An auxiliary role of deep neural network ophthalmic disease identification models in choosing medication treatment strategy.

Abstract

To evaluated a Deep Neural Network (DNN)-based ophthalmic disease diagnosis framework in facilitating personalized medication treatment plans using a prospective, single-center, randomized controlled clinical trial design. 500 patients were randomly assigned to either a DNN-aided experimental group or a control group receiving standard physician-based treatment plans. The primary outcomes were medication selection accuracy, clinical efficacy (assessed by BCVA and CMT), patient compliance, and adverse reaction management. Results showed that DNN-aided treatment plans significantly improved medication selection accuracy and treatment quality, with higher BCVA and CMT scores in the experimental group. Patients in the experimental group also demonstrated higher compliance and a trend towards lower adverse reaction rates. The study highlights the potential of DNN models to enhance ophthalmic disease management, offering precise and personalized treatment strategies with potential benefits for patient outcomes and safety as AI technology advances.