System and method for early detection of diabetic retinopathy using optical coherence tomography

摘要

A system for the imaging, processing and evaluation of tissues provides prognostic and diagnostic details regarding diseased tissue. A set of quantitative measures were developed and integrated in an image-base analysis software tool designed for OCT images. The system and methods in this invention is significant because it allows assessing the optical properties and structure morphology differences between normal healthy subjects and diabetic patients with retinopathy up to ETDRS level 35 and without retinopathy.

主权项

1. A method of predicting early diabetic retinopathy comprising:
obtaining thickness values from optical coherence tomogram (OCT) data for a retina of a subject; detecting whether an abnormal pattern is present in the thickness values based on a comparison of the thickness values to a reference data set; upon detecting the abnormal pattern in the thickness values, predicting early diabetic retinopathy in the subject based on a weighted predictor function calculated by a processor as:PEDR⁡(MD,MC)=min⁢∑i=1N⁢wi×d⁡(MDi,MCi)wherein wi is a weight coefficient (wi>=0, i=1 . . . N), d is a distance function; MDi, MCi are measured values of criteria comprising thickness, reflectance, scattering coefficient, and texture measures for the retina of the subject and normal healthy subjects, respectively; N is the number of criteria considered in the analysis (N=4); and PEDR is a natural distance measure for similarity search where w, is the extent to which MDi is matched to MCi.