Predictability of residual stromal thickness hinges on precision of instrumentation

Studies find probability of leaving less than 200 µm under a flap is approximately .05% using the Hansatome, 5.3% using the Moria LSK-One keratome.

VANCOUVER, British Columbia, Canada — The predictability of preoperative determination of postoperative residual stromal thickness for LASIK will vary to the greatest extent according to the precision and accuracy of the microkeratome being used, according to research performed with 3D very high-frequency digital ultrasound on the Hansatome (Bausch & Lomb) and the Moria LSK-One (Moria) microkeratome.

“Every single day surgeons determine whether a patient should have LASIK according to an arithmetic calculation. We all want to leave 250 µm,” said Dan Z. Reinstein, MD, MA, FRCSC, international medical director at Lasik Vision Corporation. “Although the laser ablation depth is one of the factors related to the predictability of how much tissue you are going to leave behind, frankly this is the least of our worries. In our research we have shown that the reproducibility of the ablation depth is very high, in the neighborhood of 2.7 µm. What we should really be concerned about is the fact that we are quite unaware of the actual thickness of flaps when we make them.”

Dr. Reinstein said that in most instances predicting flap thickness is based on a number provided by the manufacturer. However, “unless a physician does his own little back-room study, he is unaware of the actual mean, and of equal importance, the actual range,” Dr. Reinstein said. “Most physicians forget that if a manufacturer states that the mean is 160 µm, he must be aware that the standard deviation could be 25 or 30 µm, and hence the range is going to be about 75-100 µm. Therefore, the thinnest flaps could be 50 µm thinner and the thickest flaps will be 50 µm thicker than expected.”

For an average patient, where less than 300 µm is predicted to be left after the ablation, “you may be leaving well less than 250 µm,” Dr. Reinstein said. For example, a 520-µm cornea with a 160-µm flap and a 110-µm ablation should theoretically leave a 250-µm flap. “But what is the probability that the imprecisions of pre-op corneal thickness, flap thickness and ablation depth will leave the patient with a residual stromal thickness that is less than 250 µm?” he said.

Measuring three variables

--- 3D thickness map reconstructed from meridional VHF digital ultrasound arc-scans, showing on a color scale in microns the residual stroma under the flap of a cornea that had undergone LASIK 6 months previously. Note the thinnest point under the flap is not in the center, it is 1-mm below the center of the cornea. The use of intraoperative pachymetry may not be reliable for finding the thinnest point within the residual stromal layer intraoperatively.

Dr. Reinstein’s method for determining the imprecision of residual stromal thickness prediction in a patient preoperatively entails understanding and measuring three variables: microkeratome depth reproducibility, preoperative pachymetric reproducibility and laser ablation depth reproducibility. “We then combine these errors, or imprecisions, to derive what the chances are of getting less residual stroma than anticipated,” Dr. Reinstein said, describing a computer model that he developed for this purpose.

A very high-frequency digital ultrasound prototype from Cornell University was used in the research. “We scanned the cornea before and after LASIK, directly measuring each layer with 1-µm precision,” Dr. Reinstein said. “We’re also able to map each corneal layer before and after LASIK in three dimensions.” In essence, “we’re able to divide each cornea into its components,” he said, “determining the precision of flap thickness and the precision of laser ablation depth. We also measured the reproducibility of the Orbscan Pachscan 3D pachymetry.”

These three reproducibilities were combined to model the probability of an eye achieving less than 200 mu under the flap, despite never predicting a preoperative residual stromal thickness of 250 mu. “The model was used to demonstrate the differences in LASIK safety between the “160” head of the Hansatome and the 160-µm cutting head Moria LSK-One (labeled “130” head by the manufacturer), using the same excimer laser, the Technolas 217C (Bausch & Lomb),” said Dr. Reinstein, an assistant professor of clinical ophthalmology and research director of refractive surgery at the Weill Medical College of Cornell University in New York City.

The Hansatome achieved a mean thickness of 138 µm (standard deviation 22 µm) in the right eyes of 14 patients who underwent LASIK surgery. On the other hand, a separate study showed that the Moria LSK-One had a mean thickness of 160 µm in 20 eyes, but with a slighter higher standard deviation of 30 µm.

Average Orbscan pachymetric reproducibility varied between 2.5 µm and 10 µm, with a mean of about 7 µm. In addition, “the ablation depth was measured as having a root-mean-square error of 2.77 µm, which is effectively the precision of tissue removal,” Dr. Reinstein said. “It turns out that most of the imprecision in residual stromal calculation will come from the microkeratome.”

Specified cutoff

--- VHF ultrasound scan of a cornea 4 months post-LASIK. The epithelium and keratectomy interface are clearly visualized from one end of the corneal scan to the other. The keratome entrance point, with evidence of retraction of the cut end of Bowman’s layer under the epithelium (E); a small irregularity in the smoothness of the keratectomy is evident (I); the flap is noted to be thinner (T) nasally than temporally; the keratome track is noted to stop abruptly within the nasal stroma to produce the flap hinge (H).

The computer model predicts, given a specified target residual stromal thickness, the probability of having a residual stromal thickness below a certain cutoff. “We enter the mean corneal thickness, the reproducibility of preop mean flap thickness, the reproducibility of the corneal pachymetry, and the reproducibility of ablation depth,” Dr. Reinstein said. “You can combine reproducibilities by calculating and adding the variances and then returning to a combined standard deviation.”

A target of 250 µm was set in the model. “If a patient comes in with a minimum thickness required — for example, a predicted 250-µm residual stromal thickness — he has a 5.2% chance of having less than 200 µm left when using the Moria LSK-One keratome,” said Dr. Reinstein. In contrast, “because the Hansatome cuts 138 µm with a standard deviation of 22 µm, a patient who comes in with a predicted residual stromal thickness of 250 µms would have only a 1.5% probability of ending with less than 200 µm.”

However, the Hansatome findings “are not a direct comparison because we are not assuming preoperatively that the Hansatome is truly cutting at 138 µm,” Dr. Reinstein said. “We are assuming that it is cutting 160 µm (because of the labeling), which means that we in fact have effectively set the target residual at 272 µm, that is, 250 plus 22 (160-138=22), with the Hansatome, we effectively have more tissue than we thought. This lowers the probability to 0.046%, of a patient predicted to end up with 250-µm residual of ending up with less than 200-µm. The Hansatome is labeled as 160 µm but cuts thinner, and this is really a good thing.”

Dr. Reinstein said, “Moria users need to be aware that the company has labeled the keratome with it’s mean thickness, not it’s maximum thickness. Theoretically, with the Moria, 50% of flaps will be thicker than 160, and 16% of flaps will be thicker than 190-µm. In contrast, the Hansatome has been labeled such that only 16% of flaps OD will be thicker than 160-µm.”

In conclusion, “The prediction of the final residual stromal thickness is based upon an understanding of the imprecisions of our keratomes, pachymetry and to a lesser extent, lasers. The analysis also shows that thinner flaps are better for the patient from a risk of ectasia standpoint, although requiring more surgical skill to handle,” Dr. Reinstein said. “I also believe it may be wise for the manufacturers of keratomes to provide calibration certificates for each individual keratome that is sold, in the way that Jose Ignacio Barraquer, MD, suggested in his 1980 text, so that we can precisely calculate, minimize and inform our patients of the risks we are taking.”

For Your Information:

• Dan Z. Reinstein, MD, Ste. 100, 4259 Canada Way, Burnaby, British Columbia V5G 1H1, Canada. (604) 639-4222; fax: (604) 639-4206; e-mail: danreinstein@compuserve.com.