Keratoconus is characterized by bilateral, progressive, and non-inflammatory corneal ectasia that results in irregular astigmatism and reduced vision quality due to corneal thinning and protrusion. Spectacles and contact lenses are used early in the course of the disease, whereas surgery is indicated when these modalities fail to correct vision or in case the patient develops intolerance to contact lenses.1 Although penetrating keratoplasty (PK) has a high success rate in keratoconus with frequently clear grafts in most patients,2 approximately 30% of patients develop immune rejection3 and have the risk of graft failure in the long term.4 Intrastromal corneal ring segments (ICRSs) have been recently introduced for use in the surgical treatment of keratoconus in patients with contact lens intolerance and a clear cornea. They act by an arc-shortening effect and flatten the center of the cornea, while additionally providing a biomechanical support for the keratoconic eye.5 There are three types of commercially available ICRSs: Intacs (Addition Technologies, Fremont, CA), Keraring (Mediphacos, Belo Horizonte, Brazil), and Ferrara ring (Ferrara Ophthalmics, Belo Horizonte, Brazil).6–8 The Keraring intrastromal ring segment used in this study has a triangular cross-section that induces a prismatic effect on the cornea. This prismatic effect causes light reflection at the implant base, avoiding glare and obfuscation.8
ICRSs can be implanted mechanically or with the aid of femtosecond laser technology. When the latter method is used, an intracorneal tunnel is created using a femtosecond laser at a predefined depth considering the thickness of the cornea at its thinnest point. Then, an ICRS is implanted into this tunnel. However, complications related to the depth of the tunnel or the ring segments within the cornea may develop intraoperatively or during the postoperative period. For example, intraoperative endothelial perforation has been reported while creating the tunnel with femtosecond laser.9 Depth-related complications such as superficial movement of the ring segment, corneal melting, and ring segment extrusion may be seen during the postoperative period.9–12 Therefore, implantation of ring segments at an appropriate depth in the cornea and monitoring their intracorneal localization remains an important issue.
Slit-lamp impression as assessed by the operator, which may occasionally yield misleading results, has been the conventional method used for the evaluation of the depth of the intrastromal ring segment. Recently introduced devices such as the Pentacam rotating Scheimpflug camera (Oculus, Wetzlar, Germany)13 and anterior segment optical coherence tomography (AS-OCT)14 are able to measure the ring segment depth quantitatively. AS-OCT supplies a high-resolution image using low-coherence interferometry at a 1,310-nm wavelength without any contact or penetration.15–17 It is a new method for the diagnosis, treatment, and follow-up of anterior segment diseases and interventions.18,19
The aim of this study were to examine the use of high-resolution AS-OCT in determining the exact localization and position of the implanted Keraring segments in reference to the anterior and posterior corneal surfaces and to find out if the rings were lying at the same depth as predicted intraoperatively.
Patients and Methods
This study prospectively examined 17 eyes of 13 patients with keratoconus (4 men and 9 women) who underwent uneventful Keraring segment implantation with the aid of femtosecond laser (IntraLase Corporation, Irvine, CA) at the Yeditepe University Department of Ophthalmology between January 2008 and December 2009. The tenets of the Declaration of Helsinki were followed for all study procedures and all patients provided informed consent prior to study entry.
The mean age of the patients was 37.7 ± 11.6 years (range: 22 to 59 years). All eyes had keratoconus: 14 had manifest keratoconus, 2 had forme fruste keratoconus (patients 10 and 11), and one had high astigmatism after penetrating keratoplasty due to keratoconus (patient 13). One patient was lost to follow-up. Femtosecond laser parameters, correction values, and ring properties of each eye are presented in the table.
Table: Femtosecond Laser Parameters for Intrastromal Tunnel Creation, Correction Values, and Ring Characteristics of Each Eye
All patients were intolerant of spectacles and contact lenses with a central corneal thickness of at least 350 μm at the thinnest point and at least 450 μm at the incision site. Patients with herpetic keratitis, acute hydrops, apical scars, and systemic autoimmune or connective tissue disease were excluded. Eyes that had undergone uneventful Keraring implantation were evaluated by AS-OCT (Visante OCT; Carl Zeiss Meditec, Inc., Dublin, CA) at the end of the third postoperative month. One patient who did not return for the follow-up visit was excluded.
Visante OCT has an internal fixation target and the system provides real-time display of video camera and OCT images of the area under examination. While the patient is fixating with the eye that is being examined, the examiner centers the scan by clicking the Purkinje reflex on the real-time video image. High-resolution corneal quad images were saved and analyzed. The measurements were taken in four quadrants at eight localizations using the software’s calipers (software version: 188.8.131.52). The ring segments were also evaluated qualitatively by rotating the direction of the scan 360° to see if there was any irregularity in the position of the Keraring segment, tunnel, or both.
Keraring Intracorneal Segment
The Keraring segment has a triangular cross-section in the form of an isosceles triangle with a long basis width and a short height whereby the height determines its thickness in microns. A schematic diagram of the cross-section is demonstrated in Figure 1. Figure 2 shows a cross-section of implanted ring segments where A represents the apex, B the outer corner, and C the inner corner with respect to the corneal center. The distance from the apex of the triangle to the anterior corneal surface and the distances from the two basal corners to the posterior corneal surface were measured radially using the software’s calipers.
Figure 1. Schematic representation of Keraring segment (Mediphacos, Belo Horizonte, Brazil) cross-section: an isosceles triangle with a fixed basis width of 0.6 mm (600 μm). The thickness ranges from 150 to 350 μm with 50-μm increments.
Figure 2. Postoperative high-resolution anterior-segment optical coherence tomography image of the left eye of a 45-year-old woman from the 135° to 315° meridian showing a cross-section of implanted ring segments where A represents the apex, B the outer corner, and C the inner corner with respect to the corneal center.
All patients were evaluated by a detailed ophthalmological examination including slit-lamp and fundus examinations, manifest and cycloplegic refractions, keratometry, ultrasonic pachymetry, Orbscan (Orbscan II; Bausch & Lomb, Rochester, NY), and AS-OCT (Visante OCT).
The surgical procedure and postoperative treatment regimen was the same for each patient. The tunnel was created using a femtosecond laser (IntraLase, 30 kHz) under sterile conditions and topical anesthesia. Before placing the suction fixation ring, the Purkinje reflex was marked by a Sinskey hook. Corneal thickness was measured using ultrasonic pachymetry (Sonogage, Cleveland, OH) along a 5-mm ring location mark. The target tunnel depth was set at 70% of the thinnest corneal thickness on the 3- to 5-mm zone of the AS-OCT pachymetry map. The following femtosecond laser parameters were the same for all patients: energy cut length = 1.20 mm, energy cut thickness = 1 μm, ring energy = 2.50 μJ, and entry cut energy = 1.50 μJ. The incision was made on the steepest topographic axis. Other femtosecond laser parameters are listed for each eye in the table. The eye was draped in the usual manner under sterile conditions. After opening the incision site and entering the tunnel in both directions by Suarez spreader, the intracorneal ring segments, which were selected preoperatively according to the manufacturer’s nomogram, were implanted with the aid of an Albertazzi forceps. The ring segments were finally positioned by a Sinskey hook. A bandage contact lens was placed for 24 hours and antibiotic eye drops were prescribed for 10 days in the postoperative period. In the patient who had previous penetrating keratoplasty (patient 13), the geometric center of the graft was taken as the reference point for femtosecond laser tunnel creation. No surgical difficulties were observed during the rest of the femtosecond laser procedure. The implantation of the ring segments was also without complexity.
All operations were performed by the same surgeon (RBK). The follow-up examinations were performed 1 day, 1 week, 1 month, and 3 months after the operation.
The surgeon (RBK) performed all high-resolution corneal scans composed of 512 A-scans, each with an acquisition time of 0.25 second per tomogram.
All biometric measurements were performed by the same examiner (EG). The cross-sections of Keraring segments from four tomograms (tomogram 1 at 0° to 180°, tomogram 2 at 45° to 225°, tomogram 3 at 90° to 270°, and tomogram 4 at 135° to 315°) were evaluated. Most of the images included cross-sections of both ring segments simultaneously (Fig. 2). Because four tomograms each containing two cross-sections were evaluated, a sum of eight localizations was included in the analysis: nasal (0° for the right eye and 180° for the left eye), superior nasal (45° for the right eye and 135° for the left eye), superior (90° for both eyes), superior temporal (135° for the right eye and 45° for the left eye), temporal (180° for the right eye and 0° for the left eye), inferior temporal (225° for the right eye and 315° for the left eye), inferior (270° for both eyes), and inferior nasal (315° for the right eye and 225° for the left eye) (Fig. 3).
Figure 3. Demonstration of the meridians projected on the cornea for the four recorded tomograms. Red color represents the axial information for the right eye and blue color for the left eye.
The distance from the apex of the triangular cross-section to the anterior corneal surface was defined as AA, the distance from the inner basal corner to the posterior corneal surface was defined as CP, and the distance from the outer basal corner to the posterior corneal surface was defined as BP (Fig. 4).
Figure 4. High-resolution anterior segment optical coherence tomography images of the right eye of patient 3 from four meridians with cross-sections of the Keraring segment (Mediphacos, Belo Horizonte, Brazil). Images obtained from horizontal 180° to 0° (a), oblique 225° to 45° (b), vertical 270° to 90° (c), and oblique 135° to 315° (d) meridians show the original measurements using the software’s calipers. Note that image (c) does not contain the cross-section of the ring but the tunnel can be faintly recognized.
The AS-OCT assessment of the patient with previous penetrating keratoplasty (patient 13) did not present any difficulty and was similar to the rest of the study eyes.
Statistical Package for Social Sciences version 15.0 software (SPSS, Inc., Chicago, IL) was used for the analysis of data. Normality of data was confirmed using the Kolmogorov–Smirnov test and graphical methods. A P value of less than .05 was considered an indication of statistical significance. Depending on the normality of data, the paired sample t test or Wilcoxon sign rank test was used for the comparisons of BP versus CP distances and target versus actual depths of the rings.
Localization of the Ring Apex
The distance between the anterior corneal surface and ring apex (ring depth, AA) ranged between 160 and 350 μm. The thinnest corneal segment was inferior temporal in 11 eyes, temporal in 4 eyes, and nasal in 1 eye. The actual ring depth after implantation (263.1 ± 42.9 μm) was significantly smaller than the target ring depth calculated based on the thinnest corneal segment (356.9 ± 35.7 μm, P < .001). On the other hand, tunnel depth as assessed before the implantation of the ring was similar to the target ring depth (336.7 ± 23.5 vs 336.7 ± 25.0 μm, P = 1.000) based on the analysis of 9 eyes with available intraoperatively measured tunnel depth data.
The mean ring depths at the upper and lower poles were similar for both temporal and nasal rings: temporal rings = 256.0 ± 54.5 versus 256.0 ± 45.5 μm, P = 1.000, respectively; nasal rings = 261.4 ± 51.6 versus 250.0 ± 45.1 μm, P = .10, respectively.
BP Versus CP Distances
The distance between the outer basal corner and the posterior corneal surface (BP) was significantly greater than the distance between the inner basal corner and the posterior corneal surface (CP) in all of the eight segments, indicating the ring was tilted inward rather than being parallel to the posterior corneal surface. The mean differences between the two distances were as follows: nasal = 88.0 ± 29.1 μm, P = .001; superior nasal = 92.1 ± 21.9, P < .001; superior = 103.3 ± 19.7 μm, P < .001; superior temporal = 94.4 ± 20.1 μm, P = .007; temporal = 85.0 ± 30.1 μm, P < .001; inferior temporal = 94.7 ± 31.8 μm, P < .001; inferior = 111.7 ± 20.4 μm, P < .001; and inferior nasal = 96.0 ± 21.7 μm, P < .001. The minimum BP and CP distances were 100 and 80 μm, respectively.
In addition to several studies reporting on optical and refractive outcomes of ICRS implantation in eyes with keratoconus,8,20–23 imaging of the ICRS has been reported using low-speed retinal OCT, high-frequency digital ultrasound, high-speed corneal OCT, and Pentacam.13,14,24,25 This study evaluated the localization of implanted Keraring segments through measurement of the distances of the ring from anterior and posterior corneal surfaces using high-resolution AS-OCT. To the best of our knowledge, this is the first study to report on such a detailed quantitative imaging of intracorneal Keraring segments postoperatively.
Postoperative complications related to the depth of ring segments such as superficial movement of ring segments, corneal melting, and ring segment extrusion have been reported in association with ICRS implantation.9–12 Kamburoglu et al.13 reported on the postoperative evaluation of Intacs segments using the Pentacam Scheimpflug imaging system 1 year after surgery and demonstrated that the segments were 6.5 to 69.0 μm deeper than the intended tunnel depth. They also reported that the Intacs segment tended to shallow over time, which may be a factor contributing to the development of depth-related complications. Therefore, follow-up of patients by an imaging technique such as AS-OCT may be valuable for predicting future complications related to ICRS.
Depth of the ring segment is estimated conventionally by slit-lamp examination. However, the slit-lamp examination may not adequately show the level of the ring segments in corneal tissue quantitatively; on the contrary, high-resolution AS-OCT can show the ring segments much more accurately in terms of depth and position assessment.26 AS-OCT examination is a non-contact and easy imaging technique causing minimum discomfort and contamination for the patient.16–19 This method allows rapid evaluation of ring segments with quantitative measurement of the segment depths from any desired axial cross-section of the cornea. Especially in case of a change in the position of a ring segment, such as superficial movement, the exact localization of the segment in the cornea can be readily demonstrated and measured by high-resolution AS-OCT. Lai et al.14 used AS-OCT to assess ICRS depth in four keratoconic eyes and compared the findings with the slit-lamp impressions. They found that the slit-lamp estimate appeared to depend on the location from which the observer measured the segment depth. The authors concluded that corneal and anterior segment OCT might provide more accurate and objective measurement of ring segment depth.
In this study, we evaluated the localization of implanted Keraring segments with an apical diameter of 5 mm and a basis width of 0.6 mm, thus providing a 5-mm diameter optical zone.20 They were implanted with the aid of femtosecond laser technology, which creates photodisruption at the molecular level without heat on the tissue surface, using a 1,053-nm wavelength laser that burns the tissue for 10−12 of a second.27 The actual ring apex depth after implantation was on average 93.8 μm (range: 60 to 140 μm) shallower than the target ring depth estimated preoperatively, indicating that the apex of the Keraring cross-section was located closer to the anterior corneal surface than the intended tunnel depth.
Although it can be speculated that this difference may be due to creation of a more superficial tunnel by femtosecond laser than the targeted depth, evaluation of patients with available intraoperative tunnel depth measurement by high-resolution AS-OCT performed just after creation of the tunnel failed to show any significant difference between targeted and observed tunnel depths. Furthermore, it has also been shown that flap thickness is predictable in laser in situ keratomileusis surgery with the IntraLase femtosecond laser.28 Therefore, we have interpreted this phenomenon as the apex having a pushing and compacting effect on the intrastromal collagen lamellae overlying the Keraring segment. To minimize this lifting effect, the cornea may respond by flattening the outer layers to preserve its form and shape. This phenomenon may also explain superficial movement of the ring segments, which may cause corneal melting in the late postoperative period. Lai et al.14 suggested that shallower ring segments might 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 can lead to gradual stromal breakdown. In vivo confocal microscopy of the ICRS has shown epithelial cells with highly reflective nuclei over the segments, which may indicate increased biological stress caused by these segments.29
The base of the Keraring segments was not parallel to the posterior corneal surface, representing another significant finding of this study. Because femtosecond laser-assisted tunnel creation is performed after total applanation of the corneal plane, the tunnel configuration is expected to be parallel to both anterior and posterior corneal surfaces. Even if the outer corner is assumed to be at the ideal depth in the cornea (70% of the pachymetric measurement of that specific point), the inner corner may be close to the posterior corneal surface and might cause endothelial cell damage. Conversely, when the inner corner is at the ideal depth and distance to the posterior corneal surface, the outer corner will be more superficial; in other words, closer to the anterior corneal surface. Whether such positioning represents an additional risk factor for the superficial movement of the Keraring segment in the long term is certainly worth investigating.
A previous study using AS-OCT to examine the depth of Intacs intracorneal ring segments, which were implanted with the aid of mechanical dissectors, showed that segment depth decreased with increasing distance from the incision site. The authors hypothesized that the weaker and more flexible inferior cornea may bow downward ahead of the mechanical dissector, causing the channel to be progressively shallow during the dissection process.14 In another study, intrastromal depth of Intacs segments implanted with the aid of a femtosecond laser were measured using Pentacam and tunnel depth was similar across different points.13 In this study, the depth of the Keraring segment was similar at the upper and lower ends of the segment (ie, at the closest and farthest points to the incision site). As previously reported by Kamburoglu et al.,13 using a femtosecond laser for channel dissection provides a precise depth at each point because tunnel creation is performed after total applanation of the corneal plane.
A recent article compared mechanical insertion and femtosecond laser insertion of Intacs segments and depth predictability using Visante OCT.30 The authors concluded that both mechanical and femtosecond laser-assisted techniques showed a more superficial Intacs placement than predicted. They found no statistically significant difference in the implantation depth between the two groups. Although these studies have some similarities to our study, it is of note that implanted segment rings were of an earlier design with hexagonal cross-section and these segments were implanted more peripherally.
This study covers only the findings at the end of the third postoperative month. In our routine practice, we perform high-resolution AS-OCT at every follow-up visit for patients who received ICRS implantation and analysis of long-term data will be available in the near future. Examination of Keraring segments by AS-OCT at later postoperative periods may demonstrate the superficial movement of the ring segments, which may cause corneal melting. In such a case, the surgeon may be obliged to decide on an explantation with the guidance of AS-OCT. It was recently reported that segment displacement occurred in 0.8% of the patients with Keraring implantation.9 Corneal melting due to superficial placement of ring segments was evident in 0.2% of these patients and all of them had to be removed.
Although new technologies and implant designs seem to represent a major advance in the management of keratoconus, interactions of the ring segments with structural components of the cornea and long-term outcomes of the implants are still unknown. This study showed that ring segments were implanted at a closer depth from the anterior corneal surface than what was intended. This may well contribute to the segment extrusion in the long term. In addition, the inner basal corner of the rings was close to the posterior corneal surface, which may also cause clinically relevant changes. Larger studies with longer follow-up would provide additional information on the long-term outcomes of these implants.
High-resolution AS-OCT seems to be a valuable imaging modality for the measurement of the depth and identifying the exact position of Keraring segments. Routine use of this imaging technique in the follow-up of patients with implanted ICRSs may help predict the ring-related complications.