Utilizing a responsive web portal for studying disc tracing agreement in retinal images.

Glaucoma is a leading cause of blindness worldwide whose detection is based on multiple factors, including measuring the cup to disc ratio, retinal nerve fiber layer and visual field defects. Advances in image processing and machine learning have allowed the development of automated approached for segmenting objects from fundus images. However, to build a robust system, a reliable ground truth dataset is required for proper training and validation of the model. In this study, we investigate the level of agreement in properly detecting the retinal disc in fundus images using an online portal built for such purposes. Two Doctors of Optometry independently traced the discs for 159 fundus images obtained from publicly available datasets using a purpose-built online portal. Additionally, we studied the effectiveness of ellipse fitting in handling misalignments in tracing. We measured tracing precision, interobserver variability, and average boundary distance between the results provided by ophthalmologists, and optometrist tracing. We also studied whether ellipse fitting has a positive or negative impact on properly detecting disc boundaries. The overall agreement between the optometrists in terms of locating the disc region in these images was 0.87. However, we found that there was a fair agreement on the disc border with kappa = 0.21. Disagreements were mainly in fundus images obtained from glaucomatous patients. The resulting dataset was deemed to be an acceptable ground truth dataset for training a validation of models for automatic detection of objects in fundus images.

Face and content validity of an artificial eye model for Ab-Interno Goniotomy.

PURPOSE: To determine the face and content validity of an artificial eye model for ab-interno goniotomy (SimulEYE KDB model, InsEYEt, Westlake Village, CA) by surveying ophthalmologists with varying experience using a Kahook Dual Blade (KDB; New World Medical, Rancho Cucamonga, CA, USA) following a 90-min wet-lab course using the model. PARTICIPANTS: Overall 13 ophthalmologists participated following a surgical simulation session on goniotomy using the goniotomy blade at the 2019 Canadian Ophthalmological Society annual meeting. METHODS: A 17-question survey to assess the face and content validity of the model was given immediately following the surgical simulation session on goniotomy using the goniotomy blade. Responses to each survey question were recorded on a 5-point Likert scale ranging from (1) strongly agree to (5) strongly disagree. RESULTS: Respondents rated statements regarding the model with a median response of 1 (Strongly Agree) to 3 (Neither agree or disagree). Mann-Whitney U nonparametric analysis revealed no significant difference in responses between instructor vs. non-instructor or between prior experience vs. no prior experience for any of the survey statements. The model received highest survey ratings for utility in training residents, acquisition of surgical skills, accessibility, and higher likelihood of success with the procedure than theory and observation alone. Lowest ratings were for realism of the model compared to a human cadaveric eye. CONCLUSION: Our results suggest the SimulEYE KDB model is a reasonably cost-effective solution for simulating angle-based surgeries. Additionally, our project shows that experienced ophthalmologists found the artificial eye models useful and helpful for angle-based surgery training.