A classification approach to prostate cancer localization in 3T Multi-Parametric MRI

Multiparametric-magnetic resonance imaging (mp-MRI) has demonstrated, in many studies, its potential in prostate cancer detection and analysis. We propose a supervised classification approach based on mp-MRI data base of 20 patients, in order to localize prostate cancer and to achieve a cartographic representation of the prostate voxels based on classification results. Proposed method provides a computer aided detection (CAD) software for prostatic cancer. For that, we have extracted varied features providing functional, anatomical and metabolic information helping the classifier to distinguish between three different classes (''Healthy'', ``Benign'' and ``Pathologic''). We started by evaluating Support Vector Machine (SVM) ability to separate healthy and pathologic voxels. We obtained an error rate of 0.99%, specificity 99.25% and sensitivity 98.85%. Then, by introducing ``Benign'' voxels, SVM gave an error rate of 26% using MRSI, Diffusion-Weighted MRI and Dynamic Contrast-Enhanced MRI. Next, we evaluated Random Forest performances which gave error rate of 24.60% when separating three different classes using MRSI, T2-MRI, Diffusion-Weighted MRI and Dynamic Contrast-Enhanced MRI. Finally, we presented color-coded maps based on classification results.