Search

PurposeTo compare performance of independently developed deep learning algorithms for detecting glaucoma from fundus photographs and to evaluate strategies for incorporating new data into models.MethodsTwo fundus photograph datasets from the Diagnostic Innovations in Glaucoma Study/African Descent and Glaucoma Evaluation Study and Matsue Red Cross Hospital were used to independently develop deep learning algorithms for detection of glaucoma at the University of California, San Diego, and the University of Tokyo. We compared three versions of the University of California, San Diego, and University of Tokyo models: original (no retraining), sequential (retraining only on new data), and combined (training on combined data). Independent datasets were used to test the algorithms.ResultsThe original University of California, San Diego and University of Tokyo models performed similarly (area under the receiver operating characteristic curve = 0.96 and 0.97, respectively) for detection of glaucoma in the Matsue Red Cross Hospital dataset, but not the Diagnostic Innovations in Glaucoma Study/African Descent and Glaucoma Evaluation Study data (0.79 and 0.92; P < .001), respectively. Model performance was higher when classifying moderate-to-severe compared with mild disease (area under the receiver operating characteristic curve = 0.98 and 0.91; P < .001), respectively. Models trained with the combined strategy generally had better performance across all datasets than the original strategy.ConclusionsDeep learning glaucoma detection can achieve high accuracy across diverse datasets with appropriate training strategies. Because model performance was influenced by the severity of disease, labeling, training strategies, and population characteristics, reporting accuracy stratified by relevant covariates is important for cross study comparisons.Translational relevanceHigh sensitivity and specificity of deep learning algorithms for moderate-to-severe glaucoma across diverse populations suggest a role for artificial intelligence in the detection of glaucoma in primary care.

Ripasudil, a rho-associated protein kinase inhibitor ophthalmic solution, shows a protective effect in preventing excessive scarring in vitro. This study aims to evaluate the safety and efficacy of ripasudil for glaucoma patients submitted to the needling procedure. In this prospective, multicenter, single-arm study, we included 20 eyes of 20 patients with glaucoma who underwent the needling procedure without antimetabolites. All patients administered ripasudil after needling for three months. The primary endpoint of this study was the safety of ripasudil in patients, and the secondary endpoint was the change in IOP at 12 weeks after the needling procedure. No serious complications were found in the patients. One eye experienced pruritus and conjunctival follicle, while another eye had conjunctival follicle. These complications were transient and resolved quickly after discontinuation of ripasudil. The mean preoperative IOP was 14.6 ± 4.6 mmHg, which decreased to 11.0 ± 4.7 mmHg (p = 0.0062) at 1 week postoperatively. The IOP reduction effect continued to 12 weeks (11.8 ± 3.1 mmHg; p = 0.0448). The administration of the ROCK inhibitor, ripasudil, after the needling procedure is safe and effective in maintaining IOP for 12 weeks. [ABSTRACT FROM AUTHOR]

Purposes: To investigate associations between Corvis ST-measured corneal biomechanical parameters and glaucomatous optic nerve head (ONH) morphology. Methods: In total, 118 eyes of 70 patients with open-angle glaucoma were examined in this retrospective cross-sectional study. We measured Heidelberg retina tomograph and Corvis ST values in all eyes. We used the linear mixed model in four sectors (temporal superior, TS; temporal inferior, TI; nasal superior, NS; and nasal inferior, NI) to detect associations between six ONH-related parameters and 14 Corvis ST–related parameters, controlling for age, intraocular pressure, axial length, and central corneal thickness. We calculated the ONH temporal and nasal sector vertical asymmetries (TS-TI and NS-NI asymmetries) and identified the optimal linear mixed models to describe them using model selection with the second-order bias corrected Akaike Information Criterion. Results: The Corvis ST A2 velocity was negatively associated with the rim volume in the NS sector (p < 0.05). The optimal model for TS-TI asymmetry was TS-TI asymmetry = − 3.22 + 0.15 × HC time + 0.88 × HC deflection amplitude, whereas that for NS-NI asymmetry was 0.49–0.048 × axial length − 2.45 × A2 velocity. Conclusion: Glaucomatous ONH superior-inferior asymmetries were associated with biomechanical properties measured with Corvis ST. Eyes with superior-dominant rim volume reduction of ONH were associated with small deformations and slow recovery of the cornea. [ABSTRACT FROM AUTHOR]

There are numerous surgical procedures for glaucoma. Minimally invasive glaucoma surgery is becoming popular; however, the disadvantage is the high incidence of anterior chamber hemorrhage. Heavy bleeding can also lead to increased intraocular pressure (IOP) postoperatively. Gonio scratch is a surgical procedure that improves aqueous humor outflow by rubbing off deposits on the trabecular meshwork with a Diamond Dusted Sweeper. As the conjunctiva and trabecular meshwork are not incised, no postoperative bleeding is expected, and the IOP spike will be minimal. We designed this study to determine the efficacy and safety of gonio scratch. This is an on-going multicenter, prospective, clinical trial. Patients who are scheduled for glaucoma surgery with or without cataract surgery are being enrolled. A total of 80 eyes will be recruited in the Hiroshima University Hospital, Miyoshi Eye Clinic, Yokoyama Retina Clinic, and Kusatsu Eye Clinic. All patients will undergo gonio scratch. When combined with cataract surgery, gonio scratch is performed after the intraocular lens is inserted. The primary study endpoint is the change in IOP from baseline to 1 year after surgery. The secondary endpoints are complications, number of glaucoma medications, surgical time, and changes in visual acuity and the visual field. This study protocol was approved by the institutional review board of Hiroshima University. The trial results will be shared with the scientific community at international conferences and by publication in a peer-reviewed journal. Trial registration number is jRCTs062200003., Competing Interests: None.

Importance: Exfoliation syndrome is a systemic disorder characterized by progressive accumulation of abnormal fibrillar protein aggregates manifesting clinically in the anterior chamber of the eye. This disorder is the most commonly known cause of glaucoma and a major cause of irreversible blindness., Objective: To determine if exfoliation syndrome is associated with rare, protein-changing variants predicted to impair protein function., Design, Setting, and Participants: A 2-stage, case-control, whole-exome sequencing association study with a discovery cohort and 2 independently ascertained validation cohorts. Study participants from 14 countries were enrolled between February 1999 and December 2019. The date of last clinical follow-up was December 2019. Affected individuals had exfoliation material on anterior segment structures of at least 1 eye as visualized by slit lamp examination. Unaffected individuals had no signs of exfoliation syndrome., Exposures: Rare, coding-sequence genetic variants predicted to be damaging by bioinformatic algorithms trained to recognize alterations that impair protein function., Main Outcomes and Measures: The primary outcome was the presence of exfoliation syndrome. Exome-wide significance for detected variants was defined as P < 2.5 × 10-6. The secondary outcomes included biochemical enzymatic assays and gene expression analyses., Results: The discovery cohort included 4028 participants with exfoliation syndrome (median age, 78 years [interquartile range, 73-83 years]; 2377 [59.0%] women) and 5638 participants without exfoliation syndrome (median age, 72 years [interquartile range, 65-78 years]; 3159 [56.0%] women). In the discovery cohort, persons with exfoliation syndrome, compared with those without exfoliation syndrome, were significantly more likely to carry damaging CYP39A1 variants (1.3% vs 0.30%, respectively; odds ratio, 3.55 [95% CI, 2.07-6.10]; P = 6.1 × 10-7). This outcome was validated in 2 independent cohorts. The first validation cohort included 2337 individuals with exfoliation syndrome (median age, 74 years; 1132 women; n = 1934 with demographic data) and 2813 individuals without exfoliation syndrome (median age, 72 years; 1287 women; n = 2421 with demographic data). The second validation cohort included 1663 individuals with exfoliation syndrome (median age, 75 years; 587 women; n = 1064 with demographic data) and 3962 individuals without exfoliation syndrome (median age, 74 years; 951 women; n = 1555 with demographic data). Of the individuals from both validation cohorts, 5.2% with exfoliation syndrome carried CYP39A1 damaging alleles vs 3.1% without exfoliation syndrome (odds ratio, 1.82 [95% CI, 1.47-2.26]; P < .001). Biochemical assays classified 34 of 42 damaging CYP39A1 alleles as functionally deficient (median reduction in enzymatic activity compared with wild-type CYP39A1, 94.4% [interquartile range, 78.7%-98.2%] for the 34 deficient variants). CYP39A1 transcript expression was 47% lower (95% CI, 30%-64% lower; P < .001) in ciliary body tissues from individuals with exfoliation syndrome compared with individuals without exfoliation syndrome., Conclusions and Relevance: In this whole-exome sequencing case-control study, presence of exfoliation syndrome was significantly associated with carriage of functionally deficient CYP39A1 sequence variants. Further research is needed to understand the clinical implications of these findings.

Purposes: To investigate associations between Corvis ST-measured corneal biomechanical parameters and glaucomatous optic nerve head (ONH) morphology., Methods: In total, 118 eyes of 70 patients with open-angle glaucoma were examined in this retrospective cross-sectional study. We measured Heidelberg retina tomograph and Corvis ST values in all eyes. We used the linear mixed model in four sectors (temporal superior, TS; temporal inferior, TI; nasal superior, NS; and nasal inferior, NI) to detect associations between six ONH-related parameters and 14 Corvis ST-related parameters, controlling for age, intraocular pressure, axial length, and central corneal thickness. We calculated the ONH temporal and nasal sector vertical asymmetries (TS-TI and NS-NI asymmetries) and identified the optimal linear mixed models to describe them using model selection with the second-order bias corrected Akaike Information Criterion., Results: The Corvis ST A2 velocity was negatively associated with the rim volume in the NS sector (p < 0.05). The optimal model for TS-TI asymmetry was TS-TI asymmetry = - 3.22 + 0.15 × HC time + 0.88 × HC deflection amplitude, whereas that for NS-NI asymmetry was 0.49-0.048 × axial length - 2.45 × A2 velocity., Conclusion: Glaucomatous ONH superior-inferior asymmetries were associated with biomechanical properties measured with Corvis ST. Eyes with superior-dominant rim volume reduction of ONH were associated with small deformations and slow recovery of the cornea.