Research

CAD LUNGS

The aim of this project is to create and validate a prototype CAD system for the diagnosis of lung cancer, which will be able to process CT thorax scans, identify and indicate a certain number of suspect areas, and assist the radiologist in interpreting the results and compiling the reports. The system is made up of a hardware component consisting of a high-performance computer, and a software component able to receive the test results, process them, generate a list of nodule candidates, then  present everything to the radiologist in an efficient and easily interpreted form.

• The project is pursuing the following objectives:
• Identification of pulmonary nodules
• Evaluation of nodule dimensions
• Comparison of nodules with follow-up scans

Cancer biomarkers

Identification of new tumor biomarkers in collections of biomedical images.
Clinical diagnoses and the screening programs for oncological pathologies gather enormous quantities of radiodiagnostic images, which require a committed effort on the part of the radiologist in order to spot the characteristic signs of cancer. Present computer-assisted diagnosis systems that alleviate the work of the radiologist are based on prior knowledge of the biomarkers concerned. However, much of the information contained in the images is unused because its potential importance remains unknown. Making use of this information resource could contribute significantly to cancer prevention.
The aim of this project is to develop methods and instruments for identifying new tumor biomarkers, using a statistical approach which does not work from a priori hypotheses, but examines the correlation between parameters extracted from clinical and anamnestic data and images.
The statistical approach to the researching of new biomarkers and the extraction of these biomarkers by image processing and analysis methods introduces a significant innovation in clinical research methodology

Virtual gastroscopy

Virtual endoscopy is a new and developing technology based on the analysis of 2-dimensional (2D) and 3-dimensional (3D) images. Virtual endoscopy is a new method that could deliver effective diagnoses when used alongside the other diagnoses already obtained by standard endoscopy or video endoscopy. The rapid spread of CT technology (Computed Tomography) allows the rapid and accurate acquisition of images of the healthy and the pathological stomach, including the state of lymph node stations and possible secondary localisations in remote abdominal areas which cannot be evaluated by endoscopy.
The research project entitled ?Virtual Gatroscopy? aims to verify the development and application of a CAD system (Computer Aided Detection) for the 3D representation and automatic identification of gastric lesions through the analysis of CT images. The automatic identification of areas with suspected pathologies could allow radiologists to focus their attention during image analysis on areas where there is possible deterioration of the hollow organ, showing these in heightened detail and with greater sensitivity, and creating a new clinical method aimed at improving sensitivity and patient compliance