Machine Learning
Fast Gene Selection for Microarray Data Using SVM-Based Evaluation Criterion
BIBM '08 Proceedings of the 2008 IEEE International Conference on Bioinformatics and Biomedicine
ICCEA '10 Proceedings of the 2010 Second International Conference on Computer Engineering and Applications - Volume 02
| Hi-index | 0.00 |
In recent years, endovascular aneurysm repair (EVAR) has proved to be an effective technique for the treatment of abdominal aneurysm. However, complications as leaks inside the aneurysm sac (endoleaks) can appear, causing pressure elevation and increasing the danger of rupture consequently. Computed tomographic angiography (CTA) is the most commonly used examination for medical surveillance, but endoleaks can not always be detected by visual inspection on CTA scans. The aim of this work was to evaluate the capability of texture features obtained from CT images, to discriminate evolutions after EVAR. Regions of interest (ROIs) from patients with different post-EVAR evolution were extracted by experienced radiologists. Three different techniques were applied to each ROI to obtain texture parameters, namely the gray level co-occurrence matrix (GLCM) , the gray level run length matrix (GLRLM) and the gray level difference method (GLDM). In order to evaluate the discrimination ability of textures features, each set of features was applied as input to support vector machine (SVM) classifier. The performance of the classifier was evaluated using 10-fold cross validation with the entire dataset. The average of accuracy, sensitivity, specificity, receiving operating curves (ROC) and area under the ROC curves (AUC) were calculated for the classification performances of each texture-analysis method. The study showed that the textural features could help radiologists in the classification of abdominal aneurysm evolution after EVAR.