We developed an unsupervised machine learning algorithm and applied it to big corneal parameters to identify and monitor keratoconus stages. A big dataset of corneal swept source optical coherence tomography (OCT) images of 12,242 eyes acquired from SS-1000 CASIA OCT Imaging Systems in multiple centers across Japan was assembled. A total of 3,156 eyes with valid Ectasia Status Index (ESI) between zero and 100% were selected for the downstream analysis. Four hundred and twenty corneal topography, elevation, and pachymetry parameters (excluding ESI Keratoconus indices) were selected. The algorithm included three major steps. 1) Principal component analysis (PCA) was used to linearly reduce the dimensionality of the input data from 420 to eight significant principal components. 2) Manifold learning was used to further reducing the selected principal components nonlinearly to two eigen-parameters. 3) Finally, a density-based clustering was applied to the eigen-parameters to identify eyes with keratoconus. Visualization of clusters in 2-D space was used to validate the quality of learning subjectively and ESI was used to assess the accuracy of the identified clusters objectively. The proposed method identified four clusters; I: a cluster composed of mostly normal eyes (224 eyes with ESI equal to zero, 23 eyes with ESI between five and 29, and nine eyes with ESI greater than 29), II: a cluster composed of mostly healthy eyes and eyes with forme fruste keratoconus (1772 eyes with ESI equal to zero, 698 eyes with ESI between five and 29, and 117 eyes with ESI greater than 29), III: a cluster composed of mostly eyes with mild keratoconus stage (184 eyes with ESI greater than 29, 74 eyes with ESI between five and 29, and 6 eyes with ESI equal to zero), and IV: a cluster composed of eyes with mostly advanced keratoconus stage (80 eyes had ESI greater than 29 and 1 eye had ESI between five and 29). We found that keratoconus status and severity can be well identified using unsupervised machine learning algorithms along with linear and non-linear corneal data transformation. The proposed method can better identify and visualize the keratoconus stages.

Purpose: To investigate factors associated with cystoid macular edema (CME) after Descemet membrane endothelial keratoplasty (DMEK) in Asian eyes. Methods: In this retrospective, interventional, consecutive case series, 77 eyes of 65 patients who underwent DMEK were evaluated; in 53 eyes, cataract surgery was performed 1 month before DMEK (staged DMEK), and 24 eyes underwent DMEK alone (simple DMEK). Central retinal thickness, incidence of CME, postoperative best-corrected visual acuity, central corneal thickness, and corneal endothelial cell density were assessed at 1, 3, and 6 months after surgery. Multiple regression analysis and stepwise variable selection were performed for parameters such as type of surgery, iris damage scores, age, sex, axial length, preoperative visual acuity, rebubbling, air volume in the anterior chamber on postoperative day 1, history of diabetes, and endothelial cell density loss rates at 6 months after surgery. Results: CME occurred in 12 (15.6%) of 77 eyes. There was no significant difference in best-corrected visual acuity between eyes with and without CME (P = 0.27). Multivariable analysis revealed that the difference in iris damage scores between before and after DMEK (P < 0.001), air volume in the anterior chamber (P = 0.012), simple DMEK (P = 0.020), and rebubbling (P = 0.036) were significantly associated with CME. Stepwise variable selection indicated that iris damage (P < 0.001) was the most important risk factor for CME. Conclusions: Iris damage due to DMEK might be a possible risk and aggravating factor for the development of CME after DMEK. Surgeons should attempt to minimize damage to the iris.

To investigate the factors associated with endothelial survival after Descemet’s membrane endothelial keratoplasty (DMEK) in eyes of Asian patients with bullous keratopathy (BK). In this retrospective, consecutive interventional case series, 72 eyes of 72 patients who underwent DMEK were evaluated. Best corrected visual acuity (BCVA) and corneal endothelial cell density (ECD) were assessed at 12 months postoperatively. Multiple regression analysis was performed to assess parameters such as age, sex, axial length, preoperative visual acuity, re-bubbling, the ratio of graft to cornea area, iris damage scores, types of filling gases, air or SF₆ volume in the anterior chamber (AC) on postoperative day 1, and ECD loss rates at 12 months postoperatively. BCVA improved significantly at 12 months after DMEK (P < .001). The rate of ECD loss at 12 months after DMEK was 54.4 ± 16.1%. Multiple linear regression analysis showed that a larger ratio of graft to corneal area (P = 0.0061) and higher donor ECD (P = 0.042) were the primary factors for a lower ECD loss rate at 12 months after DMEK. A relatively larger graft size compared to the host cornea and more donor ECD might help endothelial survival in patients with BK. Moreover, for such patients, the surgeon should attempt to use a relatively larger graft size when performing DMEK, particularly in Asian eyes.

Purpose: To investigate the incidence of intraocular inflammation (IOI) and its risk factors following intravitreal injections of brolucizumab for neovascular age-related macular degeneration in Japan. Methods: A total of 1,351 Japanese consecutive patients with neovascular age-related macular degeneration who were treated with brolucizumab from May 2020 to May 2022 at 14 institutions were examined. The variables analyzed were the number of brolucizumab injections, time to onset of IOI, and risk factors. Results: Intraocular inflammation developed in 152 eyes (11.3%). Retinal vasculitis and/or retinal occlusion occurred in 53 eyes (3.9%). Ninety-four patients received bilaterally, bilateral IOI occurred in five patients (5.3%). Sixteen eyes (1.2%) had irreversible visual acuity loss and nine eyes (0.67%) had visual loss of three lines or more due to retinal vasculitis and/or retinal occlusion. The cumulative IOI incidence was 4.5%, 10.3%, and 12.2% at 30, 180, and 365 days (1-year), respectively. History of IOI (including retinal vasculitis) and/or retinal occlusion (odds ratio [OR], 5.41; P = 0.0075) and female sex (OR, 1.99; P = 0.0004) were significantly associated with IOI onset. Conclusion: The 1-year cumulative incidence of IOI in Japanese neovascular age-related macular degeneration patients treated with brolucizumab was 12.2%. History of IOI (including retinal vasculitis) and/or retinal occlusion and female sex were significant risk factors.

This single-center retrospective cohort study analyzed the 1-year real-world treatment outcomes of 63 consecutive eyes (of 60 patients) with neovascular age-related macular degeneration (nAMD) that were switched from intravitreal brolucizumab (IVBr) to intravitreal faricimab (IVF) and managed on a treat-and-extend regimen with discontinuation criteria. After the switch, patients opted to continue IVF, to switch back to IVBr, or receive photodynamic therapy (PDT). Thirty-eight patients continued IVF, 16 patients were switched back to IVBr, 2 patients received PDT, and 4 patients paused treatment. Best-corrected visual acuity (BCVA), central subfield thickness (CST), subfoveal choroidal thickness (sf-CT), and injection intervals were compared immediately before and 1 year after the initial IVF. Whereas there was no change in BCVA and CST; 0 [− 0.0969 to 0.125, P = 0.58], − 1.5 [− 27.8 to 13.5, P = 0.11] µm, respectively, sf-CT decreased significantly; − 19.5 [− 45.5 to 7.75, P = 0.015] µm. The patients switched back showed no significant change in sf-CT. The injection interval extended significantly in the IVF continuation and the switch-back group (2.0 and 3.0 weeks, respectively; [P = 0.0007 and 0.0078]) in eyes with a pre-switching interval of less than 12 weeks. Faricimab shows promise as a safe and effective alternative to brolucizumab for treating nAMD.