A System for Grading Filtration Blebs Following Trabeculectomy

Abstract

Abstract: : Purpose: To devise an immediate, universally applicable grading system, capable of accurately describing the range of blebs found after trabeculectomy. Methods: A Bleb Grading Protocol was developed. Graders made judgements about the appearance of the bleb from standard mono photograph of superior conjunctiva. There are six criteria to assess: Area. 1a: the central demarcated area of the bleb, compared with total conjunctival area in the picture. 1b: the maximal area of the bleb, compared with total conjunctival area in the picture (both scored of 1-5 depending on extension). 2: Height. The highest point of the bleb is considered (score of 1-4). Vascularity 3a: the central demarcated area of the bleb (score 1-5). 3b: the peripheral part of the bleb (score 1-5). 3c: the non-bleb conjunctiva (score 1-5). An area of subconjunctival haemorrhage larger than the scleral trap door is noted (yes/no). Each part is compared to standard photographs provided. The number given to each part relates to the photo that best fits the criteria being graded. 100 serial, standard photographs taken 4 weeks after trabeculectomy surgery were supplied to the MEH Reading Centre. Grading of the photos was based on the Grading Protocol. Independent graders performed magnified image grading. Results: High level of agreement was found between the 2 readers. Percentage agreement: Central bleb area (1a): 91%. Maximal bleb area (1b): 91%. Height (2): 94%. Central bleb vascularity (3a): 93%. Maximal bleb vascularity (3b): 85%. Non bleb vascularity (3c): 86%. Subconjunctival haemorrhage: 99%. In the remaining % the disagreement between the readers was no greater than one score. Kappa scores: 1a: 0.870, 1b: 0.857, 2:0.903, 3a: 0.899, 3b: 0.864 3c: 0.779, blood: 0.98. Conclusions: This system allows the main features of filtration blebs to be accurately described without expensive photographic equipment. Information can be collected from remote sites with images transferred electronically to grading centres. The reliability of the system is confirmed by high interobserver agreement.

Keywords: imaging/image analysis: clinical • clinical research methodology • wound healing