Glaucoma is a multifactorial disease where various systemic features are involved in the progression of the disease. Based on initial systemic profiles in electronic medical records, this study aimed to develop a model predicting factors of long-term rapid retinal nerve fiber layer (RNFL) thinning over 5 years in 505 patients with primary open-angle glaucoma. Eyes with faster or slower RNFL thinning were stratified using a decision tree model, and systemic and ophthalmic data were incorporated into the models based on random forest and permutation methods, with the models interpreted by Shapley additive explanation plots (SHAP). According to the decision tree, a higher lymphocyte ratio (> 34.65%) was the most important systemic variable discriminating faster or slower RNFL thinning. Higher mean corpuscular hemoglobin (> 32.05 pg) and alkaline phosphatase (> 88.0 IU/L) concentrations were distinguishing factors in the eyes with lymphocyte ratios > 34.65% and < 34.65%, respectively. SHAP demonstrated larger baseline RNFL thickness, greater fluctuation of intraocular pressure (IOP), and higher maximum IOP as the strongest ophthalmic factors, while higher lymphocyte ratio and higher platelet count as the strongest systemic factors associated with faster RNFL thinning. Machine learning-based modeling identified several systemic factors as well as previously acknowledged ophthalmic risk factors associated with long-term rapid RNFL thinning.
|Publication status||Published - 2023 Dec|
Bibliographical noteFunding Information:
Supported by Seoul National University Bundang Hospital Research Fund (no. 14-2022-0024), the Patient-Centered Clinical Research Coordinating Center, funded by the Ministry of Health & Welfare, Republic of Korea (Grant Nos. HI19C0481, HC19C0276), and by the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science, and Technology (No. 2016R1D1A1B02011696). The funding organizations had no role in the design or conduct of this research.
© 2023, The Author(s).
All Science Journal Classification (ASJC) codes