An Artificial Intelligence Enabled System for Retinal Nerve Fiber Layer Thickness Damage Severity Staging.

Purpose: To develop an objective glaucoma damage severity classification system based on OCT-derived retinal nerve fiber layer (RNFL) thickness measurements. Design: Algorithm development for RNFL damage severity classification based on multicenter OCT data. Subjects and Participants: A total of 6561 circumpapillary RNFL profiles from 2269 eyes of 1171 subjects to develop models, and 2505 RNFL profiles from 1099 eyes of 900 subjects to validate models. Methods: We developed an unsupervised k-means model to identify clusters of eyes with similar RNFL thickness profiles. We annotated the clusters based on their respective global RNFL thickness. We computed the optimal global RNFL thickness thresholds that discriminated different severity levels based on Bayes' minimum error principle. We validated the proposed pipeline based on an independent validation dataset with 2505 RNFL profiles from 1099 eyes of 900 subjects. Main Outcome Measures: Accuracy, area under the receiver operating characteristic curve, and confusion matrix. Results: The k-means clustering discovered 4 clusters with 1382, 1613, 1727, and 1839 samples with mean (standard deviation) global RNFL thickness of 58.3 (8.9) µm, 78.9 (6.7) µm, 87.7 (8.2) µm, and 101.5 (7.9) µm. The Bayes' minimum error classifier identified optimal global RNFL values of > 95 µm, 86 to 95 µm, 70 to 85 µm, and < 70 µm for discriminating normal eyes and eyes at the early, moderate, and advanced stages of RNFL thickness loss, respectively. About 4% of normal eyes and 98% of eyes with advanced RNFL loss had either global, or ≥ 1 quadrant, RNFL thickness outside of normal limits provided by the OCT instrument. Conclusions: Unsupervised machine learning discovered that the optimal RNFL thresholds for separating normal eyes and eyes with early, moderate, and advanced RNFL loss were 95 µm, 85 µm, and 70 µm, respectively. This RNFL loss classification system is unbiased as there was no preassumption or human expert intervention in the development process. Additionally, it is objective, easy to use, and consistent, which may augment glaucoma research and day-to-day clinical practice. Financial Disclosures: Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Characteristics of a Large, Labeled Data Set for the Training of Artificial Intelligence for Glaucoma Screening with Fundus Photographs.

Purpose: Significant visual impairment due to glaucoma is largely caused by the disease being detected too late. Objective: To build a labeled data set for training artificial intelligence (AI) algorithms for glaucoma screening by fundus photography, to assess the accuracy of the graders, and to characterize the features of all eyes with referable glaucoma (RG). Design: Cross-sectional study. Subjects: Color fundus photographs (CFPs) of 113 893 eyes of 60 357 individuals were obtained from EyePACS, California, United States, from a population screening program for diabetic retinopathy. Methods: Carefully selected graders (ophthalmologists and optometrists) graded the images. To qualify, they had to pass the European Optic Disc Assessment Trial optic disc assessment with ≥ 85% accuracy and 92% specificity. Of 90 candidates, 30 passed. Each image of the EyePACS set was then scored by varying random pairs of graders as "RG," "no referable glaucoma (NRG)," or "ungradable (UG)." In case of disagreement, a glaucoma specialist made the final grading. Referable glaucoma was scored if visual field damage was expected. In case of RG, graders were instructed to mark up to 10 relevant glaucomatous features. Main Outcome Measures: Qualitative features in eyes with RG. Results: The performance of each grader was monitored; if the sensitivity and specificity dropped below 80% and 95%, respectively (the final grade served as reference), they exited the study and their gradings were redone by other graders. In all, 20 graders qualified; their mean sensitivity and specificity (standard deviation [SD]) were 85.6% (5.7) and 96.1% (2.8), respectively. The 2 graders agreed in 92.45% of the images (Gwet's AC2, expressing the inter-rater reliability, was 0.917). Of all gradings, the sensitivity and specificity (95% confidence interval) were 86.0 (85.2-86.7)% and 96.4 (96.3-96.5)%, respectively. Of all gradable eyes (n = 111 183; 97.62%) the prevalence of RG was 4.38%. The most common features of RG were the appearance of the neuroretinal rim (NRR) inferiorly and superiorly. Conclusions: A large data set of CFPs was put together of sufficient quality to develop AI screening solutions for glaucoma. The most common features of RG were the appearance of the NRR inferiorly and superiorly. Disc hemorrhages were a rare feature of RG. Financial Disclosures: Proprietary or commercial disclosure may be found after the references.

Ocular motility changes and diplopia in sutured versus unsutured implantation of the Baerveldt glaucoma device.

PURPOSE: To investigate the effect of two surgical techniques in primary Baerveldt glaucoma implant (BGI) surgery, that is the sutured technique and the unsutured (free) plate technique, on the ocular motility and prevalence of diplopia. We hypothesize that the free plate technique results in a lower diplopia prevalence. METHODS: We performed a prospective study of patients who underwent BGI surgery with the free plate technique and compared them with patients from a previous study who had undergone BGI surgery with the sutured technique. Their ductions, ocular alignment and fusion range and the prevalence of diplopia were measured before surgery and at 3 months, 6 months and 1 year postoperatively. RESULTS: We analysed 57 free plate and 51 sutured plate patients. One year postoperatively, we found no statistically significant difference in the prevalence of diplopia between the two techniques. All duction changes between baseline and 1-year follow-up were restrictions and occurred statistically significantly more frequently in the free plate than in the sutured plate group (p = 0.03; 60% versus 34%). About the ocular alignment, in the horizontal direction, a change in exodirection was more common in both groups, while in the vertical direction, a hyperdeviation of the operated eye was more common. The vertical ocular alignment change was smaller in the free plate group than in the sutured plate group (p = 0.04 at near and p = 0.02 at distance). CONCLUSIONS: One year postoperatively, the prevalence of diplopia was not significantly different between patients with the sutured plate and patients with the free plate technique. Both surgical techniques induce diplopia and changes in ocular motility and/or in ocular alignment.

Motility Changes and Diplopia After Baerveldt Glaucoma Drainage Device Implantation or After Trabeculectomy.

PURPOSE: The purpose of this study was to quantify any diplopia and motility changes after the implantation of a Baerveldt glaucoma implant (BGI) or after trabeculectomy (TE). METHODS: We analyzed 51 patients with a BGI and 52 patients with a TE from a prospective cohort study. To quantify any diplopia, we asked patients about the presence of diplopia at 1 year after surgery. To quantify any ocular motility changes, we measured ductions in eight gaze directions, the patients' ocular alignment and their fusion range before and 1 year after surgery. RESULTS: In the BGI group, 14 patients (28%) experienced diplopia compared with one patient (2%) in the TE group (P < 0.001). Duction changes were more commonly observed in the BGI group (35%) than in the TE group (19%). In the BGI group, ductions were mostly restricted in elevation (13%; P < 0.001), in abduction (13%), in elevation in 25° adduction (13%; P = 0.044), and in elevation in 25° abduction (25%; P < 0.001). In 32% of the patients, their near horizontal ocular alignment shifted, notably in exodirection (P = 0.04). The fusion range decreased significantly in the horizontal direction (-12.6° ± 10.3°, mean ± standard deviation; P = 0.01). CONCLUSIONS: BGI surgery was significantly associated with postoperative diplopia and impaired eye motility (reduced ductions), mostly present in abduction, elevation, elevation in 25° adduction, and elevation in 25° abduction. Even without impaired ductions, diplopia could come about. TRANSLATIONAL RELEVANCE: By studying diplopia across glaucoma patients prospectively with diplopia questionnaires and extensive orthoptic measurements, we gain better insight into its occurrence.

Modeling of scanning laser polarimetry images of the human retina for progression detection of glaucoma.

The development of methods to detect slowly progressing diseases is often hampered by the time-consuming acquisition of a sufficiently large data set. In this paper, a method is presented to model the change in images acquired by scanning laser polarimetry, for the detection of glaucomatous progression. The model is based on image series of 23 healthy eyes and incorporates colored noise, incomplete cornea compensation and masking by the retinal blood vessels. Additionally, two methods for detecting progression, taking either one or two follow-up visits into account, are discussed and tested on these simulated images. Both methods are based on Student's t-tests, morphological operations and anisotropic filtering. The images simulated by the model are visually pleasing, show corresponding statistical properties to the real images and are used to optimize the detection methods. The results show that detecting progression based on two follow-up visits greatly improves the sensitivity without adversely affecting the specificity.