Nomogram for excimer laser platform shown to be robust in re-treatment analyses

Key Points

Dr. MacRae and colleagues developed the nomogram for the system. He reviewed outcomes for a series of 1,088 eyes treated from November 2004 to May 2007.

With all eyes having a minimum follow-up of 3 months, only 60 (5.5%) underwent re-treatment using criteria based on a low threshold to select patients for enhancement. Among the 1,088 eyes in the overall series, none lost more than one line of best-corrected visual acuity (BCVA), and including enhancements, all patients achieved uncorrected visual acuity (UCVA) of 20/20 or better in at least one eye. In only eight of the 1,088 eyes (0.7%) was UCVA of 20/20 or better not achieved.

Manifest refraction

The nomogram treats for sphere based on individual eye parameters (preoperative higher-order aberrations [HOAs] and manifest refraction) rather than using a constant surgeon offset as initially recommended by the manufacturer. Its superior performance in achieving better accuracy is highlighted by a comparison of its associated UCVA and refractive results versus those achieved in the FDA clinical trial of the platform, noted Dr. MacRae.

"One of the reasons we think the nomogram function is so robust is that it uses the preoperative manifest refraction instead of the wavefront data. While this seems counterintuitive at first, we realized that preoperative manifest refraction is really the gold standard because it is what the patient perceives in terms of neural interpretation," he said.

"In addition, the nomogram compensates for interactions between preoperative HOAs and postoperative sphere and cylinder as well as for myopia degree and astigmatism coupling. These are interactions that occur with virtually all excimer laser platforms," Dr. MacRae said.

He was the surgeon for all 1,088 eyes treated with the nomogram. Mean spherical equivalent (SE) for the group was –4.95 D, and mean cylinder was –0.76 D. Eyes with low myopia and smaller pupils received a conventional treatment using proprietary software (Zyoptix PlanoScan, Bausch & Lomb) rather than wavefront-guided ablation. Eyes treated for monovision or with <20/20 preoperative BCVA were not included.

Of the 1,088 eyes, only 34 (3.1%) had UCVA <20/20 after the primary surgery, and only one eye (0.1%) had UCVA <20/40. For the latter eye, BCVA was 20/20 and the refraction was –1.00 –0.25 ×37°.

"This eye represents our worst outcome using the [nomogram]," said Dr. MacRae.

Re-treatment criteria

Of the 60 eyes that were re-treated, only 26 (43%) had UCVA of 20/20 or better prior to re-treatment. The criteria for re-treatment are based on low levels of residual sphere and/or cylinder.

"Originally, eyes were re-treated if sphere is over- or undercorrected by 0.5 D or more and/or if they have residual cylinder of 0.5 D. However, now I will even re-treat some eyes with residual sphere as low as 0.25 D, and I tend to treat residual cylinder of 0.5 D or more and even 0.25 D in some patients," he said.

"Even in patients with low residual refractive error, if I put up a lens in a trial frame and the patient is happier with the vision, then I consider re-treatment," Dr. MacRae added.

Analyses to predict factors associated with re-treatment showed no correlations with preoperative SE, astigmatism, or HOAs.

"This tells us our nomogram is pretty robust and has been well-optimized to compensate for these factors. However, with only 60 eyes needing re-treatment, we probably do not have adequate power to identify a predisposing factor," Dr. MacRae said.

Dr. MacRae, Manoj Venkiteshwar, PhD, and their optics colleagues in the Customized Vision Correction Lab at the University of Rochester who have collaborated in developing the nomogram are continuing to investigate their theory that an interaction occurs between coma and astigmatism.

"Optical analysis shows that when the pupil diameter decreases, coma transforms into cylinder," he said. "However, the mathematical analysis to understand this complex interaction is probably the most difficult we've done so far because it involves polar coordinate mapping to decipher the relationship. We've not been entirely successful yet in breaking the code."