Post by Alexander Stupnicki

👀 From Headlines to Clinics: Can artificial intelligence transform how we diagnose and manage Fuchs Endothelial Corneal Dystrophy (FECD)? That’s the question we set out to explore in our systematic review, now published in Translational Vision Science & Technology (TVST). 🔗 Current Applications of Artificial Intelligence for Fuchs Endothelial Corneal Dystrophy: A Systematic Review: https://lnkd.in/dTNxucXB 📌 What we found: - Use cases; segmentation of corneal endothelium, detection and quantification of corneal oedema and prediction of graft detachment after corneal transplantation (mostly DMEK). - Most models achieved strong performance metrics (AUC > 0.90, Dice > 0.85), but generalisability remains a major issue, with only 3/19 studies performing external validation. - Many models were trained on small, single-centre datasets, raising concerns about bias and reproducibility. 📈 What this means for the future: There’s huge potential for AI to support earlier and more accurate diagnosis. But to get there, we need: ✅ Large, diverse datasets ✅ External and prospective validation ✅ Explainable AI (XAI) to support clinician trust ✅ Integration of multimodal data (imaging, clinical, genomic) I’m incredibly grateful to Assoc. Prof. Nikolas Pontikos, Shane (Siyin) Liu, and Lynn Kandakji for the opportunity to contribute to this project and learn from such a brilliant team at the UCL Institute of Ophthalmology and NIHR Moorfields Biomedical Research Centre. Excited to see how these insights shape the future of AI-assisted ophthalmic care. #AIinHealthcare #Ophthalmology #Cornea #FECD #DeepLearning #MedicalImaging #DigitalHealth #AcademicResearch #VisionScience #MachineLearning