Search

Aims and objectives To investigate the value of advanced multiparametric MR imaging biomarker analysis based on radiomic features and machine learning classification, in the non-invasive evaluation of tumor heterogeneity towards the differentiation of Low Grade vs. High Grade Gliomas. Methods and materials Forty histologically confirmed glioma patients (20 LGG and 20 HGG) who underwent a standard 3T-MRI tumor protocol with conventional (T1 pre/post-contrast, T2-FSE, T2-FLAIR) and advanced techniques (Diffusion Tensor and Perfusion Imaging, 1H-MR Spectroscopy), were included. A semi-automated segmentation technique, based on T1W-C and DTI, was used for tumor core delineation in all available parametric maps. 3D Texture analysis considered 12 Histogram, 11 Co-Occurrence Matrix (GLCM) and 5 Run Length Matrix (GLRLM) features, derived from p, q, MD, FA, T1W-C, T2W-FSE, T2W-FLAIR and raw DSCE data. Along with 1H-MRS metabolic ratios and mean rCBV values, a total of 581 attributes for each subject were obtained. A Support Vector Machine – Recursive Feature Elimination (SVM-RFE) algorithm and SVM classifier were utilized for feature selection and classification, respectively. Results Three different SVM classifiers were evaluated with consecutively SVM-RFE feature subsets. Linear SMO classifier demonstrated the highest performance for determining the optimal feature subset. Finally, 21 SVM-RFE top-ranked features were adopted, for training and testing the SMO classifier with leave-one-out cross-validation, achieving 95.5% Accuracy, 95% Sensitivity, 96% Specificity and 95.5% Area Under ROC Curve. Conclusion Results demonstrate that quantitative analysis of phenotypic characteristics, based on advanced multiparametric MR neuroimaging data and texture features, utilizing state-of-the-art radiomic analysis methods, can significantly contribute to the pre-treatment glioma grade differentiation.

Introduction Diffusion Weighted Imaging (DWI) has demonstrated increased potential in identifying imaging-based biomarkers for cancer diagnosis, prognosis and monitoring response to therapy. Purpose This study investigates the feasibility of apparent diffusion coefficient (ADC) histogram statistics in identifying highly aggressive breast carcinomas, by investigating correlation to established prognostic indices. Materials and methods The dataset consists of 43 histologically verified invasive ductal carcinomas (IDC) of patients undergoing breast DW-MRI at 3 T (b-values 0, 900 s/mm2). ADC maps were generated for a slice representative of lesion largest diameter. An expert radiologist delineated lesion contour on ADC map, defining the lesion region of interest to be subjected to histogram analysis. Feature extraction considered mean, standard deviation, skewness, kurtosis, entropy, maximum, minimum and range of ADC. The ability of ADC histogram features in identifying tumor grade, estrogen receptor (ER) and progesterone receptor (PR) status was investigated. Results Median value of feature ADC mean was 1.034 × 10−3 mm2/s, 0.869 × 10−3 mm2/s and 0.971 × 10−3 mm2/s in grade I, grade II and grade III lesions, respectively. Median value of ADC kurtosis was 0.176, 1.165 and 1.289 in grade I, grade II and grade III lesions, respectively. ER positive lesions demonstrated increased median value of ADC entropy (5.828), reflecting increased ADC heterogeneity, as compared to ER negative ones (5.480). Median value of ADC entropy was 5.842 and 5.486 in PR positive and PR negative lesions, respectively. Conclusion ADC histogram analysis may contribute in breast cancer prognosis by identifying high grade tumors and predicting ER and PR status.

Introduction Conventional breast MRI has shown high diagnostic sensitivities for the detection of breast cancer, whereas relatively low specificities have been reported, resulting in many unnecessary biopsies of benign lesions. Purpose The main purpose of this study was to investigate whether diffusion measures of anisotropy (DWI & DTI) can improve the discrimination between benign and malignant lesions. Moreover, using detailed DTI microstructure analysis, we sought to investigate the contribution of the architecture of breast fibroglandular tissue in the differential diagnosis of malignancy. Methods and materials The study included 58 women with a total of 86 breast lesions. DTI and DWI were performed complementary to dynamic contrast MRI. Apparent diffusion coefficient (ADC), mean diffusivity (MD) and fractional anisotropy (FA) were measured for lesions and contralateral breast parenchyma of each patient. All values were compared between malignant, benign lesions and the normal parenchyma by univariate and multivariate analyses. Results The FA values showed high variation and hence moderate statistical significance. Nevertheless the analysis of DTI vector maps and parametric maps, revealed significantly lower values of the tumor’s orthogonal diffusion coefficients λ1, λ2, λ3 comparing to normal breast and benign tissue. Despite the significant overlap the FA values of malignant tissue were significantly higher than normal and benign tissue (p = 0.002). ROC curve analysis was also performed for every discrimination factor studied. Conclusion ADC measures have accurate discrimination ability in all cases. Careful microstructural analysis revealed significantly lower values of the tumor’s prime diffusion coefficient λ1, and may be used as a potential indicator.