13 Spectral-Domain Optical Coherence Tomography Evaluation of Pre-Descemet Endothelial Keratoplasty Graft
Pre-Descemet endothelial keratoplasty (PDEK), a recent modification of endothelial keratoplasty, involves transplantation of the pre-Descemet layer (Dua layer) along with the Descemet membrane (DM) with endothelium. In this selective tissue transplantation, the pre-Descemet layer provides additional thickness to the thin DM. Dua and colleagues identified the pre-Descemet layer as a tough, fibrous layer about 10.15 ± 3.6 µm thick. 1 The principal advantages of the technique are easy intraoperative tissue handling and less injury to the donor harvested graft. The initial results showed good postoperative outcomes and fewer surgical complications. 2 The technique inherited the basic advantages of early visual rehabilitation and lowered graft rejection, similar to Descemet membrane endothelial keratoplasty (DMEK). 3 , 4 The postoperative graft position, although seen clinically by slit lamp, high-resolution spectral-domain optical coherence tomography (SD OCT) provides additional information for the configuration of endothelial grafts. In this chapter, we show the postoperative graft configuration using SD OCT.
13.1 Pre-Descemet Endothelial Keratoplasty
A corneoscleral disc with an approximately 2-mm scleral rim is dissected from the whole globe or obtained from an eye bank. A 30-gauge needle attached to a syringe is inserted from the limbus into the midperipheral stroma (Fig. 13.1, upper left). Air is slowly injected into the donor stroma until a type 1 big bubble is formed (Fig. 13.1, upper left and central). Trephination is done along the margin of the big bubble. The bubble wall is penetrated at the extreme periphery, and trypan blue is injected to stain the graft, which is then cut with a pair of corneoscleral scissors and covered with the tissue culture medium (Fig. 13.1, upper right).
With the patient under peribulbar anesthesia, a trephine mark is made on the recipient cornea respective to the diameter of the DM to be scored on the endothelial side. A 2.8-mm tunnel incision is made at the 10 o’clock position near the limbus. The anterior chamber (AC) is formed and maintained by saline injection or infusion. The margin of recipient DM to be removed is scored with a reverse Sinskey hook and then peeled (Fig. 13.1, middle left and central). Donor lenticule (endothelium-DM-pre-Descemet layer) roll is inserted in the custom-made injector (Fig. 13.1, middle right) and injected in a controlled fashion into the AC (Fig. 13.1, lower left). The donor graft is oriented endothelial side down and positioned onto the recipient posterior stroma by careful, indirect manipulation of the tissue with air and fluid (Fig. 13.1, lower central). Once the lenticule is unrolled, an air bubble is injected underneath the lenticule to lift it toward the recipient posterior stroma (Fig. 13.1, lower right). The AC is completely filled with air for the next 30 minutes, followed by an air-liquid exchange to pressurize the eye. The eye speculum is finally removed, and the AC is examined for air position. The patient is advised to lie in a strictly supine position for next 3 hours.
13.2 Spectral-Domain OCT
Postoperative SD OCT scans were performed by experienced examiner, and the scans were evaluated by expert ophthalmologists. An anterior-segment five-line raster pattern in an axis of 0 to 180 degrees and 90 to 270 degrees was used. The raster scan had five lines (length, 3 mm) and a distance of 250 μm between the lines. The scan was centered at the corneal vertex for a central 3-mm scan. Additional inferior, temporal, nasal, and superior positional scans were also taken. Graft thickness was measured in microns using the tool caliper in the SD OCT. Graft detachment was graded as group I when grafts were completely attached or with a minimal edge detachment; as group II for graft detachments less than one-third of the graft surface area, not affecting the visual axis; as group III when graft detachments comprised more than a third of the graft surface area; and as group IV when grafts were completely detached. Epithelial thickness and recurrence of bulla were noted. Graft-host junction was visualized for interface opacification. Graft split is defined as the separation of pre-Descemet layer and DM. Postoperative central corneal thickness (CCT) was also measured in all follow-ups. Twelve eyes of 12 patients whose mean age was 65 ± 3.8 years were evaluated (nine women, three men). The donor age ranged from 1 to 56 years. Graft size ranged from 7.5 to 8 mm. All eyes had preoperative pseudophakic bullous keratopathy as the indication for endothelial transplantation.
13.3 PDEK Graft in OCT
The mean graft thickness in PDEK was 37.3 ± 3.5 µm (range, 32–44 µm). The graft undergoes minimal dehydration in the postoperative period. The mean graft thickness on days 7, 30, and 90 was 35.5 ± 3.4 µm (32–40 µm), 33 ± 1.8 µm (32–36 µm), and 30.3 ± 2.6 µm (28–36 µm), respectively. The difference over the time period was significant (Friedman test, P = 0.000) (Fig. 13.2), but there was no significant difference (P = 1.000) between the central (3 mm) and peripheral (4–6 mm) graft thickness.
The graft adhered well (Fig. 13.3) in 9 of 12 eyes on day 1. Two eyes had group II detachment (Fig. 13.4), and one eye had group III graft detachment (Fig. 13.5). One eye with grade III detachment underwent air injection. The graft was well apposed on day 1 after air injection; however, there was redetachment on day 12, and rebubbling was done subsequently. In group I eyes with well-adhered graft, a small, shallow peripheral detachment was seen in the inferior (two eyes) and nasal quadrants (one eye). The mean detachment depth was 24.6 ± 8.3 µm. Descemet folds were noted in two eyes on day 1 (Fig. 13.6), but they resolved on day 7 with medical management and supine position. Smooth concave configuration of the posterior cornea was obtained in all eyes by 1 month. None of the eyes had complete graft detachment or lenticule drop.
Eleven of 12 eyes had smooth graft-host interface. One eye had minimal interface haze by postoperative 1 month. After a course of intense steroid treatment, the graft-host interface haze decreased in one eye (Fig. 13.7). Separation of graft into two linear hyperreflective lines was seen by OCT in two eyes. The posterior layer was 16 µm, and the anterior layer was 12 µm. Both eyes had mild corneal edema on day 1 that resolved with strict supine position and medical management.
All eyes had central epithelial defect on postoperative day 1. Epithelial healing was complete in all eyes by 48 hours. The mean epithelial thickness was 44.4 ± 9.8 µm by week 1 and reduced to 37.5 ± 6.2 µm in the last follow-up. There was a significant reduction in thickness (P = 0.003) over the time period. No difference in the central and peripheral epithelium was seen in 11 eyes. On postoperative day 1, mean CCT was 612 ± 46.4 µm. By day 7, significant resolution of corneal edema was noted (P = 0.001). Grade 3 cellular reaction was seen in one eye, and one eye had grade 4 fibrinous reaction. Shallow detachment was seen in the eye with fibrin (Fig. 13.8). Intense steroid treatment attained good graft adherence by 2 weeks.
13.3.1 Correlations and Associations
No significant correlation was found between graft thickness and the best corrected vision at day 1 (P = 0.409) and day 90 (P = 0.661) or between the mean CCT reduction and the graft thickness reduction from day 1 to day 90 (P = 0.645, r = 0.149). No correlation was seen between the corneal edema and graft thickness on day 1 (P = 0.374, r = 0.282), nor was an association seen between the corneal thickness on day 1 and the graft detachment (chi-square test, P = 0.285). No association was found between the graft thickness on day 1 and the graft detachment (chi-square test, P = 0.167). There was strong association with graft adherence and best corrected visual acuity (chi-square, P = 0.007). Eyes with early detachment showed poorer visual outcome.