A better understanding of various diseases that affect the retina depends partially on an improved in vivo visualization of small retinal changes on a cellular level. However, state of the art optical imaging technologies are limited by imperfections of the eye optics and imaging with sufficient resolution for visualizing individual cells is not possible. In order to overcome this limitation, an active correction of aberrations introduced by imperfections of the eye (mainly the anterior segment) using adaptive optics (AO) technology is required. By manipulating the phase of the imaging light at an imaging plane conjugated to the pupil plane of the eye, unwanted distortions can be removed resulting in a diffraction limited transversal resolution at the posterior segment of the eye (retina).
The improved transverse resolution allows visualizing structures on a micrometer-scale such as cone photoreceptors or microvasculature. The cells and vasculature that build up the retina are organized in layered thin structures and represents another challenge for imaging of this organ. The semi-transparent property of the retina is an additional challenge as a high sensitivity of any imaging technology is essential in order to visualize it. In recent years, a non-invasive interferometric imaging technique has been developed that is perfectly suited to image the retina. The method is called optical coherence tomography (OCT) and provides intensity profiles in depth with micrometer axial resolution. The combination of AO with OCT is very attractive for retinal imaging as high transverse resolution is provided by AO and high axial resolution is achieved by OCT leading to high isotropic image resolution in all three dimensions. The attempt to combine these two techniques has been performed successfully in the past showing a great improvement on image quality. Despite the improved image quality provided by AO-OCT the impact in clinical routine is still questionable as these instruments usually are expensive, bulky and difficult to use. The work performed within the framework of this thesis addresses technical challenges for translating AO-OCT technology into a clinical environment and on the path investigates advantages of AO-OCT in comparison to state of the art retinal imaging methods.
In the first paper of this thesis, a compact AO-OCT instrument is presented, which combines a fundus camera and a spectrometer based OCT system, both powered with AO correction. The proposed combination allows an easy determination and localization of regions of interest in the retina with the aid of the AO fundus camera by changing the location of a fixation target during real time image display. Once the region of interest is found, the system can then be switched into an AO-OCT mode that allows recording of three-dimensional image data of this region. Thanks to the internal fixation target and the overview AO-fundus camera display the same location on the retina can be found at each patient visit that paves the way for longitudinal studies. The system was used to image patients with different diseases such as diabetic retinopathy (DR), age related macular degeneration (AMD) at different stages, Stargardts disease, central serous retinopathy (CSR), among others.
In the second paper of the thesis, the capabilities of the prototype introduced in the first paper were increased by implementation of OCT angiography. Thereby several OCT cross sectional images are recorded at exactly the same location on the retina in order to visualize vessels. As a result, high-resolution angiograms of the retinal vasculature are obtained. The work resulted in interesting findings concerning the improvement of vessel contrast and reduction of image projection artefacts. Finally, a comparison of AO-OCT vessel images with images produced by commercial instruments is presented together with first AO-OCT angiography in patients.
In the third and final paper of the thesis, a new AO-OCT instrument operating in the 1050nm wavelength band is introduced. The optical design was based on the experience gained by operation of the first prototype. One drawback of the initial prototype was the need of complementary data such as wide field OCT volumes in order to determine and localize regions of interest. This need was eliminated by building an instrument capable of acquiring image data in two different modes: A large field of view (FoV) OCT mode and a high-resolution AO-OCT mode (with limited FoV). This dual imaging capability allows the direct comparison of retinal images recorded with different transverse resolutions. The influence of transverse resolution is of specific interest, as it was shown in the second paper, for the interpretation of angiographic data recorded with OCT.