Multiparametric deep learning tissue signatures for a radiological biomarker of breast cancer: Preliminary results.

PURPOSE: Deep learning is emerging in radiology due to the increased computational capabilities available to reading rooms. These computational developments have the ability to mimic the radiologist and may allow for more accurate tissue characterization of normal and pathological lesion tissue to assist radiologists in defining different diseases. We introduce a novel tissue signature model based on tissue characteristics in breast tissue from multiparametric magnetic resonance imaging (mpMRI). The breast tissue signatures are used as inputs in a stacked sparse autoencoder (SSAE) multiparametric deep learning (MPDL) network for segmentation of breast mpMRI. METHODS: We constructed the MPDL network from SSAE with 5 layers with 10 nodes at each layer. A total cohort of 195 breast cancer subjects were used for training and testing of the MPDL network. The cohort consisted of a training dataset of 145 subjects and an independent validation set of 50 subjects. After segmentation, we used a combined SAE-support vector machine (SAE-SVM) learning method for classification. Dice similarity (DS) metrics were calculated between the segmented MPDL and dynamic contrast enhancement (DCE) MRI-defined lesions. Sensitivity, specificity, and area under the curve (AUC) metrics were used to classify benign from malignant lesions. RESULTS: The MPDL segmentation resulted in a high DS of 0.87 ± 0.05 for malignant lesions and 0.84 ± 0.07 for benign lesions. The MPDL had excellent sensitivity and specificity of 86% and 86% with positive predictive and negative predictive values of 92% and 73%, respectively, and an AUC of 0.90. CONCLUSIONS: Using a new tissue signature model as inputs into the MPDL algorithm, we have successfully validated MPDL in a large cohort of subjects and achieved results similar to radiologists.

Diffusion-weighted imaging improves the diagnostic accuracy of conventional 3.0-T breast MR imaging.

PURPOSE: To evaluate the incremental value of diffusion-weighted (DW) imaging and apparent diffusion coefficient (ADC) mapping in relation to conventional breast magnetic resonance (MR) imaging in the characterization of benign versus malignant breast lesions at 3.0 T. MATERIALS AND METHODS: This retrospective HIPAA-compliant study was approved by the institutional review board, with the requirement for informed patient consent waived. Of 550 consecutive patients who underwent bilateral breast MR imaging over a 10-month period, 93 women with 101 lesions met the following study inclusion criteria: They had undergone three-dimensional (3D) high-spatial-resolution T1-weighted contrast material-enhanced MR imaging, dynamic contrast-enhanced MR imaging, and DW imaging examinations at 3.0 T and either had received a pathologic analysis-proven diagnosis (96 lesions) or had lesion stability confirmed at more than 2 years of follow-up (five lesions). DW images were acquired with b values of 0 and 600 sec/mm(2). Regions of interest were drawn on ADC maps of breast lesions and normal glandular tissue. Morphologic features (margin, enhancement pattern), dynamic contrast-enhanced MR results (semiquantitative kinetic curve data), absolute ADCs, and glandular tissue-normalized ADCs were included in multivariate models to predict a diagnosis of benign versus malignant lesion. RESULTS: Forty-one (44%) of the 93 patients were premenopausal, and 52 (56%) were postmenopausal. Thirty-three (32.7%) of the 101 lesions were benign, and 68 (67.3%) were malignant. Normalized ADCs were significantly different between the benign (mean ADC, 1.1 x 10(-3) mm(2)/sec +/- 0.4 [standard deviation]) and malignant (mean ADC, 0.55 x 10(-3) mm(2)/sec +/- 0.16) lesions (P < .001). Adding normalized ADCs to the 3D T1-weighted and dynamic contrast-enhanced MR data improved the diagnostic performance of MR imaging: The area under the receiver operating characteristic curve improved from 0.89 to 0.98, and the false-positive rate decreased from 36% (nine of 25 lesions) to 24% (six of 25 lesions). CONCLUSION: DW imaging with glandular tissue-normalized ADC assessment improves the characterization of breast lesions beyond the characterization achieved with conventional 3D T1-weighted and dynamic contrast-enhanced MR imaging at 3.0 T.