This year marks the 10th anniversary of the FDA's first approval of an excimer laser for refractive surgery. More than a decade of technical research and development as well as extensive clinical use outside the US preceded this decision. The impressive progress in this field deserves recognition.

EARLY EXPLORATIONS

In the mid-1980s, investigators considered two refractive applications for the excimer laser. The first tested used the laser to improve the most popular refractive surgical procedure at that time, RK. Some researchers believed that the laser would allow surgeons to create a more accurate and effective radial corneal incision and thus render RK safer and more effective. Investigators also espoused the highly controversial theory that ablating a large area of tissue would alter the corneal curvature and provide a more accurate, safer procedure (similar to keratomileusis) than RK. Both types of laser systems were built, and they were described at the First International Workshop on Excimer Refractive Surgery in 1986.

Marguerite McDonald, MD, and her team at the LSU Eye Center in New Orleans carried out extensive animal studies that led to their performing the first successful myopic treatment in humans with computer-controlled PRK in 1998. Additionally, investigators compared the results of ablating through the epithelium versus the theoretically more precise method of removing the epithelium to minimize the total amount of tissue ablation.

THE TECHNOLOGY
The First Lasers
Excimer lasers in the early 1980s produced large beams of partially incoherent energy. One early engineering decision was to develop systems that would control these beams rather than create a small spot scanning system in which changes in software could modify the pattern of ablation. The decision was based on a desire for more rapid ablation and the difficulty of maintaining the registration of small spots over the relatively extended period required for removing tissue. The excimer laser beams were homogenized and rotated to produce rotational symmetry, and variable apertures were used to control the beam's size. The first systems could only correct simple spherical myopia, which has radial symmetry and does not require a precisely centered ablation. Subsequent systems used small spot scanning or the compromise of a scanning slit-beam across a fixed aperture. These later developments required the introduction of rapid pulse rates, scanning techniques, and automatic eye tracking.

Astigmatic Treatment
In the late 1980s, investigators took two approaches to treating astigmatism. One involved a variable slit-shaped aperture to create a cylindrical beam to treat astigmatism, a process analogous to the variable circular aperture to treat myopia. Alternatively, cylindrical-shaped masks were designed to be interposed between the laser beam and the front of the cornea.
In either case, more precise centration and localization of the ablation pattern were necessary. Investigators quickly recognized that errors of only a few degrees in the alignment of an astigmatic ablation would leave substantial residual astigmatism. Larger ablation zones were also necessary to produce a transition at the ends of the cylindrical shape and to increase the width of the resulting elliptical ablation. The introduction of fiduciary lines projected on the patient's face somewhat improved alignment, especially when used in conjunction with ink marks placed on the eye. It is only in this past year that the issue of rotational alignment was solved by aligning the diagnostic and the therapeutic beams to the image of the iris.

Hyperopic Treatment
The development of treatment algorithms for hyperopia required a significantly different approach. The pattern of ablation removes no tissue from the center of the treated area while taking away the most tissue at the edge of the optical zone. Studies in the early 1990s showed that a transition zone extending as far as 9mm was necessary to minimize optical regression due to healing effects in the cornea. One subsequent approach used an axicon element, which is a specially designed lens that creates an annulus of light to produce a doughnut-shaped (annular) beam. A more versatile technique was to introduce an inline-scanning element that moved the large beam off axis. The large ablation areas for hyperopia required long ablation times, which taxed both the patient's and the doctor's ability to maintain proper fixation and centration. The need for assistance in the form of automatic eye tracking was becoming critical.

Eye Trackers
The first automatic eye tracking systems developed had a high signal-to-noise ratio to minimize the effect of artifacts. Sampling speeds consistent with the pulse rate of the laser system were employed. Three-dimensional systems could detect eye movement out of the focal plane of the laser. These tracking systems reduced the anxiety and fatigue of both the patient and the doctor and allowed for more accurate ablation patterns that improved visual outcomes.
Once the large-beamed systems had scanning capabilities, investigators conducted extensive research to evaluate the ablation profile for each diameter of beam. With this information, they developed new algorithms that took advantage of the savings in ablation time of the larger beam sizes while providing the resolution of smaller sized beams. This new technique was called variable spot scanning, and it allowed the ablation area to increase to meet the needs of the procedure while minimizing the total procedural time.

Aberrometry
The introduction of tools capable of measuring optical aberrations more complex than sphere and astigmatism opened the door for further improvements in the results of laser vision correction. These systems could sample hundreds of points over the pupil and determine the wavefront error of the entire corneal surface. Based on this information, surgeons could determine an ablation pattern that would reduce or eliminate a given eye's aberrations. The requirements for an accurately aligned eye and ablation pattern significantly increased. The eye's cyclotorsional rotation as the patient moved from a seated to a supine position sufficiently compromised the accuracy of simple axial-alignment techniques to thwart accurate registration of the treatment.

Ink marks placed during the examination help, but crosschecking the wavefront error against identifying marks on the iris and programming the excimer laser system to recognize these marks yields more reproducible results. Current systems that use iris or limbal registration may significantly reduce aberrations.

CONCLUSION

Technical improvements in excimer laser systems for refractive surgery have dramatically increased the range and types of refractive errors that can be treated successfully. The clinical results are dramatically better today than when the first system was approved by the FDA 10 years ago. Continued research to improve engineers' and surgeons' understanding of the basic ablation process and of optical aberrations will drive advancements in the technology that will produce even better results. It is unlikely that new developments will occur at the dramatic rate of the past decade, however.

Charles Munnerlyn, PhD, is cofounder of Visx, Incorporated and is currently retired. He may be reached at (408) 238-3920; charlesrm@comcast.net.
Stephen L. Trokel, MD, is a Clinical Professor at Columbia-Harkness in New York City. He is a consultant for Advanced Medical Optics, Inc., in Santa Ana, California. Dr. Trokel may be reached at (212) 326-3320; trokel@columbia.edu.