Standardized semiquantitative analysis of corneal neovascularization using projected corneal photographs--pilot study after perforating corneal keratoplasty before immune reaction

BACKGROUND: A semiquantitative scheme for analysis of corneal neovascularization using projected corneal photographs is demonstrated and tested in a pilot study to analyze occurrence of corneal neovascularization in patients after perforating keratoplasty which subsequently developed transplant rejection. METHODS: Corneal photographs on the slit lamp with diffuse frontal illumination were obtained in a standardized technique. Slides were projected with 100 x magnification and analyzed twice with a 2 months interval. Corneal vessels were graded by two independent observers in each of 12 corneal sectors in a standardized fashion (grade 0: no vessels beyond limbus, 1: vessels between limbus and outer end of a double-running diagonal suture; 2: vessels between outer suture end and graft-host junction; 3: vessels reaching graft-host junction; 4: vessels within donor cornea). All patients with endothelial graft rejection of the prospective Erlangen non-high-risk keratoplasty study were included in a pilot study (1/1997-6/2000: 13 of 325; 4%). One patient without photographs available was excluded. Corneal photographs taken prior to surgery (n = 10), at the last 3 monthly-routine control before (10), at rejection episode (12) and one year later (10) were evaluated for corneal neovascularization. RESULTS: Interobserver correlation at the two assessments was 0.79 and 0.86 (Kendall's Tau B). Correlation between the assessments at the two analyses 2 months apart was 0.8. New vessels with diameter up to 6 microns can be detected. 8 of 12 analyzed patients (67%) with immune reaction after keratoplasty developed corneal neovascularization within 1 year after operation prior to transplant rejection in at least one corneal sector (2.1 +/- 1.9 sectors; 1-6). At time of rejection, new vessels reached the graft-host junction in 2 patients, in 1 patient vessels grew into the donor cornea, whereas in 8 the vessels were seen beyond the outer suture end without reaching host-graft junction (grade I: 1 patient). New vessels usually pointed to the outer suture ends of the double-running suture. CONCLUSIONS: Development of corneal neovascularization e.g. after keratoplasty can be assessed reliably using projected slides of corneal photographs at 100 x magnification. This method has the advantage of being more objective, precise and available compared to simple evaluation at the slit lamp. Postkeratoplasty corneal neovascularization seems to be common in non-high-risk eyes later developing transplant rejection. However, new vessels usually do not reach the host-graft junction. Whether neovascularization after keratoplasty demonstrates a risk factor for subsequent transplant rejection remains to be analyzed in a greater study.