Artificial intelligence across the cataract care continuum: a systematic scoping review introducing the CAICC framework and the surgical complexity prediction gap

Background Cataracts remain a leading cause of reversible blindness, and artificial intelligence (AI) applications now span the cataract care pathway. However, the distribution of these applications and their validation maturity remain unclear. This review mapped AI across the care continuum, summarized task-specific performance, characterized validation maturity, and identified evidence gaps. Methods The Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews (PRISMA-ScR) was followed. PubMed, Scopus, the Cochrane Library, WorldCat, and the American Academy of Ophthalmology Meeting Archive were searched; the final search was completed on 20 July 2025. Eligible studies were English-language primary research published from 1 January 2020 through 20 July 2025. A review-specific cataract artificial intelligence clinical continuum (CAICC) evidence-mapping framework was developed during data charting and synthesis. Two reviewers independently conducted study selection, data extraction, endpoint adjudication, and validation-maturity assessment. Performance was summarized descriptively; validation maturity characterized evaluation independence and clinical stage, not risk of bias or overall methodological quality. Results Of 361 reports sought for retrieval, 122 (33.8%) were not retrieved; 113/122 (92.6%) were considered plausibly eligible from available metadata. The final synthesis included 185 studies: detection ( n = 52), characterization ( n = 50), personalized surgical planning ( n = 27), intelligent surgery ( n = 28), recovery ( n = 24), and patient engagement ( n = 4). Among 181 validation-applicable included studies, internal testing was the most frequent highest-maturity category (123/181, 68.0%), followed by external testing (33/181, 18.2%), prospective single-center evaluation (6/181, 3.3%), and prospective multicenter evaluation (2/181, 1.1%); 17/181 (9.4%) had no independent or unclear testing. Among the retrieved and included studies, no comprehensive clinically validated multimodal preoperative cataract surgical-complexity model was identified. Conclusions Most included studies remained at internal validation stages, and prospective multicenter evaluation was rare. Because 122/361 reports sought for retrieval were unavailable and most were considered plausibly eligible, retrieval bias may affect the comprehensiveness of the evidence map. Future research should prioritize multimodal preoperative surgical-complexity prediction, independent external testing, and prospective multicenter evaluation of locked models. Registration Open Science Framework (OSF), DOI 10.17605/OSF.IO/ZKM86.