Why this review matters for people with RP
Retinitis pigmentosa (RP) is not one disease but a group of inherited retinal conditions in which light-sensing cells, called photoreceptors, progressively lose function. For patients and families, the central challenge is that vision loss may continue even when the underlying genetic cause is known. Researchers are therefore pursuing approaches that could protect remaining retinal cells, replace damaged cells, or potentially restore part of retinal function.
A 2026 review in Cells, “Artificial Intelligence as a Catalyst for Cell-Based Therapies in Retinal Disease,” examines how artificial intelligence (AI) could help make cell-based retinal treatments more consistent, personalized, and practical to deliver. The review considers RP alongside age-related macular degeneration and Stargardt disease. Its central message is hopeful but measured: AI may become valuable supporting infrastructure for retinal cell therapy, but it has not yet been proven necessary or clinically validated as part of these treatments.
Cell therapies and the hurdles they face
Cell-based therapies aim to address retinal degeneration by using living cells. These may include retinal pigment epithelium (RPE) cells, which support photoreceptors, or cells derived from induced pluripotent stem cells (iPSCs). iPSCs are adult cells that have been reprogrammed into a stem-cell-like state and can then be directed toward retinal cell types.
For RP, such strategies may someday be relevant when photoreceptors or their supporting environment have been severely affected. Some approaches may seek to preserve surviving retinal tissue; others could aim to replace lost cells. Gene editing tools such as CRISPR-Cas9 may also have a role in correcting particular disease-causing genetic changes before cells are used for therapy.
However, producing and transplanting cells is complex. Researchers must determine which patients are most likely to benefit, make cell products consistently, verify that cells have developed as intended, perform delicate retinal surgery, and monitor whether transplanted cells remain safe and functional over time. The review identifies these steps as major areas where AI tools could help.
Where AI could fit into the RP cell-therapy pathway
AI is a broad term that includes machine learning, deep learning, and computer vision—systems that identify patterns in large and complicated datasets. In retinal care, these methods may be useful at multiple points.
Better patient profiling and selection
People with RP can have different genetic causes, rates of progression, retinal structures, and remaining vision. AI could potentially combine genetic information, visual function testing, and retinal images to identify meaningful patient subgroups. This could help researchers design more focused clinical studies and, eventually, match a person to the approach most appropriate for their retinal condition.
Improving stem-cell production
A cell therapy depends on the quality of the cells being made. AI-assisted image analysis could help evaluate whether stem cells are growing and differentiating into the desired retinal cell type in a consistent way. This may support quality control and improve manufacturing at larger scale, where variation between batches is an important concern.
Assessing retinal organoids and gene editing
Retinal organoids—three-dimensional laboratory-grown tissues that model aspects of retinal development—are increasingly used to study disease and test therapies. Computer vision could assist in characterizing their structure and development.
The review also highlights AI’s possible role in evaluating CRISPR-Cas9 editing, particularly in identifying potential “off-target” changes: unintended edits elsewhere in the genome. Reducing these risks is essential if edited cells are to be considered for clinical use.
Monitoring after transplantation
Retinal imaging produces detailed information that can be difficult to assess consistently across many visits and patients. AI could potentially analyze multimodal imaging—such as scans that show retinal layers and photographs that document retinal appearance—to track changes after a transplant. This may help detect structural changes, evaluate graft-related features, and connect imaging findings with visual outcomes over time.
Important limits: promise is not proof
This publication is a review and a forward-looking framework, not evidence that an AI-enabled cell therapy for RP is ready for routine care. The authors carefully distinguish direct evidence in retinal cell therapy from related work in other retinal fields, other medical specialties, and conceptual proposals.
Several obstacles remain. Retinal cell-therapy datasets are often small, while AI systems generally need large, high-quality, well-labeled datasets. Algorithms must also be interpretable enough for clinicians, researchers, and regulators to understand how they reach conclusions. Privacy, regulatory requirements, reproducibility, and prospective clinical validation will all be critical.
Most importantly, AI cannot replace the biological foundations of treatment development. A model may identify patterns, but rigorous laboratory studies, careful surgical methods, and well-designed clinical trials are still needed to establish safety and benefit.
What this could mean for the future of RP research
The review proposes an end-to-end AI-supported pathway: patient profiling could inform gene editing and cell production; quality-control data could guide transplantation; and long-term imaging and visual outcomes could feed back into future treatment decisions. This connected approach is an appealing vision for personalized retinal medicine, but it remains a framework that must be tested in real clinical settings.
For the RP community, the near-term significance is that AI may help researchers overcome practical bottlenecks that have slowed complex cell-based treatments. Over time, better ways to select participants, standardize cell products, and monitor outcomes could make studies more informative and therapies more reproducible. The field is moving toward a future in which biological insight, advanced imaging, genetics, cell science, and responsible AI work together—not as substitutes for one another, but as complementary tools in the effort to preserve and restore vision.