Computer aided staging of lymphoma patients with FDG PET/CT imaging based on textural information

We have designed a computer aided diagnosis (CADx) system to assess the presence of cancer in FDG PET/CT exams of lymphoma patients. Detection performances of the random decision forest (RDF) and support vector machine (SVM) classifiers were assessed based on a feature set including 115 PET and CT first order and textural parameters. An original feature selection method based on combining different filter methods was proposed. The evaluation database consisted of 156 lymphomatous (M for malignant), 158 physiologic (N for normal) and 32 inflammatory (NS for normal suspicious) regions of interest. An optimization study was performed for each classifier separately to select the best combination of parameters considering the two problems of discriminating the {M} and {NS+N} classes and the {M} and {NS} classes. Promising classification performance was achieved by the SVM combined with the 12 most discriminant features with AUC values of 0.97 and 0.91 for the first and second problem respectively.