Artificial intelligence applications facilitate decision-making in cataract surgery for highly myopic patients

Abstract

Background: Surgical decision-making for highly myopic cataracts requires a high level of expertise. We, therefore, aimed to develop a preliminary artificial intelligence (AI) model for surgical decision-making in highly myopic cataracts, based on previous deep learning models. Materials and methods: We first established a highly myopic cataract decision-making AI model by integrating cataract grading and postoperative visual acuity prediction models of highly myopic eyes, which we had developed previously, with surgical decision logic. The outcomes of surgical decision-making were classified into four categories: surgery not advised, cataract surgery recommended, retinal surgery recommended, and combined cataract–retinal surgery recommended. The gold standard for surgical decision is defined as the decision jointly made by two professional ophthalmologists together (X.Z. and Y.W.). If the decision-makings regarding highly myopic cataract surgery were not fully consistent, a final judgment was made by a third expert (Y.L.). Subsequently, we evaluated the accuracy of AI model’s surgical decision-making against the gold standard and doctors at different levels, using both internal (107 highly myopic eyes from Eye and ENT Hospital, Fudan University) and external (55 highly myopic eyes from Wuhan Aier Eye Hospital) test datasets. Results: In the internal and external datasets, according to the Lens Opacities Classification System (LOCS) III international standards for cataract grading, 99.07% and 87.27% of automatic nuclear grading, along with 88.79% and 61.82% of automatic cortical grading, respectively, had an absolute prediction error of ≤1.0 compared with the gold standard. The mean postoperative visual acuity prediction error was 0.1560 and 0.3057 logMAR in the internal and external datasets, respectively. Finally, the consistency of the AI model’s surgical decisions with the gold standard for highly myopic cataract patients in the internal and external datasets was 96.26% and 81.82%, respectively. AI demonstrated substantial agreement with the gold standard (Kappa value = 0.811 and 0.556 in the internal and external datasets, respectively). Conclusion: The AI decision-making model for highly myopic cataracts, based on two deep learning models, demonstrated good performance and may assist doctors in complex surgical decision-making for highly myopic cataracts.