Optimización de modelos gráficos de reconstrucción de estructuras biológicas mediante algoritmos evolutivos computacionales. Aplicación práctica en oftalmología clínica

Laburpena

Currently, it is common in the field of Clinical Ophthalmology to use various diagnostic instruments capable of providing morpho-geometric information about the corneal structure. This information allows for the characterization of the cornea and the detection of various pathologies, such as keratoconus, a degenerative disease that is currently among the most studied in the anterior segment of the human eye, due to its significant impact on the visual quality of the patient. This thesis addresses the study of a new method for reconstructing the corneal Surface using an evolutionary computational algorithm based on surface topographic data, and its subsequent application for the detection and classification of patients with keratoconus. To do this, an in-depth study has been conducted at three clearly distinguished levels: the analysis of corneal morphology, focusing on the study of the anatomy of the cornea in both healthy and pathological states, and the different morpho-geometric indices that characterize it, as well as the associated instrumentation; secondly, an analysis of the knowledge base of evolutionary computational algorithms and their application to geometric surface reconstruction; and finally, the study associated with computer-aided design (CAD) representation of geometric surfaces obtained from topographic data reconstruction. The methodology used in the thesis proposes the use of a genetic algorithm (GA) in the modal reconstruction of the corneal surface from topographic data of 270 corneal surfaces. The data is obtained from a cohort study of corneal topography involving 102 patients divided into a group of healthy patients and three pathological groups. The employed genetic algorithm is of the non-dominated multivariable type (MV-NDAG). The modal function of the obtained surface has been represented using CAD tools and digital lithography, in order to facilitate the interpretation of the results. The results obtained from the reconstructions performed have been analyzed using statistical tools to assess the goodness of the reconstruction and its potential to characterize the pathology by obtaining corneal morpho-geometric indices that could be used in the diagnosis and subsequent classification of keratoconus. The proposed method has been validated against other reconstruction methods based on the mean square error (MSE). Finally, a valorization of the obtained results has been carried out, comparing them with similar studies conducted in the field of Ophthalmology, as well as with the results of the application of the method in similar applications of biological structure reconstruction. This study allows for the establishment of a new method for modal reconstruction of the corneal surface from topographic data. Unlike other modal methods, this reconstruction allows for a valid solution in incomplete datasets, even improving its accuracy. Furthermore, the values of the obtained morpho-geometric indices allow for an accurate diagnosis and classification of keratoconus, making it an additional tool in the analysis and diagnosis of corneal ectasias.