown that there is a high degree of inter-rater and intra-rater variability associated with visual assessment of stenosis severity when compared to objective measurement techniques.12-16 Based on these data, it is possible that the employment of objective lesion measurement techniques could improve PCI outcomes. The purpose of this study was to determine how accurate interventional cardiologists are at visual lesion length assessment and subsequent stent length selection.
Interventionalists today are using visual estimates of something that can be done more accurately with a computer or a robotic-type device. The results of our study didn’t surprise me.
If interventionalists don’t take on self-improvement, in today’s healthcare climate, there is the strong likelihood that someone else will attempt to step in and regulate.
We need to democratize, standardize, reduce variation, and be reproducible. We want to provide access to high-quality interventions so that you can actually get a consistent, standardized result. It is my belief that robotic and computer innovation ought to be used in healthcare in order to reduce variation and give us consistent, reproducible results. At the conclusion of a robotic case I have arithmetic reassurance that I’ve measured the lesion down to a tenth of a millimeter. I can tell the patient, your lesion was measured precisely and they received the best stent I could give them, in the right place, in a controlled fashion, and that they should have the best possible outcome. Innovation involves using computers, robots and devices where we can actually get a consistent, standardized result. What is going to be required in healthcare is standardization. You want high-quality intervention and standardized procedures, which will eventually change how people train. There are things in medicine that can be done better with computers and robotics. We should embrace the technology.
Any final thoughts?
The study confirms what has been shown in the past: visual estimation is not optimal. I presented a similar study at the CRT2015 meeting in Washington, D.C. in February that showed comparable findings. It was a single-center look at our first 60 cases with the CorPath System.17 Operators first made a visual assessment, then made a robotic measurement, and then determined whether the measurement actually changed what they would have done. In two-thirds of the cases, the robotic measurement did change the subsequent stent length selection. We were actually accurate in visual lesion measurement about a third of the time, short about a third of the time, and long about a third of the time. It has been published in abstract form and is being submitted for full publication. Further, these data showed that the use of the robotic system prevented the use of extra stents in 8.3% of the cases. Measurement of lesions with robotic PCI may reduce measurement errors, need for extra stents, and longitudinal geographic miss. Moreover, the addition of this innovative technology to our cath lab has increased the technical skills of our staff, improved staff morale and led to enhanced teamwork.
- Campbell PT, Mahmud E, Marshall JJ. Interoperator and intraoperator (in)accuracy of stent selection based on visual estimation. Catheter Cardiovasc Interv. 2014 Dec 15. doi: 10.1002/ccd.25780. [Epub ahead of print]
- Costa M a, Angiolillo DJ, Tannenbaum M, Driesman M, Chu A, Patterson J, Kuehl W, Battaglia J, Dabbons S, Shamoon F, Flieshman B, Niederman A, Bass T a. Impact of stent deployment procedural factors on long-term effectiveness and safety of sirolimus-eluting stents (final results of the multicenter prospective STLLR trial). Am J Cardiol. 2008; 101: 1704–1711.
- Mauri L, O’Malley a J, Cutlip DE, Ho KKL, Popma JJ, Chauhan MS, Baim DS, Cohen DJ, Kuntz RE. Effects of stent length and lesion length on coronary restenosis. Am J Cardiol. 2004; 93: 1340-1346.
- Mauri L, O’Malley a J, Popma JJ, Moses JW, Leon MB, Holmes DR, Teirstein PS, Cutlip DE, Donahoe D, Kuntz RE. Comparison of thrombosis and restenosis risk from stent length of sirolimus-eluting stents versus bare metal stents. Am J Cardiol. 2005; 95: 1140–1145.
- Venkitachalam L, Lei Y, Stolker JM, Mahoney EM, Amin AP, Lindsey JB, Kennedy KF, Pencina MJ, Lopez JJ, Kleiman NS, Cohen DJ. Clinical and economic outcomes of liberal versus selective drug-eluting stent use: insights from temporal analysis of the multicenter Evaluation of Drug Eluting Stents and Ischemic Events (EVENT) registry. Circulation. 2011; 124: 1028-1037.
- Weiss G, Metzger C, Caputo R, Delgado J, Marshall J, Vetrovec G, et al. Safety and Feasibility of Robotic Percutaneous Coronary Intervention: PRECISE Study. J Am Coll Cardiol. 2013 Apr 16; 61(15): 1596-1600.
- Urban P, Gershlick AH, Guagliumi G, Guyon P, Lotan C, Schofer J, et al. Safety of coronary sirolimus-eluting stents in daily clinical practice: one-year follow-up of the e-Cypher registry. Circulation. 2006; 113: 1434–1441.
- Moses JW, Leon MB, Popma JJ, Fitzgerald PJ, Holmes DR, Shaughnessy CO, et al. Sirolimus-eluting stents versus standard stents in patients with stenosis in a native coronary artery. N Engl J Med. 2003; 349: 1315-1323.
- Schofer J, Schlüter M, Gershlick AH, Wijns W, Garcia E, Schampaert E, Breithardt G. Sirolimus-eluting stents for treatment of patients with long atherosclerotic lesions in small coronary arteries: double-blind, randomised controlled trial (E-SIRIUS). Lancet. 2003; 362: 1093–1099.
- Stone GW, Ellis SG, Greenberg JD, Spriggs D, Shaughnessy CDO, Demaio S, et al. Comparison of a polymer-based paclitaxel-eluting stent with a bare metal stent in patients with complex coronary artery disease: a randomized controlled trial. JAMA. 2005; 294: 1215–1223.
- Leon MB, Mauri L, Popma JJ, Cutlip DE, Nikolsky E, O’Shaughnessy C, et al. A randomized comparison of the Endeavor zotarolimus-eluting stent versus the TAXUS paclitaxel-eluting stent in de novo native coronary lesions 12-month outcomes from the ENDEAVOR IV trial. J Am Coll Cardiol. 2010; 55: 543–554.
- Detre KM, Wright E, Murphy ML, Takaro T. Observer agreement in evaluating coronary angiograms. Circulation. 1975; 52: 979–986.
- DeRouen TA, Murray JA, Owen W. Variability in the analysis of coronary arteriograms. Circulation. 1977; 55: 324–328.
- Galbraith JE, Murphy ML, de Soyza N. Coronary angiogram interpretation. Interobserver variability. JAMA. 1978; 240: 2053–2056.
- Goldberg RK, Kleiman NS, Minor ST, Abukhalil J, Raizner AE. Comparison of quantitative coronary angiography to visual estimates of lesion severity pre and post PTCA. Am Heart J. 1990; 119: 178–184.
- Fleming RM, Kirkeeide RL, Smalling RW, Gould KL. Patterns in visual interpretation of coronary arteriograms as detected by quantitative coronary arteriography. J Am Coll Cardiol. 1991; 18: 945–851.
- Campbell PT, Kruse KR, Kroll CR, Patterson JY, Esposito MB. The impact of precise lesion length measurement on stent length selection: ramifications for stent savings. J Am Coll Cardiol Intv. 2015; 8(2_S): S26. doi:10.1016/j.jcin.2014.12.101.