
FLORetina-ICOOR 2026: AI moves from the image to the trial, the clinic, and the operating room
Six instructional courses at this year's congress in Florence follow artificial intelligence through retinal imaging, drug development, geographic atrophy care, and robotic surgery.
When the retina community meets in Florence next month, artificial intelligence (AI) will be more than a single session topic. The FLORetina International Congress on OCT Angiography, En Face OCT, and Advances in OCT (FLORetina-ICOOR) runs December 3 to 6 at the Fortezza da Basso.1 The course listings published so far include 2 courses tagged AI and 4 more under an AI and robotics heading.2,3
Taken together, the courses suggest a shift in how the field is talking about AI. At last year's meeting, much of the discussion was about what algorithms could do.4,5 The 2026 courses spend more time on how to use AI tools in practice: how to validate them, put them into clinic workflows, build them into trial endpoints, and explain them to patients.
The groundwork: Pearls and pitfalls
The course "Pearls and Pitfalls of Artificial Intelligence for Retinal Research and Practice," co-chaired by Daniela Ferrara, MD, PhD, FASRS, and Usha Chakravarthy, MD, PhD, and co-moderated by Giovanni Staurenghi ,MD, is a broad starting point.2 Its faculty covers the main areas where AI is used in retina. Stela Vujosevic, MD, PhD, FEBO, will speak on OCT-based algorithms and Daniel Ting, MD, PhD, on color fundus photography. Pearse Keane, MD, FRCOphth, will discuss oculomics, the use of retinal imaging as a window to systemic health. Valeria De Filippis, MD, PhD, will cover AI-based mining of electronic medical records, and Nadia Waheed will address AI in retinal research and clinical trials.2
Lajos Csincsik will discuss validation approaches for clinical imaging algorithms.2 The course description says that implementing AI successfully depends on careful validation, including whether results generalize across devices and populations, and on meeting regulatory standards. It also names multimodal and agentic AI as emerging approaches.2 For clinicians judging the AI products now being marketed to them, the validation talk may be the most practically useful part of the course.
Digital twins and synthetic control arms
A second Ferrara-led course, co-chaired with Waheed, looks at drug development. "Artificial Intelligence for Clinical Trials in Retinal Diseases: From Imaging Biomarkers to Digital Twins" makes a strong claim in its description: AI is becoming a mandatory technology in drug development for retinal diseases.2 The talks include imaging biomarkers as trial endpoints (Waheed), treatment response prediction in exudative macular disease (Virginia Mares, MD, MSc, PhD), and GA progression prediction (Oliver Leingang, PhD).2
The later talks look further ahead. Ferrara will present on AI-enabled digital twins, Hrvoje Bogunovic, PhD, on external control arms and real-world data, and Marion Munk, MD, PhD, on the move toward fully digital, in silico trials.2 According to the organizers, digital twins could make synthetic control arms possible, especially in early-phase trials. They also note barriers, including data standards, governance, cybersecurity, and regulatory frameworks they describe as outdated.2 Anat Loewenstein, MD, MHA, who discussed home OCT monitoring at FLORetina 2025,⁵ will cover AI-enabled home OCT monitoring in clinical trials. That talk comes as home-based platforms such as Notal Vision's SCANLY, which received FDA De Novo authorization in 2024, move into research use.6
SAIVO's second year: implementation
SAIVO was founded in May 2025 and ran its first course at FLORetina last year.4 Its 2026 symposium, "Implementing AI in retina and posterior segment diseases: triumphs and challenges," is again moderated by founding members Munk and Quan Dong Nguyen.3 This year the talks are about deployment. Mario Romano, MD, PhD, will discuss lessons from Vision Code 2026. Niranjan Manoharan, MD, will present an imaging-integrated prescreening pipeline for GA trial recruitment, and Roomasa Channa, MD, will describe putting an automated GA area algorithm into clinical workflows. Rita Wolf, MD, will report on validation and outcomes of autonomous AI-based retinopathy screening in children.3
The program also extends past the macula. Talks cover AI quantification of hyperreflective vitreous particles in uveitis, human-in-the-loop triage of choroidal lesions for possible melanoma on ultra-widefield imaging, and foundation models for external-eye imaging in oculoplastics.3 The symposium closes with Mary K. Daly, MD, "Informed Consent for AI in an Epistemic Void."3 The title is a reminder that patients are often asked to accept AI-supported decisions that even their physicians cannot fully explain.
A stepwise roadmap for geographic atrophy
For clinicians who want something they can use right away, "AI in Clinical Practice: A Practical Patient-Centered Roadmap for Geographic Atrophy Management," organized by Roy Schwartz, MD, and Caroline Baumal, MD, may be the most applied course of the meeting.3 It follows the GA care pathway in 5 steps: differential diagnosis against genetic mimics (Dinah Zur, MD), AI-based risk stratification (Gregor Reiter, MD, MSc, PhD), measuring and monitoring from baseline to treatment response (Aude Ambresin, MD, PhD), functional monitoring (Baumal), and predicting individual patient trajectories (Schwartz). Omer Trivizki, MD, MBA, closes the course with a look at future needs.3 The organizers promise live demonstrations and real-world cases.3
The timing makes sense. The United States now has 2 approved GA therapies, but specialists have pointed out that AI-driven imaging analysis for GA has not been cleared for routine US clinical practice, while some such tools are already used in Europe.7 Ophthalmology Times has also reported on investigational AI tools, such as functional OCT, that aim to estimate retinal sensitivity from structural scans. Those tools have not yet been cleared by regulators.8 The course may help show how much of this AI-supported GA workflow is available in clinics today and how much is still in development.
AI and robotics in the vitreoretinal OR
The two surgical courses round out the program. "Artificial Intelligence in Vitreoretinal Surgery: A crash course," organized by Rodrigo Anguita Henriquez, MD, and Martin Zinkernagel, MD, PhD, is written for surgeons who say they lack practical AI training.3 It starts with the basics of machine learning and neural networks (Philippe Valmaggia, MD) and asks whether no-code AI is ready for clinicians (Lorenzo Ferro Desideri, MD). It then covers proliferative diabetic retinopathy (Aude Couturier, MD), predictive modeling (Zinkernagel), and real-time computer vision for intraoperative guidance, presented by congress president Stanislao Rizzo, MD.3 Anguita Henriquez, MD, MBBS ends with "From generative to agentic AI: the road to surgical intelligence."3
"Robotics, AI and the Future of Eye Surgery," organized by Jean-Pierre Hubschman, MD, may be the most forward-looking session of the meeting.3 Fanny Nerinckx, MD, will present first-in-human experience with the Luca robotic system. In 2024, Luca was used to perform 7 vitrectomies at Ghent University Hospital in Belgium, and the company reported no device-related adverse events.9,10 Nerinckx has said the platform operates with 20 times less tremor than a human hand.⁹ Other speakers include Marc de Smet, MD, CM, PhD, BSc(hon), FARVO, FRCSC, FRCOphth, FEBOphth, FMH, on moving robotic platforms from the lab to the operating room, Khalid Al Sabti on building a new robotic system, Rodolfo Mastropasqua, MD, PhD, FEBO, on intraoperative OCT and robotic automation, and Aya Barzelay Wollman, MD, PhD, on machine learning in robotic instruments. Steven Schwartz will address barriers to adoption.3 The course ends with a panel whose title asks the question many surgeons and administrators will have: "Is robotic eye surgery ready—and who will pay for it?"3
What to watch
Across all 6 courses, the faculty seem less interested in whether AI works than in what it takes to deploy it responsibly. That means validation across devices and populations, workflow integration, regulatory pathways, reimbursement, and patient consent. Ophthalmology Times will be on site in Florence to cover the AI sessions and the rest of the FLORetina-ICOOR 2026 program.
References
FLORetina ICOOR 2026. Floretina ICOOR Meeting. Accessed November 18, 2026. https://floretinaicoormeeting.com/
Discover our courses: AI. FLORetina ICOOR 2026. Accessed November 18, 2026. https://floretinaicoormeeting.com/courses?tag=AI
Discover our courses: Artificial Intelligence & Robotics. FLORetina ICOOR 2026. Accessed November 18, 2026. https://floretinaicoormeeting.com/courses?tag=Artificial+Intelligence+%26+Robotics
FLORetina 2025: AI in practice, discussed by Marion Munk and Quan Dong Nguyen. Ophthalmology Times. December 17, 2025. Accessed November 18, 2026. https://www.ophthalmologytimes.com/view/floretina-2025-ai-in-practice-discussed-by-marion-munk-and-quan-dong-nguyen
FLORetina 2025 recap: the art of retina on display in Florence. Ophthalmology Times. February 6, 2026. Accessed November 18, 2026. https://www.ophthalmologytimes.com/view/floretina-2025-recap-the-art-of-retina-on-display-in-florence
Positive results published from pivotal study of Notal Vision's AI algorithm in home-based OCT system. Optometry Times. Accessed November 18, 2026. https://optometrytimes.com/view/positive-results-published-from-pivotal-study-of-notal-vision-s-ai-algorithm-in-home-based-oct-system
Could AI improve currently available treatments for geographic atrophy? Healio. March 16, 2026. Accessed November 18, 2026. https://www.healio.com/news/ophthalmology/20260320/could-ai-improve-currently-available-treatments-for-geographic-atrophy
Functional OCT takes retinal imaging where structure alone cannot go. Ophthalmology Times. July 25, 2026. Accessed November 18, 2026. https://www.ophthalmologytimes.com/view/functional-oct-takes-retinal-imaging-where-structure-alone-cannot-go
Robotic system tested in retinal surgery. Retinal Physician. July/August 2025. Accessed November 18, 2026. https://retinalphysician.com/issues/2025/julyaugust/robotic-system-tested-in-retinal-surgery
AcuSurgical. V- LUCA: from first surgery to a new era in retinal precision. Accessed November 18, 2026. https://acusurgical.com/v-luca-from-first-surgery-to-a-new-era-in-retinal-precision/
Related to this article








