RESUMO
BACKGROUND: A fair amount of microcalcifications sent for biopsy are false positives. The study investigates whether quantitative radiomic features extracted from digital breast tomosynthesis (DBT) can be an additional and useful tool to discriminate between benign and malignant BI-RADS category 4 microcalcification. METHODS: This retrospective study included 252 female patients with BI-RADS category 4 microcalcifications. The patients were divided into two groups according to micro-histopathology: 126 patients with benign lesions and 126 patients with certain or possible malignancies. A total of 91 radiomic features were extracted for each patient, and the 12 most representative features were selected by using the agglomerative hierarchical clustering method. The binary classification task of the two groups was carried out by using four different machine-learning algorithms (i.e., linear support vector machine (SVM), radial basis function (RBF) SVM, logistic regression (LR), and random forest (RF)). Accuracy, sensitivity, sensibility, and the area under the curve (AUC) were calculated for each of them. RESULTS: The best performance was achieved using the RF classifier (AUC = 0.59, 95% confidence interval 0.57-0.60; sensitivity = 0.56, 95% CI 0.54-0.58; specificity = 0.61, 95% CI 0.59-0.63; accuracy = 0.58, 95% CI 0.57-0.59). CONCLUSIONS: DBT-based radiomic analysis seems to have only limited potential in discriminating benign from malignant microcalcifications.
RESUMO
Few reports in literature describe triple negative breast cancer (TNBC) imaging findings. Aim of the study is to determine MR-features of TNBC compared to receptor positive cancer (nTNBC). From May 2014 to May 2015, we retrospectively enrolled 31 consecutive patients with histological diagnosis of TNBC and a control group of 31 consecutive nTNBC observed in the same period, out of 602 cancer, diagnosed in our department in the same year. Histopathological analysis and MR-features of TNBC (31 patients) were compared to nTNBC (31 patients). MR-features included dimension, fibroglandular tissue (FGT), background parenchimal enhancement (BPE), mass shape, margins, presence of rim, intratumoral signal intensity in T2w, uni-multifocality, kinetic curves. All patients were examined with MR 1,5T (Magnetom Simphony Tim, Siemens Healthcare) performing T2w fat-sat and contrast enhanced high temporal and spatial resolution T1w before and after injection of Gadolinium. 62 staging MR were reviewed. Median age was 50 (30-78ys) with a standard deviation of 10,9. TNBC showed 3 MR features in concordance with current literature: rim enhancement, hyperintensity in T2 sequence and unifocality. Rim enhancement was shown in 67.7% of TNBC (21/31) and 29% of nTNBC (9/31). Higher T2w values were shown in 83.9% of TNBC (26/31) and 58.1% of nTNBC (18/31). Cancer was multifocal in 7/31 (22.6%) of TNBC and 19/31 (61.3%) nTNBC. No correlation was found for dimension (p=0.12), FGT (p=0.959), BPE (p=0.596), homogeneity of enhancement (p=0.43), margins (p=0.671) and kinetic (p=0.37). Multivariate analysis demonstrated that rim enhancement and unifocality correlated independently with TNBC group. Area under ROC curve of our model is 0.835. Furthermore, we evaluated the clinical outcome of all 31 TNBC patients in a follow-up time ranging from 24months to 36months separating them in a free-survival group (23 women) and a recurrence group (8 women with local recurrence or distant metastasis): only kinetic curves resulted to be significantly higher in recurrence group (p=0.042).
