View abstract

This abstract is assigned to sessions
Presentation preference Oral presentation
TitleGrading Image Scale (G.I.S) for anterior chamber cell count in patients with anterior chamber inflammation
PurposeTo develop a systematic tool to be used by clinicians during diagnosis and follow-up of ocular anterior chamber inflammation. Anterior uveitis is composed of few inflammatory conditions (i.e. traumatic, post-surgical, Idiopathic, associated with rheumatic diseases and other autoimmune diseases) with a common denominator of white cells infiltrating the anterior chamber (AC) of the eye. Often, damaged AC vessels secrete proteins, resulting in milky appearance of the anterior chamber called flare. The prognosis of untreated anterior uveitis is guarded, often leading to severe visual impairment and even blindness. One of the major challenges in uveitis is a proper evaluation of disease severity based on the extent of cell infiltration to the anterior chamber, and to a lesser extent, changing levels of flare when relevant. Moreover, lack of systematic tools of AC cell counting result in a significant challenge in clinical studies accurately measuring the impact of drug therapy on anterior chamber inflammation.
MethodsA set of images representing different scores of Anterior Chamber Cells (ACC) grades were created according to the following scale: Grade 0 – zero cells; grade 1 – 1-5 cells; grade 2 – 6-15 cells; grade 3 – 16-30 cells and grade 4 – more than 30 cells. For each grade, 3 different images were created. Tarsius Grading Image Scale (GIS) software was implemented using Python. User software interaction was through intuitive graphical user interface and fluid transition between (1) test phase in which the user is learning the score mechanics, and (2) main scoring phase in which the user grade randomized images representing different ACC grades. Uveitis experts used the software to grade the different images and the consistency between experts was calculated. The consistency threshold was set to > 75%.
ResultsGrading results of 11 uveitis experts were analyzed and the scoring of the images were found to be consistent between experts and above the set threshold. The higher consistency was found for the lower ACC grades of 0, 1 and 2 (98%, 93% and 91% respectively) and lower consistency for ACC grades 3 and 4 (77% and 83% respectively).
ConclusionTarsius ACC Grading Image Scale is an important step toward a systematic AC cell counting. Further analysis of the applicability of this tool for clinical studies is needed.
Conflict of interestNo
Authors 1
Last nameDE SMET
Initials of first name(s)M
Authors 2
Last nameNEUMANN
Initials of first name(s)R
Authors 3
Initials of first name(s)D
Authors 4
Last nameMILMAN
Initials of first name(s)Z
Authors 5
Last nameKREMER
Initials of first name(s)M