Optic disc edema can arise from a variety of possible causes, some benign and some life threatening. For timely appropriate medical intervention, or to reduce patient anxiety in the event none is needed, it is critical that the cause and severity of a swollen optic disc be determined as soon as possible.
In this doctoral work, several algorithms are pieced together to determine the cause of optic disc edema. The process of determining the cause of swelling involves the extraction of several features, many of which are relatively new to the field of ophthalmology. Included among these are the shape of Bruch's membrane as found semi-automatically from SD-OCT images and the presence and orientation of folds in the retina, which are also most visible in SD-OCT images, and some selected features from 2D fundus images.
One specific cause of optic disc edema, called papilledema, is due to raised intracranial pressure. This, too, has a variety of possible causes, and often urgency (or severity) is rated by the Frisén scale, which is a 0-5 ordinal rating of severity (with 0 being normal). In the event papilledema is found to be the cause of swelling, this doctoral work also seeks to implement a more robust measurement of severity than the Frisén scale. Specifically, the total retinal volume (TRV) of the optic disc has been computed from SD-OCT images in other work. It is believed that the TRV serves as a more reliable means of assessing papilledema severity, as it is a continuous, repeatable measurement that is not subject to observer interpretation. As part of this doctoral work, the TRV is estimated from fundus images, which are faster and cheaper to obtain than SD-OCT images.
Thus, the aims of this thesis consist of finding and quantifying folds in the retina, using folds (and other features) to distinguish between the causes of a swollen optic disc, and, in the event an optic disc is swollen due to papilledema, to assess the severity of the swelling by estimating the TRV from fundus images. While the ultimate goal of this work would be to entirely diagnose a patient with optic disc edema from fundus images, that is beyond the scope of a single thesis. Thus the efforts here are to build towards that.