Reducing False-Positive Biopsies using Deep Neural Networks that Utilize both Local and Global Image Context of Screening Mammograms.

Breast cancer is the most common cancer in women, and hundreds of thousands of unnecessary biopsies are done around the world at a tremendous cost. It is crucial to reduce the rate of biopsies that turn out to be benign tissue. In this study, we build deep neural networks (DNNs) to classify biopsied lesions as being either malignant or benign, with the goal of using these networks as second readers serving radiologists to further reduce the number of false-positive findings. We enhance the performance of DNNs that are trained to learn from small image patches by integrating global context provided in the form of saliency maps learned from the entire image into their reasoning, similar to how radiologists consider global context when evaluating areas of interest. Our experiments are conducted on a dataset of 229,426 screening mammography examinations from 141,473 patients. We achieve an AUC of 0.8 on a test set consisting of 464 benign and 136 malignant lesions.

Fatty acid composition in mammary adipose tissue measured by Gradient-echo Spectroscopic MRI and its association with breast cancers.

PURPOSE: To assess the association of fatty acid levels in mammary adipose tissue of postmenopausal women with the presence of breast cancer using the Gradient-echo Spectroscopic Imaging (GSI). MATERIALS AND METHODS: Unilateral GSI was performed at 3 T in 61 postmenopausal women undergoing breast MRI exams. The study included 19 women with breast cancer, 23 women with benign/high risk lesions, and 19 women with a history of cancer. Voxel-wise spectral analysis of fatty acids was conducted to measure relative portions of monounsaturated (MUFA), polyunsaturated (PUFA), and saturated fatty acids (SFA) in each voxel. The voxels within mammary adipose tissue were automatically selected and their median fatty acid fractions were used for quantitative analysis. Statistical analyses were performed using χ2 test, one-way analysis of variance (ANOVA) with Tukey-Kramer multiple comparison tests, and linear regression. RESULTS: Postmenopausal women with malignancies had significantly higher SFA (0.336 ± 0.038) in mammary adipose tissue compared to those with benign disease (0.283 ± 0.046, p = 0.0008) and to those with a history of breast cancer (0.287 ± 0.050, p = 0.0038). Postmenopausal women with malignant lesions had significantly lower MUFA (0.352 ± 0.041) compared to those with benign disease (0.401 ± 0.043, p = 0.0032) and with history of breast cancer (0.388 ± 0.055, p = 0.0484). The history of cancer group had a significant correlation (r = 0.60, p = 0.006) between SFA and BMI, and the cancer group had a significant correlation (r = 0.57, p = 0.010) between PUFA and BMI. CONCLUSIONS: Fatty acid composition of mammary adipose tissue, particularly higher SFA and lower MUFA, may be associated with breast cancer. The GSI method utilizes an automated voxel-based analysis to measure fatty acid composition, and may be used to assess the role of mammary adipose tissue in cancer development and progress.