The slitlamp impression of corneal ring segment depth was not simply related to the OCT measurements (). The slitlamp estimate appeared to depend on where the observer chose to measure the segment depth. The slitlamp impression was closer to the apparent depth A′ in superficially positioned rings (RSSN 1, 2, 7, and 8), where the examiner estimated the corneal depth anterior to the ring segment in the middle of the ring width and compared it with the corneal thickness at the inner edge of the ring segment. For more deeply positioned rings (RSSN 3, 4, 5, and 6), the slitlamp impressions were closer to the true depth, where the examiner estimated the corneal depth anterior to the inner edge of the ring segment and compared it with the full-corneal thickness at the same position. Overall, the slitlamp impression exaggerated the shallowness of shallow segments and the depth of deeply placed segments. Corneal and anterior segment OCT may provide more accurate and objective measurement of ring segment depth.
The distal portion of the ring segments tends to be shallower. This trend was not seen with the segment implanted in the superior cornea; however, it was obvious for the inferior portions of nasal and temporal segments and very pronounced for the segments in the inferior cornea. All patients had inferiorly positioned cones. We hypothesize that the weaker and more flexible inferior cornea may bow downward ahead of the mechanical dissector, causing the channel to progressively shallow during the dissection process. Since the inferior cornea is already thinner in these cases, shallowing is of greater concern.
Based on these results, keratoconic eyes with a very thin inferior cornea may be at greater risk for depth-related complications. The surgeon may have to exercise caution in placing segments in the inferior cornea to treat inferior ectasia in keratoconus, keratectasia, and pellucid marginal degeneration. In addition, the manufacturer may consider modifying the bevel on the leading edge of the dissector to encourage deepening rather than shallowing of the channel during dissection. Femtosecond laser channel dissection at a constant depth may be a better option.15
The allocation of ring segment thickness and position has been an area of active discussion. Swanson reported good results with asymmetric ring placement (with a thicker superior ring and a thinner inferior ring) in 38 patients with pellucid marginal degeneration: 70% of their patients experienced 3 or more lines of improvement in visual acuity, and 90% of their patients showed cone displacement centrally (M. Swanson, “Corneal Architecture Remodeling with Intacs for Pellucid Marginal Degeneration,” presented at the ASCRS Symposium on Cataract, IOL and Refractive Surgery, San Diego, California, USA, May 2004). The authors hypothesize that placing the thicker segment superiorly may exert a pulling force that displaces the cone centrally. Future studies could explore asymmetric placement with a thicker superior segment, or even placement of a single segment superiorly. Good results have also been reported with placement of the thicker segment inferiorly over a mean follow-up of 9 months.5
Our results suggest that if only a single segment is to be implanted, superior placement might be safer.
Our study showed a significant correlation between segment depth and anterior stromal compression. This is to be expected because the tension from corneal ring insertion is borne by a thinner layer of anterior stroma, leading to greater tensile stress (force per cross-sectional area). Greater stress leads to more strain, and more transverse strain naturally causes compression in the depth dimension. The left-most point in (RSSN 7, patient 4) was an exception. This ring segment was shallow, but the cornea anterior to it was not compressed. We believe the lack of anterior stromal compression in this case was caused by placement of the channel in the flap space. The flap could easily move to drape over the segment rather than experience localized strain. The detailed OCT image clearlyshows the flap being lifted in this case (). This unintended effect may actually have been protectived there was no epithelial breakdown at the shallow distal portion of the segment and a stable BCVA of 20/20 - was achieved.
Shallower ring segments may result in more complications, such as epithelial and stromal breakdown and ring extrusion, because the anterior stromal compression is greater. The greater tensile strain on the anterior stroma could lead to gradual stromal breakdown in a process similar to keratoconus disease progression. In addition, these regions may experience more forward bowing of the anterior corneal surface over the implant. The greater anterior surface curvature may explain the epithelial breakdown. Superficially placed segments also may compromise diffusion of nutrients to the epithelium. In vivo confocal microscopy of intrastromal ring segments have shown epithelial cells with highly reflective nuclei in regions over the segment, which may indicate increased biologic stress caused by the device.16
Imaging of ring segments has been reported using very-high-frequency ultrasound17
and lower speed retinal OCT systems18
in patients with low myopia. This study is the first to use a high-speed CAS-OCT system to assess ring segment depth. The high speed allowed us to obtain a quick survey of ring segment depth greater than 360 degrees. The measurement was rapid and noncontact and allowed us to detect the change in channel depth along the insert path.
In conclusion, high-speed CAS-OCT provided precise assessment of intrastromal corneal ring segment depth and may help identify patients at greater risk for depth-related complications. The Visante OCT system (Carl Zeiss Meditec, Inc.) is similar to the prototype used in this study. The Visante, approved by the U.S. Food and Drug Administration in October 2005, may provide a readily available tool for noncontact measurement of ring segment depth after intrastromal corneal ring implantation.