Quantitative methods for computer aided decision support systems in confocal laser endomicroscopy imaging of the gastrointestinal tract

摘要

The mucosa of the gastrointestinal tract represents the mainudbarrier between the inner body and the external world. Audlayer of cells runs from the esophagus to the rectum, playinguda key role in preventing access to environmental hostileudfactors that could cause inflammation. Alterations in suchudmucosa are caused or can cause severe problems to patients,udamong others celiac disease, irritable bowel disease, Crohn’suddisease, ulcerative colitis and Barrett’s esophagus. The goldudstandard for evaluating such diseases requires biopsies to beudperformed on the patient, often following the random fourquadrantudprotocol, other than positive serology. Quantitativeudmethods for evaluating in-vivo these diseases, by exploitinguddistinctive image features that vary according to the grade ofudthe disease, would improve the way clinical examinations areudperformed. This could in the long run lead to virtual biopsiesudwith a single endoscopy examination. We propose a ComputerudAided Decision Support System for endoscopic examinationsudperformed using Confocal Laser Endomicroscopy forudceliac disease and irritable bowel syndrome that, exploitingudimage features extracted in an automatic way, can assist theudphysician in its diagnosis and help him in selecting and identifyingudthe areas that most require attention during an examination.udExploiting image features that are well-investigatedudin the literature, our tool outputs valuable information aboutudthe mucosa under examination with a friendly user interface.udWe hope with such solution to increase the attention towardsudthe need of quantitative methods in this medical field.