Identifying retinal complications of diabetes with autofluorescent imaging: A scoping review

Abstract

Background

It is estimated that approximately 460 million people world-wide are living with diabetes mellitus. The number of cases has increased over several years and is still expected to grow. As a consequence of diabetes mellitus, there are several acute and chronical complications, and the complications often occur after accumulation of lipids in the blood vessels (Frizziero et al., 2021). A common complication is diabetic retinopathy and diabetic macular oedema (DME), which is a leading cause of vision loss and irreversible blindness. Autofluorescent imaging may be a useful tool to detect and monitor different stages of diabetic retinopathy or diabetic macular oedema, due to lipofuscins property of emitting and absorbing light (Frizziero et al., 2021; Markan et al., 2020; Midena and Bini, 2016).

Methods

The study is a literature review with the retrieval of articles from five databases: Ovid (Medline), Embase, Cochrane, Cinahl complete and Web of science. The search was limited to articles published between the year 2000 – 2021, conducted on humans and written in English or a Scandinavian language. Two researchers screened the titles and abstracts independently, for thereafter one researcher read the full text of the included publications. The aim for this scoping review is to map the use of autofluorescent imaging in the retinal examination of patients with diabetes.

Results

The search in databases resulted in 1005 articles after removing duplicates, and after further screening it ended up with 44 research papers that was included in this scoping review. The literature review resulted in several different hallmarks on how to analyse and interpret fundus autofluorescence (FAF) images at patients with diabetic retinopathy or diabetic macular oedema. Several of the included studies compared different types of treatment for diabetic retinopathy or DME and evaluated these with FAF images. Many of the articles had similar results with retinal hypo-autofluorescent spots right after treatment. The same spots had become hyper-autofluorescent after one week and remained like this for months afterwards. Closer to a year after treatment, the spots became hypo-autofluorescent again. Depending on the treatment, lasers with lower intensity did not leave any noticeable signs at retina, evaluated with the use of FAF. However, the more conventional high-intensity lasers clearly resulted in visible scar tissue with the use of FAF. Distinct FAF patterns correlate to cystoid spaces, and several studies suggest a classification of the different patterns seen at FAF images. When using modern methods as fluorescence lifetime imaging ophthalmoscopy or flavoprotein autofluorescence, they all indicate a distinct longer autofluorescence within the patient group. It also seems that different wavelengths are more suitable to detect different pathological signs, also when compared to colour fundus photography.

Conclusion

Fundus autofluorescence is an easy-to-use, fast, non-invasive method which has the potential to detect and monitor several different stages of diabetes. It can also be used to confirm laser treatment, and in a case where the patient needs retreatment, it can assist with positioning the laser spots. The newer technique with fluorescence lifetime imaging ophthalmoscopy seems to be able to significantly differentiate between patients and healthy controls. For now, FAF seems to be a good ancillary test in detecting and monitoring of pathology, but further research and some enhancements of the quality, combined with a universal agreement on how to classify the image is still recommended before it is acknowledged as a stand-alone screening tool.