Rudolf Virchow, the father of modern pathology, stated that pathology is the very basis of medicine. Starting with morphology, pathology has gradually developed to include techniques of physiology, genetics, and molecular biology; today, it incorporates all of those disciplines and is referred to as general medicine. Pathological research provides a rich assortment of information. However, when performed for humans, it is difficult to obtain tissue samples, and limitations in terms of continuously observing changes are encountered. Recently, there have been major advances in the equipment used in ophthalmic examination. In particular, optical coherence tomography (OCT) has become an indispensable tool in the observation of retinochoroidal disease; this can be attributed to its ability to obtain high-resolution images within a short period. Further, the scanning laser ophthalmoscope (SLO), which facilitates ultra-wide field retinal imaging, has also been used. When combined with techniques, such as anterior ocular segment OCT, it is possible to capture a wide range of imaging data pertaining to the eyeball. It is important to try to apply pathology to clinical medicine, which are linked by imaging advances in the field of ophthalmology.
One of the major breakthroughs in imaging research achieved via the use of OCT is that it is now possible to obtain images of the living choroid. Initially begun with choroidal thickness analysis and then developed into structural analysis, it is now possible to analyze using en-face image with black and white conversion. Hence, pachychoroid spectrum disease, a new concept in age-related macular degeneration (AMD), has been suggested in this context. Lack of uniformity, etc., in the fundus owing to dilated choroidal (pachyvessels) and choroidal outflow tract vessels have been suggested to be involved with AMD pathogenesis. The choroid has abundant blood vessels, and although detailed analysis considered difficult, we have developed proprietary choroidal vessel structural analysis software using en-face image. The software made it posiible to quantify the choroidal vessel area and variations in vascular diameter and blood vessel orientation. Because we analyzed patients with normal eyes as well as those with central serous chorioretinopathy or polypoidal choroidal vasculopathy, which are pachychoroid spectrum diseases, we could quantify the dilation of vessels in the Haller's layer and the deterioration in blood vessel orientation symmetry.
There exist diurnal variations in the choroid, and changes in thickness can reportedly be achieved using treatments, such as anti-vascular endothelial growth factor therapy and photodynamic therapy. One of the reasons due to which chronological and cross-sectional choroidal analysis is difficult is the lack of a set method for determining the depth at which choroidal analysis should be performed to ensure reproducibility. Therefore, we used machine learning to study the feature quantity of choroidal en-face image, created a support vector regression model, and then, created a model to automatically determine the start points of the Sattler's and Haller's layers. Then, we conducted the analysis.
One of the basic tenets of medicine is "first do no harm." Recently, this idea is attracting much attention through the term "noninvasive." The Optos^Ⓡ SLO nakes it possible to simultaneously perform wide-field fundus OCT. Furthermore, because it is noninvasive, it reduses the load on the patient. We devised a method of noninvasively observing the choroidal large vessels images obtained at a wavelength of 635 nm. Although indocyanine green fundus angiography is said to be necessary for observing the choroidal vessels, the devised method provides an added advantage that the choroidal vessels can be observed repeatedly and noninvasively. We discovered a technique of observing vortex veins in patients with pachychoroid spectrum disease, who have received little attention to date, and were able to discover that the vortex veins of these patients expanded more than those in normal healthy eyes.
Currently, efforts to create databases of clinical information are ongoing worldwide. Further, with the recent advances in artificial intelligence, much attention is being given to how this technology will change the face of medicine in the future. In the field of ophthalmology, image information can play a central role, and the creation of outstanding image databases will contribute to major advances in medicine in Japan. This research can help contribute to the formulation of the basic principles and to the technology that will be involved.
Nippon Ganka Gakkai Zasshi (J Jpn Ophthalmol Soc) 123: 260-283, 2019.