Fuchs' endothelial corneal dystrophy (FECD) increases the risk of corneal decompensation after cataract surgery, often leading to endothelial keratoplasty (EK). Reliable prediction of EK is essential for surgical planning, yet traditional corneal biomarkers have limited prognostic value. We present a novel deep learning framework based on Scheimpflug tomographic imaging to improve FECD prognosis. The approach integrates three components to address data limitations: (1) clinical domain knowledge, (2) ensemble learning, and (3) self-supervised learning (SSL). A hybrid convolutional neural network (CNN) is proposed, combining a RANSAC-based algorithm for estimating best-fit sphere (BFS) elevation maps with a dual-branch design that incorporates Polar Pooling to mimic clinical reasoning. Robustness is enhanced through bootstrap aggregation, where multiple models are trained on different data subsets and their predictions averaged. To further reduce reliance on annotated data, we introduce a self-supervised contrastive pretraining task. This task distinguishes images from the same eye across time, using a weighted contrastive loss that emphasizes corneal irregularities. On a multi-center dataset, the framework achieved an AUC of 0.94, outperforming biomarker-based models and demonstrating strong generalizability across hospitals. The proposed methodology provides a robust and interpretable decision-support tool for FECD management and cataract surgery planning.
23 July 2026 in Archivos de la Sociedad Española de Oftalmología
To evaluate the clinical value of postoperative examinations at 24 hours and 30 days following uncomplicated phacoemulsification, and to identify subgroups who may safely benefit from remote or simplified follow-up pathways. A retrospective observational study was conducted, including all patients who underwent uncomplicated phacoemulsification at a high-volume tertiary public hospital during September 2024. Demographic data, ophthalmologic history, surgical characteristics, day-1 and day-30 findings, visual outcomes, and emergency visits were extracted from electronic medical records. Postoperative complications associations were evaluated using bivariate tests and multivariate binomial logistic regression. Of 296 patients, the most common 24-hour complication was transient intraocular pressure elevation (13.2%). Overall, 20.9% showed a postoperative finding on day 1, and 14.2% required unplanned reassessment. At 30 days, complications occurred in 6.7% of cases, mainly cystoid macular edema (3.6%). In multivariate analysis, 24-hour complications (P=0.028) and ophthalmologic comorbidities (P=0.010) independently predicted 30-day complications. Patients without ophthalmologic history, without 24-hour complications, and with early visual improvement had a markedly lower risk of adverse outcomes. Applying a risk-based follow-up strategy could reduce approximately 50% of routine 30-day in-person reviews and relieve up to 300 hours of annual clinical workload. This risk-stratified postoperative model may improve service efficiency in high-volume public hospitals, support the sustainability of ophthalmic care, and may be transferable to other public health systems. Low-risk patients could be safely monitored remotely without compromising clinical outcomes. Prospective studies are required to confirm these findings.