Abstract
Purpose: :
To develop a standardized grading scheme for corneal verticillata (CV), observed in von Hippel Lindau patients after treatment with vandetanib for renal tumors.
Methods: :
16 Von Hippel Lindau patients with renal tumors were treated with vandetanib, a dual anti-epidermal growth factor receptor and vascular endothelial growth factor receptor 2 tyrosine kinase inhibitor. Slit-lamp biomicroscopy was performed within 1 week of treatment initiation and then every 3 months. Slit-lamp images (12.3 MB per image) were obtained using a Canon EOS 20 digital camera (Tokyo, Japan) mounted on a Haag-Streit BX900 photo slit-lamp (Bern, Switzerland) and OIS WinStation 4000SL software. A grading scheme was developed to quantify subepithelial corneal stranding (grades 0, 1, 2, or 3) and CV (grades 0, 1, 2, 3, 4, or 5). Strand grading evaluated the number and distribution of strands in the cornea. CV grading documented the visibility and area of whorl deposits and their degree of transparency/opacity. Four graders independently graded a randomly selected sample of corneal photographs that were presented on two separate occasions via on-screen digital images. Intra- and inter-grader agreement were assessed using unweighted and weighted kappa statistics (k and kw).
Results: :
There were 9 females and 7 males, median age 42.5 yr (range, 27-63 yr), who underwent 1 to 5 examinations with corneal photography (median, 2.5). 86 images from 32 eyes were obtained. Visual acuity ranged from 20/16 to no light perception (median, 20/20). Although no patient reported visual symptoms (glare, haloes, or clouding) at baseline,10/16 reported them during follow-up. A random sample of 48 images (23 right/ 25 left eyes) was graded. The distribution of grades (using Grader A's grades from session 1) for stranding was: grade 0, 41.7; grade 1. 4.2%; grade 2, 2.1%; and grade 3, 52.1%. For CV, the distribution was uniform: grades 0, 1, 2, 3, 4, 5 were all 16.7%. Intragrader agreement (within-grader agreement) for stranding was moderate: k = 0.35 - 0.56; kw = 0.37 - 0.71. Intragrader agreement for CV was moderate to good: k = 0.36 - 0.63; kw = 0.65 - 0.81. Intergrader agreement (agreement among graders (session 1)) for stranding was good: k=0.53 - 0.75; kw = 0.64 - 0.86. Agreement among graders (session 1) for CV was also good: k = 0.40 - 0.58; kw = 0.71-0.78.
Conclusions: :
Intragrader agreement was in the moderate to good range for both stranding and CV, indicating acceptable reproducibility. Intergrader agreement was good for both stranding and CV and tended to be better than intragrader agreement, suggesting that intragrader variation needs to be reduced. Longitudinal studies of incidence and progression of CV may find this grading system useful to quantify this potential complication of vandetanib treatment.
Keywords: cornea: clinical science • clinical (human) or epidemiologic studies: outcomes/complications • growth factors/growth factor receptors