UC Berkeley

Clinically significant macular edema (CSME) results from the microvascular complications of diabetes in the central retina of the eye. Diabetes-induced disregulation of the intra- and extra-vascular fluid homeostasis leads to the accumulation of fluid in the retinal tissue. When left untreated, CSME can result in vision loss by disrupting normal metabolic processes, ischemia and by mechanical disruption of the intricate microstructure of the retina. CSME can be succesfully treated when detected in time. Vision loss caused by CSME is largely irreversible and, therefore, the emphasis is on early detection. Challenges in CSME detection result in a significant number of untreated patients with devastating visual consequences. Tele-medicine screening which relies on either monoscopic or stereoscopic fundus photographs is widely implemented and has good sensitivity and specificity for the detection of retinopathy levels. Contrary to that, CSME detection in fundus photographs faces substantial challenges. Stereoscopic photographs require dilation which increases the risk of vision loss due to an elevated intraocular pressure which may lead to angle closre glaucoma and is generally time consuming and incovenient for patients which further decreases compliance with screening. Non-dilated stereophotography suffers from unacceptably high number of stereophotographs of poor quality. Monoscopic images have low sensitivity for CSME detection because of the lack of stereopsis. Programs which use monoscopic images must rely on surrogate markers of CSME. There is currently no consensus on how to detect CSME in monoscopic images. Different programs use different surrogate markers primarily because there is lack of evidence in scientific literature regarding the unequivocal validity of a particular CSME detection method.

This dissertation includes three studies that share a common goal of improving the accuracy of CSME detection. The anticipated impact of this research is that we will be able to improve the efficiency with which we identify patients requiring treatment, thereby increasing their chance for reducing the risk of vision loss due to CSME.

The first study presented in Chapter 3 of this dissertation evaluates the ability of hard exudates, a surrogate marker of DME, located within one-disc diameter from the center of the macula to detect CSME. This study tests the “real world” implementation of this approach in a diabetic retinopathy screening program and attempts to resolve the discrepancies regarding the accuracy of this method that exists in the literature. The second study presented in Chapter 4 of this dissertation tests whether the proximity of hard exudates to the fovea is associated with more severe cases of CSME and thus warrants a more expedited referral and intervention. In addition, we proposed an OCT-based adaptation of CSME severity scale based on the DME severity scale derived from the Early Treatment Diabetic Retinopathy Study (ETDRS) data. This is a step forward towards developing a more repeatable and objective structural measure of sight-threatening diabetic macular edema which correlates with functional measures of retinal health, such as visual acuity and electroretinographic recordings. Finally, the third study presented in Chapter 5 builds on the findings of the first two studies and utilizes a novel approach for the detection of CSME. This approach is based on the combined measure of the proximity of hard exudates to the fovea and the areal extent of exudation in the central macula. We have also evaluated the association of the photopic 30 Hz flicker ERG measurements with CSME. Finally, the model of CSME detection which combines relevant clinical variables was evaluated as the CSME detection tool.

The results of the studies presented in this dissertation support the use of a new approach – radially arranged sectors in the central macula - for the detection of CSME. This approach was shown to accurately classify patients with and without CSME. Moreover, the increase in the number of sectors affected by hard exudates is also associated with the increase in probability of having severe CSME, justifying a more expedited referral for treatment. Finally, while significantly associated with CSME, the latency of the 30 Hz photopic flicker ERG response does not seem to offer additional power to discriminate between patients with and without CSME.