Mask-Guided Convolutional Neural Network for Breast Tumor Prognostic Outcome Prediction on 3D DCE-MR Images.

In this proof-of-concept work, we have developed a 3D-CNN architecture that is guided by the tumor mask for classifying several patient-outcomes in breast cancer from the respective 3D dynamic contrast-enhanced MRI (DCE-MRI) images. The tumor masks on DCE-MRI images were generated using pre- and post-contrast images and validated by experienced radiologists. We show that our proposed mask-guided classification has a higher accuracy than that from either the full image without tumor masks (including background) or the masked voxels only. We have used two patient outcomes for this study: (1) recurrence of cancer after 5 years of imaging and (2) HER2 status, for comparing accuracies of different models. By looking at the activation maps, we conclude that an image-based prediction model using 3D-CNN could be improved by even a conservatively generated mask, rather than overly trusting an unguided, blind 3D-CNN. A blind CNN may classify accurately enough, while its attention may really be focused on a remote region within 3D images. On the other hand, only using a conservatively segmented region may not be as good for classification as using full images but forcing the model's attention toward the known regions of interest.

Triple-negative and non-triple-negative invasive breast cancer: association between MR and fluorine 18 fluorodeoxyglucose PET imaging.

PURPOSE: To assess the relationship between parameters measured on dynamic contrast material-enhanced (DCE) magnetic resonance (MR) imaging and fluorine 18 fluorodeoxyglucose (FDG) positron emission tomography (PET)/computed tomography (CT) in primary invasive breast cancer. MATERIALS AND METHODS: This HIPAA-compliant study was a retrospective review of medical records and therefore approved by the institutional review board without the requirement for informed consent. Patients with a diagnosis of invasive breast cancer from January 2005 through December 2009 who underwent both DCE MR imaging and FDG PET/CT before treatment initiation were retrospectively identified. Fractional volumes were measured for ranges of signal enhancement ratio (SER) values from DCE MR imaging data and compared with maximum standardized uptake values (SUVmax) from FDG PET/CT data. Linear regression analysis was performed to clarify the relationship between SER and SUVmax, adjusting for tumor size, pathologic grade, and receptor status. RESULTS: Analyzed were 117 invasive breast cancers in 117 patients. Overall, a higher percentage of high washout kinetics was positively associated with SUVmax (1.57% increase in SUVmax per 1% increase in high washout; P = .020), and a higher percentage of low plateau kinetics was negatively associated with SUVmax (1.19% decrease in SUVmax per 1% increase in low plateau; P = .003). These relationships were strongest among triple-negative (TN) tumors (4.34% increase in SUVmax per 1% increase in high washout and 2.65% decrease in SUVmax per 1% increase in low plateau; P = .018 and .004, respectively). CONCLUSION: In invasive breast carcinoma, there is a positive relationship between the percentage of high washout and SUVmax and a negative relationship between the percentage of low plateau and SUVmax. These results are stronger in TN tumors. SUPPLEMENTAL MATERIAL: http://radiology.rsna.org/lookup/suppl/doi:10.1148/radiol.13130058/-/DC1.