Contribución al análisis de registros de electrorretinografía multifocal para detección precoz de glaucoma

Abstract

Glaucomatous optic neuropathy is one of the most prevalent chronic eye diseases, affecting 5 % of the world population between 40 and 60 years, and causing irreversible blindness in the patient. This is the reason that prioritizes early detection of the disease by means of sensitive and objective diagnostic methods. Current techniques used in clinic are subjective and globally show the status of the patient’s retina. Visual examination by the physician (tonometry, automated perimetry, etc.) is often carried out, so it does not show signs of glaucomatous retina in early stages of the disease. Some of these techniques, such as perimetry, provide results that depend on the subjective perception of the patient. The three characteristics that traditionally defined the presence of glaucoma in a subject were: increased intraocular pressure, changes in the optic disc and visual field defects. More recently, it has been shown that there can be a significant loss of retinal ganglion cells and optic nerve fibers, before any signs of functional loss appear in the conventional visual campimetry test. Other studies have shown that although elevated intraocular pressure is one of the main risk factors for glaucoma, glaucomatous eyes not always have shown a high pressure. Some subjects also can tolerate a higher intraocular pressure before the loss of nerve fibers by glaucoma occurs. Therefore, the presence of elevated intraocular pressure alone is not sufficient for the diagnosis of glaucoma. The multifocal electroretinogram (mfERG) is a novel technique that can provide objective information and a sensitivity map of the retina, with a high topographic resolution. However, by using the parameters that are currently being studied with the multifocal technique, i.e., amplitudes and latencies of the retinal response, it is not possible to discriminate the response of the inner layers of the retina, especially involved in the glaucoma process. In this thesis, signals from mfERG records are studied, using various mathematical techniques not applied so far in this clinical field. These techniques are the advanced morphological analysis of signals and the wavelet transform, including its continuous, discrete and wavelet packet versions. Through the morphological analysis it is possible to extract a large set of parameters from mfERG signals, which allows a better characterization of the retinal response, and then classify them as healthy or glaucomatous using a neural network. Moreover, the application of the wavelet transform to these mfERG records allows us to obtain a set of markers for glaucoma, in the presence of subtle frequency changes in early stages. In all of the cases, the sensitivity and specificity values obtained are higher than those obtained with the traditional techniques used in the clinical setting