An Individualized Model to Predict and Monitor Glaucoma Progression Using Structural and Functional Data
Glaucoma is an eye disease that can lead to blindness. We propose a new model to detect glaucoma changes over time. This model will be tailored for each patient and tested in a different group of patients. We will use this model to predict which patients’ glaucoma is likely to progress—or get worse—over time.
The objective of this project is to validate the performance of a new model designed to monitor glaucoma progression. This model uses structural and functional data jointly and updates its predictions and determinations with each new pair of data that becomes available. The predictive accuracy of the model is better than that obtained with ordinary least square linear regression, particularly in short follow-up series.
The first aim of this project is to validate the model in a cohort independent from the original sample that was used to develop it. Thus, our laboratory will use the large cohort enrolled in the longitudinal Ocular Hypertension Treatment Study (OHTS) to test and optimize the model.
The second aim is to estimate the sensitivity and the specificity of the model using data from two longitudinal studies: the Diagnostic Innovations in Glaucoma Study (DIGS) and the African Descent and Glaucoma Evaluation Study (ADAGES). The ability of the model to tease out true progression from variability is being assessed by comparing its determination of progression to that of other clinically accepted structural and functional measures of progression. My laboratory is developing individualized methods to improve the detection of glaucoma progression.
The present study is innovative in that it makes use of both structural and functional data to monitor glaucoma progression. The model is unique in that it makes no assumptions about how glaucoma progresses or about the nature of the relationship between structure and function. The model is also individualized, which is important given the large variability observed between patients, and offers an intuitive graphical representation of the progression status of each patient. At the conclusion of this study, clinicians will have a powerful method to monitor glaucoma progression and to identify those patients who are likely to progress. This will improve patient care and lead to the preservation of sight in glaucoma patients. The model proposed in this study is also highly relevant to assess the effectiveness of new treatments designed to slow or halt the progression of glaucoma.