Purchase this article with an account.
Claudio Ignacio Perez, Anwell Nguyen, Sunee Chansangpetch, Marta Mora, Mai Badr, Travis Porco, Shan C Lin; Agreement between different levels of expertise for diagnosing angle closure with Swept-Source Optical Coherence Tomography. Invest. Ophthalmol. Vis. Sci. 2018;59(9):4070. doi: https://doi.org/.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
There is need for understanding if glaucoma specialty training is necessary for evaluating iridotrabecular contact (ITC) using Swept-Source anterior segment optical coherence (AS-OCT) imaging. We performed a cross-sectional study of non-expert and expert interobserver agreement and diagnostic accuracy of AS-OCT images compared with gonioscopy for detecting angle closure.
134 eyes of 134 patients were enrolled. We included patients over 40 years old and excluded patients with prior intraocular surgery or laser and poor quality images. Patients underwent dark-room gonioscopy exam with one glaucoma expert. Angle closure was defined if the posterior trabecular meshwork was not seen without indentation in 2 quadrants. Horizontal and vertical AS-OCT images were obtained in dark conditions with CASIA 2 angle analysis protocol. Two glaucoma-trained ophthalmologists and two medical doctors (non-glaucoma trained), graded all images in an independent manner. Angle closure by AS-OCT was assessed if 2 or more quadrants had ITC. Main outcome was interobserver agreement measured with Kappa score (κ) between experts, non-experts and gonioscopy. For diagnostic performance, area under receiver operating characteristic curve (AUC) was calculated using gonioscopy as the gold standard.
Mean age of patients was 67.6 ± 10.1 years, 55.2% were female, and 24.6% had angle closure by gonioscopy. The agreement of closed angle between non-experts and gonioscopy was moderate (κ=0.45; 95% confidence interval (CI) 0.3–0.61 and κ=0.46; 95% CI 0.3–0.63, for each grader) and between experts and gonioscopy was moderate (κ=0.49; 95% CI 0.35–0.65 and κ=0.58; 95% CI 0.43–0.74, for each grader). Agreement between non-expert observers was moderate (κ=0.44; 95% CI 0.28–0.6) and between expert observers was excellent (κ=0.83; 95% CI 0.66–1.0). The diagnostic performance of the expert grader was better for detecting angle closure than the non-expert (AUC 0.85 vs. 0.78, respectively; P=0.019).
Although the level of expertise did not affect the agreement between observers and gonioscopy, the higher concordance between expert observers in comparison with non-expert observers, may explain the higher diagnostic performance for detecting angle closure in the glaucoma-trained graders. This suggests that more training may be helpful for assessing ITC in Swept-Source AS-OCT imaging.
This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.
This PDF is available to Subscribers Only