RESUMO
PURPOSE: To develop a new diabetic retinopathy severity scoring system and to determine if it can monitor changes from baseline as well as identify precise features that have changed over time. Such a grading system could potentially provide an understanding of the impact of treatments utilizing an algorithmic scoring technique. METHODS: The traditional ETDRS grading system was examined and a flow algorithm based on the grading approach was created. All visual comparative assessment points, relying on identification of features in relation to prior standard photographic images, were evaluated and quantified. A new grading form was created that provided fields that captured all relevant features required for determining the ETDRS grading score. A computer software algorithm was developed that examines all entered fields and calculates the appropriate diabetic severity score. RESULTS: This diabetic retinopathy scoring algorithm system was successful in generating a severity score comparable to traditional methods of grading images. Validation with traditionally graded images was performed, demonstrating that in a majority of cases, the severity scores were comparable. The algorithmic grading system was then used to analyze images obtained in a large clinical study of diabetic macular edema, resulting in data regarding baseline scoring values, as well as detailed features of the microvasculature that drove the severity scoring results, and changes seen during the trial. CONCLUSION: This new algorithmic diabetic severity scoring system provides a means to monitor the progression or regression of retinopathy with therapeutic intervention as well as assess the individual microvascular features that may be modified over the course of treatment.