Six surgeons share their experiences with eye tracking technology and next-generation LASIK.

In the quest for improved outcomes and higher patient satisfaction in laser surgery, eye tracking technology is the next logical step. However, this technology is relatively new, and many surgeons still see it as unproven.

To help shed some light on this increasingly important area, we've asked six surgeons with first-hand experience to share some of what they've learned about eye tracking technology and the different instruments that are now, or soon will be, available.

On the following pages:

• Dr. Brian Will
explains the basics of eye tracking technology and shares some of his experience using the LADARVision eye tracker from Summit/Autonomous.
• Dr. David R. Hardten
explains why the tracker in VISX's new Star S3 excimer laser has changed his mind about the value of eye trackers.
• Dr. Steve Trokel
shares some of his experience using the new VISX system.
• Dr. Jack Holladay
talks about helping LaserSight develop their next generation eye tracking system.
• Dr. Jorge L. Alio
of Alicante, Spain describes some of the positive results he's had using Bausch & Lomb's Technolas 217A excimer laser system in his clinical practice.
• Dr. Dan Durrie
discusses the eye trackers currently under development by Nidek and Bausch & Lomb.

Locking On: Tracking Eye Tracker Technology

By Brian Will, M.D.

Vancouver, Wash.

Eye tracking technology hasn't really come into its own just yet. But within a few years, performing laser surgery without it will almost certainly be unthinkable. By then, numerous lasers with second or third generation eye trackers will be FDA-approved. Custom ablations will probably be the norm.

But the future isn't here yet. Today, many ophthalmologists feel uncomfortable using eye trackers, even when they're available. Also, we still haven't seen a lot of hard statistical evidence showing significantly improved outcomes because of tracking technology. As a result, many of us are still skeptical of their value.

My skepticism, however, has lifted. Like many of my colleagues, I've felt a distinct aversion for an over-dependence on technical devices when performing delicate surgical maneuvers. But after performing more than 2,000 LASIK procedures using the LADARVision platform, I've become a true "tracker-holic" with a serious "tracker-dependent" personality disorder.

I have seen the future - and a significant component of that future is eye tracking!

The basic problem is one we're all familiar with: patient ocular movements during LASIK can be rapid, erratic and unpredictable. Manual fixation of the globe by the surgeon can minimize these variables, but it also introduces a significant risk for off-axis ablation. Moreover, many obese patients and smokers move significantly during breathing, and many younger or more athletic patients present with "head bobs" because of strong heartbeats. Globe fixation can't compensate for these movements. Eye tracker technology, on the other hand, has the potential to eliminate these movements as an issue during surgery.

Whether or not you're sold on eye tracking technology, this is clearly a development who's time is coming, for numerous reasons:

• the increasing popularity of small spot scanning lasers
• the increased need for precision when treating hyperopia
• the advent of custom ablation
• larger treatment zones and the concomitant increase in patient eye fatigue
• higher patient expectations regarding outcomes
• allaying patient fears about the consequences of their eye movements. (A recent survey found that 50% of respondents would only have laser surgery if a tracker was used.)

For all of these reasons, eye tracking technology is sure to become an important part of surgery such as LASIK. With that in mind, I'd like to review some of the basic information about how this technology works, and share some of my experiences using it (in particular, my experiences with the LADARVision eye tracker from Summit/Autonomous).

Many different eye trackers are currently under development around the world, although only two are FDA-approved in the United States (Autonomous' LADARVision tracker, and the tracker in VISX's new Star S3 excimer laser). Every tracker has different characteristics, and they're not always easy to compare. (In fact, some technical specifications are still proprietary information.)

An accurate comparison is also difficult because the technology is so new; only the LADARVision platform has any FDA-approved statistics about accuracy to report, and conclusive data about how much this technology improves actual patient outcomes is still incomplete. (Indeed, many surgeons feel that their current outcomes are so good that it will be difficult or impossible to prove that tracking technology has improved their results. When custom ablations become commonplace, however, this may change.)

Generally speaking, trackers can be categorized along two axes:

• Laser or video-based.
Laser tracking is at the heart of Autonomous' LADARVision tracking system. Video camera based detectors are the basic building blocks for the eye tracking systems incorporated into the Bausch and Lomb Technolas 217, the LaserSight LaserScan LSX, the VISX Star S3 and the Nidek EC-5000.
• Active or passive.
Active trackers follow the movement of the eye and change the position of the surgical laser accordingly. Passive trackers simply monitor for movement and shut the laser off if the eye moves beyond limits set by the doctor. (Many trackers do both.)

The only laser platform currently using laser radar to detect eye movement is the Summit Autonomous LADARVision system. Although I've used several tracker-guided systems, my preference is for the LADARVision system. This technology has exceeded even my own high expectations.

The LADARVision active tracking system consists of an infrared laser, a position sensor and adjusting tracking mirrors. The infrared laser emits a beam that reflects off adjusting tracker mirrors to the pupillary boundary. Light reflected off the pupillary boundary is detected by a position sensor, measuring the magnitude, velocity, rate of acceleration and direction of the eye's movement.

During surgery, the infrared laser beam automatically acquires and optimizes itself to the pupillary boundary of the patient 4,000 times each second, locking and coupling itself to the eye's saccadic movements. (Evaluation of 554 individual eyes during LADARVision surgery revealed that saccadic eye motion ranged from 0.04 mm to 1.16 mm, with a mean of 0.35 mm.)

A number of statistics confirm the reliability of laser-based tracking technology:

• Outcomes
. According to Autonomous, visual and refractive outcomes 6 months after treatment with the LADARVision system were independent of the amplitude of eye movement; patients with large eye movements during surgery ended up with visual acuity as good as those patients with small eye movements.
• Spot placement accuracy
. Autonomous reports that absolute tracker error for this system is 37 microns for saccades of 100� per second (a typical 1� saccade during fixation). Resolution to within 37 microns for a 0.9 mm spot size translates to a peak spot placement offset error of 4%. (The average error may be much smaller.)
• Signal to noise ratio.
In any electronic detection device, "noise" (produced by measurement artifact and electrical instability in the detection circuitry) effectively reduces the speed, reliability and reproducibility of the tracking system. The LADARVision system measures eye position many times faster than the laser pulse rate, giving it a very high signal to noise ratio.
• Tracker failure rate
. Autonomous reports that LADARVision successfully tracked 100% of 678 eyes in their phase III FDA clinical trials. This is confirmed by my own experience: I've managed more than 2,000 cases without a single tracker failure.

Along with the statistics reported by Autonomous, reports from other surgeons using the LADARVision system and two published reports of PRK and LASIK performed on patients with congenital nystagmus, I believe this confirms the robust nature and reliability of the laser radar detection system.

One potential limitation of the laser radar system is the requirement that the patient pupil dilate to 6.5 mm or larger in order to initiate tracking. However, in my experience, this hasn't been a problem. In more than 2,000 consecutive cases at our center, no patients have failed to dilate to that point.

A related limitation is the inability to track patients with pupils severely distorted by previous trauma, disease or intraocular surgery. (Unfortunately, this latter deficit is true for video trackers as well.)

Caution must also be exercised when treating pseudophakic eyes because the tracker can be mislead by the edge of the lens implant (or the haptic, in cases where the IOL is significantly decentered).

Infrared video tracking technology identifies the pupillary margin by using one or more video cameras to detect reflected infrared light. High speed processors use the contrast between the reflected energy from the iris and pupil to identify the pupillary margin. This information is then delivered to the laser guidance system. (Different manufactures are using variations on this theme. Some are including enhancements such as contrast stabilization and multi-camera input to create three dimensional data arrays. )

In the past, a significant limitation of video tracker technology was relatively slow tracker speed. However, faster computer technology is rapidly improving our ability to process large, complex graphical images. Newer models, such as the VISX Star S3, operate at 60 Hz; the Wavelight Allegretto is reported to reach speeds of 200 Hz or higher.

In addition, contrast enhancement techniques have minimized previous target detection problems such as reductions in corneal clarity due to the laser ablation and artifacts produced by surface reflections. Improvements also continue to occur in pupillary edge definition software.

The cornea and pupil offer the surgeon five potential targets to use as a basis for centration:

1. the entrance pupil
2. the patient's visual axis
3. the image of the laser fixation diode reflected from the cornea to the observer, representing a measure of the angle kappa
4. the anatomic center of the cornea
5. the apex of the cornea. The primary target currently tracked by both laser radar detectors and infrared video camera systems is the patient's pupil.

Most refractive surgeons currently agree that the laser ablation should be centered over the entrance pupil (the pupil that we see when we look at the patient's eye). This is actually a Purkinje image formed by the cornea; it's located about 0.5 mm in front of, and is about 14% larger than, the real pupil.

However, some surgeons believe that the ablation should be centered on the patient's visual axis. They may use the position of the corneal reflection of the laser fixation diode as an approximation of the visual axis.

Fortunately, in numerous cases the center of the entrance pupil and the corneal reflex are the same. However, many times the corneal light reflex is also observed nasal to the pupil because the fovea is temporal to the entrance pupil. (The magnitude of that deviation from the center of the entrance pupil is angle kappa.) A significantly positive angle kappa appears to be more common in hyperopia than in myopia.

As angle kappa increases, the decision of where to center the laser ablation becomes more challenging for both the surgeon and the eye tracker.

Despite the fact that both video and laser-based systems use the pupil as a tracking target, the tracking systems differ in how this information is used to center the laser treatment.

When using a laser radar tracking system, the centration point is determined prior to the surgery in a relaxed, controlled setting. The tracker captures a graphic computer image of the patient's anterior segment while the patient is focusing on a fixation diode that is coaxial with the excimer laser beam. (The relaxed setting ensures that the entrance pupil isn't affected or distorted by patient anxiety, a high adrenaline state, the neurophysiologic effects of benzodiazepines or other sedatives, iris ischemia from the LASIK suction ring, accommodative effort, or scleral expansion by the LASIK suction ring being positioned over the ciliary body.)

The surgeon reviews the graphic image and decides exactly where on the cornea the laser ablation will be centered. The attractiveness of this approach is that the position of the center of the laser ablation is chosen by the surgeon. Once this determination is made, the laser locks in the desired position and maintains that location information uniformly throughout the laser procedure.

Also, because the laser radar tracking system "remembers" the location of the center of the initial laser ablation, subsequent laser re-treatments or enhancements are centered on the identical location. This prevents "eccentric re-treatments" that can induce astigmatism and other unwanted optical aberrations.

Infrared video tracking systems also center the laser ablation on the patient's entrance pupil, but they do so while the pupil is in a dynamic state during the laser procedure.

During the ablation, the video tracking system captures multiple sequential graphic images of the anterior segment using an infrared video camera. As each video frame is processed, a computer analyzes the image and calculates the center of the patient's entrance pupil. This new data is then transferred to the aiming mechanism of the laser device.

Laser radar tracking systems require dilation of the pupil; video based tracking systems do not. There has been some debate about the significance of that difference.

Mydriasis enlarges and distorts the entrance pupil, rendering it an inappropriate target for centration. This would seem to put the laser-based eye tracker at a disadvantage. However, centration is performed before the procedure, with the patient's pupil undilated; the information is stored in the laser for reference during the laser procedure. Laser radar tracking doesn't use the dilated pupil as part of the centration protocol, nor does it use the dilated pupil as a guide for centering the laser ablation intraoperatively.

Although some doctors see the need for dilation as a disadvantage, in our practice we've found dilation of patients' pupils prior to laser surgery to be very efficient for both the patient and the surgeon, and also a means to improve surgical outcomes:

• Patients can schedule their complete ocular examination and surgery on the same day, significantly reducing time away from work and saving our staff time and effort. (This is particularly attractive for patients who travel long distances to our center.)
• Pupillary dilation allows me to perform a complete ocular examination, including detailed fundus evaluation, prior to taking the patient to the laser surgery suite. This is both efficient and first-rate patient care.
• Mydriasis greatly improves my ability to identify flap microstriae and interface debris at the time of surgery. The excellent red reflex afforded by pupillary dilation allows me to identify and correct the problem either under the laser microscope or at the slit lamp.
• Because we use neosynephrine as part of our dilation protocol, hemostasis is markedly improved, particularly for large flaps in patients where we are blending to 8 and 9 mm, or in patients with small corneas or micropannus from contact lens wear.

Today's technology is impressive, but it will continue to improve. Within a few years trackers should be able to accurately track the visual axis (instead of the pupil) and reposition mirrors at much higher speeds (currently a serious limiting factor). Accuracy and consistency will reach ever higher levels.

Today, many surgeons still consider eye tracking technology optional. Given the comparatively straightforward ablations most of us still perform, the fact that eye tracking technology is still in its infancy and the limited availability of eye trackers in the United States, this attitude has some justification. However, my experience has convinced me that eye tracking technology is already capable of making a significant difference in your practice, both in terms of improving outcomes and allaying patient fears.

Of particular concern to many ophthalmic refractive surgeons is the issue of intraoperative compensation for ocular cyclorotation, particularly for patients with astigmatism. The imminent arrival of wavefront guided laser ablation simply raises the "bar" even higher by increasing the significance of cyclorotational eye movements in even purported "spherical" corrections.

This can be more of a problem than many surgeons realize. We recently conducted our own study to evaluate he magnitude and frequency of both ocular cyclorotation that occurs when patients move from the sitting to the supine position (pre-laser cyclorotation), and the cyclorotation that occurs during the laser ablation (intra-laser cyclorotation).

1. Intra-laser cyclorotation occurred in only 4 out of 2,092 consecutive eyes; the maximum observed rotation was only 2�. Based on this data, designing a tracking system that monitors intra-laser cyclorotation would have little if any beneficial effect on clinical outcomes.
2. However, this was not the case with pre-laser cyclorotation, which we measured in 159 consecutive eyes. In this smaller series, 64.7% of eyes demonstrated pre-laser ocular cyclorotation up to 12� in magnitude; 2.5% rotated more than 10�.

In this series, pre-laser cyclodeviation was an identifiable source of error that, without compensatory action, would have adversely affected our ability to attain a predictable refractive endpoint. Clearly, a tracking system that allows the surgeon to direct the tracker to adjust for pre-laser intraoperative cyclorotation would be highly valuable.

Currently, the only eye tracking system that allows for this compensation is the laser radar system. However, as we move closer to designing laser systems capable of delivering truly customized corneal ablations, it seems likely that other eye tracking systems will incorporate similar technologies.

- Brian Will, M.D.

One of the main purposes of eye tracking technology is to make custom ablations a practical reality. With that in mind, I've been working with LaserSight to develop their AstraTrack eye tracking system (the next generation version of the AccuTrack system that's currently available as part of their LaserScan LSX excimer laser outside the United States). I've recently done 20 custom ablations in Monterey, Mexico, using the new system.

My goal is to leave the eye prolate, and LaserSight is the farthest along in terms of custom ablations with a prolate corneal outcome. (The prolate shape leaves patients with the same overall quality of vision after as before.) All 20 custom ablations were successful, with a prolate cornea following the surgery.

Brian R. Will, M.D., is president and chief executive officer of Will Vision & Laser Centers in Vancouver, Wash. He is former chief of refractive surgery at the Aker-Kasten Cataract and Laser Institute in Boca Raton, Florida and has served as project director for the International Institute for Advanced Laser Surgery. He's certified by both the American Board of Ophthalmology and the American Board of Eye Surgeons.

By David R. Hardten, M. D.

Minneapolis, Minn.

In the past, I've been nervous about using eyetrackers during surgery. I felt this way for a number of reasons:

• I'd used some of the early laser eye trackers outside of the United States. I wasn't terribly impressed with the results.
• My friends in other countries told me that they routinely turned their trackers off because they weren't quite sure what the laser was tracking.
• I've been quite happy with my surgical results without using an eye tracker. If there was any chance that I'd get in trouble by using a tracker, I didn't want to.

But the potential need for a tracker was also clear. Both my patients and I have felt uncomfortable about the issues of centering the laser ablation, maintaining fixation and compensating for random eye movements. My patients almost always ask, "Did I do okay looking at the light?" (I routinely tell them that they did a good job.)

I've had the opportunity to use the VISX STAR S3 with Active Eye Tracker for the last 2 months. During that time I've gained some knowledge of VISX's eyetracking system. I find a number of things about the VISX eyetracker appealing:

• It's simple to operate.
• It's able to track even very significant patient eye movement.
• In general, the tracker does an excellent job of locking on to the pupil. (In some eyes that are extremely deep-set, or in patients that have extremely puffy lower cheeks, the oblique camera gets blocked, and the instrument is unable to track the eye.)
• The tracker makes it easy to tell when the system is engaged. The entire reticle flashes when the tracker is unable to lock onto the pupil. The center of the reticle flashes when the tracker has found the pupil and is waiting for you to decide where in relationship to the pupil you want to lock the tracker. The reticle is totally still when it's found the pupil and you've engaged it.
• You have the option of deciding whether to center directly on the pupil, or slightly to one side or another. (This is helpful when the pupil is somewhat displaced from the geometric center or apex of the cornea.) At the same time, as a safety feature, the tracker won't let you decenter more than 0.5 mm.
• It's able to track under the LASIK flap. This is true even if you need an instrument to hold back the flap when a patient's upper lid is pushing the flap towards the center of the cornea.
• It tracks light and dark irises equally well.
• The system does an excellent job of centration. As a safety measure, if the patient looks away from the fixation light (moving farther than 0.5 mm) the laser will instantly disengage. I haven't seen any clinically evident decentrations in postoperative topography or patient symptoms. The early visual results have been excellent.
• Because the pupil doesn't need to be dilated, the fixation target is easier for the patient to see, which I believe is extremely important. (Anything that diminishes the patient's ability to see the fixation target will reduce the reproducibility of the surgery.) This also makes it easier to use the secondary clues of pupil position within the geometric center of the cornea and pupil position in relationship to the corneal apex to verify that the patient is looking at the fixation light.

One last practical note: The VISX tracker has a sensitivity calibration system that changes the sensitivity of the cameras depending on the lighting at the time of the surgery. For this reason it's important not to turn on the tracker until the lights are adjusted as you like them, and until after you lift the flap.

The VISX STAR S3 system has alleviated my fears about using an eye tracker. Results have been excellent, and my patients are very happy to know that the tracker is working.

by Daniel S. Durrie, MD

Overland Park, KS

For the last 12 years I've had the privilege of working with multiple laser platforms. During that time, all the lasers I've used have had alignment systems that were adequate for large area ablation. Eye tracking technology wasn't necessary.

Today, however, we're moving into the more automated world of custom ablation. As a result, all laser manufacturers are redesigning their systems to accommodate eye trackers.

I've been working with Nidek and Technolas on the clinical studies of their excimer lasers since 1993. Both companies have trackers for their lasers that are in use outside the United States. The main reason they haven't been introduced into the clinical platforms in the United States is that both of these lasers use a form of wide-area ablation.

Although the Nidek EC-5000 is a flying slit laser and the Technolas 217 is a flying spot laser, their ablation patterns for the correction of myopia generally start in the center of the cornea and move peripherally.

To treat myopia and astigmatism, the Technolas laser divides its program into either four or eight "mini-treatments." These are done in a multi-pass fashion with the pattern starting centrally and moving peripherally during the ablation. (In contrast, Autonomous lasers use a single "treatment" in which the laser spots are randomly dispersed around the cornea to avoid the laser plume.)

The Nidek laser uses an expanding diaphragm and expanding slit masking system to deliver its flying slit laser to the cornea. (In that respect it's similar to the VISX and Summit lasers, which have been in the United States for a significant period of time.) We've found that tracking isn't necessary when using this type of delivery system.

Both the Nidek and Technolas ablation patterns have produced excellent results - without the use of a tracker. For example, the Technolas 217's FDA approval data shows that 87.5% of eyes achieved 20/20 or better with a single treatment.

Nidek is preparing to introduce a segmental ablation pattern that will use both a flying spot and a flying slit. When that happens, they'll add the tracker to their clinical platform. Likewise, Technolas/B&L is planning to introduce Zyoptics - a wavefront-guided ablation pattern - this fall. At that time they'll also introduce their eye tracking system. (I've used this system in Europe and found it to be very satisfactory with their ablation platform.)

Given the current availability of the tracking system for the VISX laser and the excellent eye tracking system available on the Autonomous laser, the future seems clear. I believe tracking will become standard on lasers, just like cruise control has become standard on most automobiles.