Polarization sensitive OCT of retinal pathologies

The human retina is organized in several layers that perform various tasks of the complex visual process. Pathologic changes of these layers can cause various retinal diseases that can lead to vision loss and ultimately to blindness. Diagnosis of retinal diseases and understanding of the associated pathologic changes has greatly benefited from recent advances in retinal imaging. Especially optical coherence tomography (OCT) provides considerable insight into morphologic changes associated with retinal disorders by recording cross sectional images of the retina with unprecedented resolution. However, commercially available OCT has two drawbacks: it is rather slow, typically recording one cross sectional (B-scan) image per second - too slow for 3D imaging, and it provides only intensity based images. Two recent developments considerably improve OCT technology: (i) Spectral domain (SD) OCT is a very sensitive and fast variant of OCT that allows 3D image acquisition in a few seconds. (ii) Polarization sensitive (PS) OCT provides additional image contrast, allowing to visualize structures that were difficult to discern in intensity based images, and also providing quantitative information on certain tissue parameters, e.g., birefringence. We recently demonstrated that both technologies can be combined to a spectral domain PS-OCT technique that provides 3D polarization sensitive images of the human retina. This project is designed to evaluate and investigate the usefulness of the new SD PS-OCT technique for the purposes of clinical ophthalmology. In a first step, an improved instrument with a stable patient interface and improved alignment and documentation capabilities will be built. With this instrument, an extensive survey of the polarization properties of the normal human retina, especially in the macula and nerve head area, shall be performed. For this purpose, PS-OCT retinal imaging shall be performed in a representative group of healthy subjects of various ages, gender, and races. The images and data thus obtained shall be used to establish a data base of normal PS-OCT images against which the images and data obtained in patients can be compared. The main task of the project will be to evaluate the feasibility of SD PS-OCT for detection, early diagnosis, and therapy control of some of the most widespread vision threatening diseases. For this purpose, extensive imaging in patients with major retinal diseases: age related macular degeneration, other retinal pigment epithelium related or exudative retinal diseases, glaucoma, etc., will be performed. These results shall be compared to the data base obtained in normals to evaluate the diagnostic capabilities of PS-OCT. On a long term, this might help to reduce visual impairment in the affected population.

The purpose of the project was to develop a new instrument for retinal diagnostics, polarization sensitive optical coherence tomography (PS-OCT), and to evaluate its performance for imaging in patients with various retinal diseases. The human retina is organized in several layers that perform various tasks of the complex visual process. Pathologic changes of these layers can cause various retinal diseases that can lead to vision loss and ultimately to blindness. Diagnosis of retinal diseases and understanding of the associated pathologic changes has greatly benefited from recent advances in retinal imaging. Especially optical coherence tomography (OCT) provides considerable insight into morphologic changes associated with retinal disorders by recording cross sectional images of the retina with unprecedented resolution. Two recent developments have considerably improved OCT technology: (i) spectral domain (SD) OCT improved sensitivity and imaging speed by more than two orders of magnitude; (ii) PS-OCT provides intrinsic, tissue specific contrast and quantitative measurement of tissue parameters like birefringence and depolarization. In the frame of the project, a new PS-OCT instrument based on SD technology was developed that combines a robust, stable patient interface and user-friendly system control software. The instrument was used to image healthy subjects and patients with various diseases. We could demonstrate that two retinal layers that play decisive roles in the most important vision threatening diseases can be imaged with superior contrast and quantitatively analysed: the retinal pigment epithelium, which is damaged in age-related macular disease (AMD), and the retinal nerve fiber layer which is damaged in glaucoma. In addition, hard exudates which are important biomarkers in diabetic retinopathy (DR) could be clearly identified by their polarization contrast. Based on these results, we developed new advanced algorithms that can automatically detect and quantify retinal lesions that are characteristic for AMD and DR. The algorithms were evaluated by comparing segmentation results to those obtained manually by expert graders, and the correlation was excellent. The precision of our algorithms was within a few per cent. These results fully supported our basic hypothesis that the intrinsic, tissue specific contrast provided by PS-OCT improves automated detection of retinal lesions in important, vision threatening diseases. Several hundred patients have been imaged up to now by PS-OCT. We presently continue to use the technology in several clinical trials to investigate disease progression and for therapy monitoring.