Search

BackgroundThe purpose of this study was to develop a model that can predict the postoperative visual acuity in eyes that had undergone vitrectomy for an epiretinal membrane (ERM). The Light Gradient Boosting Machine (LightGBM) was used to evaluate the accuracy of the prediction and the contribution of the explanatory variables. Two models were designed to predict the postoperative visual acuity in 67 ERM patients. Model 1 used the age, sex, affected eye, axial length, preoperative visual acuity, Govetto's classification stage, and OCT-derived vector information as features to predict the visual acuity at 1, 3, and 6 months postoperatively. Model 2 incorporated the early postoperative visual acuity as an additional variable to predict the visual acuity at 3, and 6 months postoperatively. LightGBM with 100 iterations of 5-fold cross-validation was used to tune the hyperparameters and train the model. This involved addressing multicollinearity and selecting the explanatory variables. The generalized performance of these models was evaluated using the root mean squared error (RMSE) in a 5-fold cross-validation, and the contributions of the explanatory variables were visualized using the average Shapley Additive exPlanations (SHAP) values.ResultsThe RMSEs for the predicted visual acuity of Model 1 were 0.14 ± 0.02 logMAR units at 1 month, 0.12 ± 0.03 logMAR units at 3 months, and 0.13 ± 0.04 logMAR units at 6 months. High SHAP values were observed for the preoperative visual acuity and the ectopic inner foveal layer (EIFL) area with significant and positive correlations across all models. Model 2 that incorporated the postoperative visual acuity was used to predict the visual acuity at 3 and 6 months, and it had superior accuracy with RMSEs of 0.10 ± 0.02 logMAR units at 3 months and 0.10 ± 0.04 logMAR units at 6 months. High SHAP values were observed for the postoperative visual acuity in Model 2.ConclusionPredicting the postoperative visual acuity in ERM patients is possible using the preoperative clinical data and OCT images with LightGBM. The contribution of the explanatory variables can be visualized using the SHAP values, and the accuracy of the prediction models improved when the postoperative visual acuity is included as an explanatory variable. Our data-driven machine learning models reveal that preoperative visual acuity and the size of the EIFL significantly influence postoperative visual acuity. Early intervention may be crucial for achieving favorable visual outcomes in eyes with an ERM.

PurposeTo investigate the clinical features, treatment options, and visual outcomes of submacular hemorrhage (SMH) secondary to neovascular age-related macular degeneration (nAMD).DesignA retrospective, observational case series.MethodsSetting: Multicenter institutional setting. Patient Population: A total of 127 patients (127 eyes; 88 men, 39 women; (mean age, 74.2 years)) diagnosed with AMD-associated SMHs exceeding 2 disc diameters involving the fovea. Observation: The AMD types, previous treatments, treatment options, anatomic findings, and best-corrected visual acuity (BCVA) were assessed. Main Outcome Measures: Clinical features, treatment options, and visual outcomes of SMHs secondary to nAMD.ResultsThirty-two eyes had typical AMD, 94 eyes polypoidal choroidal vasculopathy (PCV), and one eye retinal angiomatous proliferation. Eighty-five eyes were treatment-naïve; 42 eyes were treated previously: anti-vascular endothelial growth factor (VEGF) therapy (n = 26), photodynamic therapy (n = 3), and combined therapy (n = 13). Treatment of SMHs included vitrectomy (36 eyes), pneumatic displacement (49 eyes), and anti-VEGF monotherapy (42 eyes). The final BCVA improved significantly in treatment-naïve cases from 0.86 to 0.62 logarithm of the minimal angle of resolution (logMAR) unit (Snellen equivalent from 20/145 to 20/83) and from 0.80 to 0.56 (Snellen equivalent from 20/126 to 20/73) in PCV cases. Meanwhile, the BCVA logMAR values improved from 1.15 to 0.75 (Snellen equivalent from 20/283 to 20/112) and from 0.87 to 0.63 (Snellen equivalent from 20/148 to 20/85) in eyes that underwent vitrectomy or pneumatic displacement, respectively. In eyes with BCVAs between 20/133 to 20/40 at SMH onset, the final VA in the pneumatic displacement group was better than in the anti-VEGF monotherapy group. One eye had a retinal detachment and 1 eye had a macular hole in the vitrectomy group, and 5 eyes had a vitreous hemorrhage in the pneumatic displacement group.ConclusionsThe recommended treatment for SMHs secondary to nAMD exceeding 2 disc area and with BCVA below 20/40 is vitrectomy or pneumatic displacement for visual improvement.

PurposeTo evaluate the clinical characteristics, treatment trends, and visual prognosis of submacular hemorrhage (SMH) secondary to neovascular age-related macular degeneration (nAMD) and retinal arterial macroaneurysm (RAM).MethodsThis retrospective study enrolled 187 Japanese patients with SMH at 10 institutions from 2015 to 2018. Medical records including SMH etiology, best-corrected visual acuity (BCVA), fundus photographs, optical coherence tomography images, and selected treatments were analyzed.ResultsMajor causes of SMH were typical nAMD (tnAMD) (18%), polypoidal choroidal vasculopathy (PCV) (50%) and RAM (29%). Age, male/female ratio, baseline BCVA, central retinal thickness, and involved retinal layers were significantly different between etiologies (all PConclusionsThe characteristics of, and treatment strategy for, SMH were different between the underlying diseases. Anti-vascular endothelial growth factor treatment with or without expansile gas was mainly chosen for SMH in tnAMD and PCV, whereas vitrectomy with gas was the most common treatment for RAM, and the higher rate for vitrectomy might result in the greater BCVA improvement in the RAM group than in the other groups. Final BCVA was better in PCV, RAM, and tnAMD, in that order, because patients with PCV were younger and had better baseline BCVA.

PURPOSE: To study the changes in the outer retinal microstructures during a six month period after the onset of acute zonal occult outer retinopathy (AZOOR)-complex by spectral-domain optical coherence tomography (SD-OCT). METHODS: Seventeen eyes of 17 patients with the AZOOR-complex were studied. The integrity of the external limiting membrane (ELM), ellipsoid zone (EZ; also called the inner/outer segment junction), and interdigitation zone (IDZ; also called the cone outer segment tips) were evaluated in the SD-OCT images obtained at the initial visit and at six months. The three highly reflective bands were divided into three types; continuous, discontinuous, and absent. The integrity of the outer nuclear layer (ONL) was also assessed. RESULTS: Among the three highly reflective bands, the IDZ was most altered at the initial visit and least recovered at six months. Fifteen of 17 eyes (88%) had a recovery of at least one of the three bands at six months in the retinal area where the ONL was intact, and these areas showed an improvement of visual field. Three eyes (18%) had retinal areas where the ONL was absent at the initial visit, and there was no recovery in both the retinal structures and visual fields in these areas. CONCLUSIONS: Our results indicate that more than 85% eyes with AZOOR-complex show some recovery in the microstructures of the outer retina during a six month period if the ONL is intact. We conclude that SD-OCT is a useful method to monitor the changes of the outer retinal microstructure in eyes with the AZOOR-complex.