While breast cancer screening recommendations vary by agency, all agencies recommend mammographic screening with some frequency over some portion of a woman’s lifetime. Temporal evaluation of these images may inform personalized risk of breast cancer. However, due to the highly deformable nature of breast tissue, the positioning of breast tissue may vary widely between exams. Therefore, registration of physical regions in the breast over time points is a critical first step in computerized analysis of changes in breast parenchyma over time. While a postregistration image is altered and therefore not appropriate for radiomic texture analysis, the registration process produces a mapping of points which may aid in aligning similar image regions across multiple time points. In this study, a total of 633 mammograms from 87 patients were retrospectively collected. These images were sorted into 1144 temporal pairs, where each combination of images of a given women of a given laterality was used to form a temporal pair. B-splines registration and multi-resolution registration were performed on each mammogram pair. While the B-splines took an average of 552.8 CPU seconds per registration, multi-resolution registration took only an average of 346.2 CPU seconds per registration. Multi-resolution registration had a 15% lower mean square error, which was significantly different than that of B-splines (p<0.001). While previous work aimed to allow radiologists to visually evaluate the registered images, this study identifies corresponding points on images for use in assessing interval change for risk assessment and early detection of cancer through deep learning and radiomics.
Given the increased need for consistent quantitative image analysis, variations in radiomics feature calculations due to differences in radiomics software were investigated. Two in-house radiomics packages and two freely available radiomics packages, MaZda and IBEX, were utilized. Forty 256 × 256-pixel regions of interest (ROIs) from 40 digital mammograms were studied along with 39 manually delineated ROIs from the head and neck (HN) computed tomography (CT) scans of 39 patients. Each package was used to calculate first-order histogram and second-order gray-level co-occurrence matrix (GLCM) features. Friedman tests determined differences in feature values across packages, whereas intraclass-correlation coefficients (ICC) quantified agreement. All first-order features computed from both mammography and HN cases (except skewness in mammography) showed significant differences across all packages due to systematic biases introduced by each package; however, based on ICC values, all but one first-order feature calculated on mammography ROIs and all but two first-order features calculated on HN CT ROIs showed excellent agreement, indicating the observed differences were small relative to the feature values but the bias was systematic. All second-order features computed from the two databases both differed significantly and showed poor agreement among packages, due largely to discrepancies in package-specific default GLCM parameters. Additional differences in radiomics features were traced to variations in image preprocessing, algorithm implementation, and naming conventions. Large variations in features among software packages indicate that increased efforts to standardize radiomics processes must be conducted.
We investigated the additive role of breast parenchyma stroma in the computer-aided diagnosis (CADx) of tumors on full-field digital mammograms (FFDM) by combining images of the tumor and contralateral normal parenchyma information via deep learning. The study included 182 breast lesions in which 106 were malignant and 76 were benign. All FFDM images were acquired using a GE 2000D Senographe system and retrospectively collected under an Institution Review Board (IRB) approved, Health Insurance Portability and Accountability Act (HIPAA) compliant protocol. Convolutional neutral networks (CNNs) with transfer learning were used to extract image-based characteristics of lesions and of parenchymal patterns (on the contralateral breast) directly from the FFDM images. Classification performance was evaluated and compared between analysis of only tumors and that of combined tumor and parenchymal patterns in the task of distinguishing between malignant and benign cases with the area under the Receiver Operating Characteristic (ROC) curve (AUC) used as the figure of merit. Using only lesion image data, the transfer learning method yielded an AUC value of 0.871 (SE=0.025) and using combined information from both lesion and parenchyma analyses, an AUC value of 0.911 (SE=0.021) was observed. This improvement was statistically significant (p-value=0.0362). Thus, we conclude that using CNNs with transfer learning to combine extracted image information of both tumor and parenchyma may improve breast cancer diagnosis.
KEYWORDS: Digital breast tomosynthesis, Breast, Convolutional neural networks, Computer aided diagnosis and therapy, Feature extraction, Breast cancer, Digital mammography, Image classification, Databases
With growing adoption of digital breast tomosynthesis (DBT) in breast cancer screening protocols, it is important to compare the performance of computer-aided diagnosis (CAD) in the diagnosis of breast lesions on DBT images compared to conventional full-field digital mammography (FFDM). In this study, we retrospectively collected FFDM and DBT images of 78 lesions from 76 patients, each containing lesions that were biopsy-proven as either malignant or benign. A square region of interest (ROI) was placed to fully cover the lesion on each FFDM, DBT synthesized 2D images, and DBT key slice images in the cranial-caudal (CC) and mediolateral-oblique (MLO) views. Features were extracted on each ROI using a pre-trained convolutional neural network (CNN). These features were then input to a support vector machine (SVM) classifier, and area under the ROC curve (AUC) was used as the figure of merit. We found that in both the CC view and MLO view, the synthesized 2D image performed best (AUC = 0.814, AUC = 0.881 respectively) in the task of lesion characterization. Small database size was a key limitation in this study, and could lead to overfitting in the application of the SVM classifier. In future work, we plan to expand this dataset and to explore more robust deep learning methodology such as fine-tuning.
We evaluated the potential of deep learning in the assessment of breast cancer risk using convolutional neural networks (CNNs) fine-tuned on full-field digital mammographic (FFDM) images. This study included 456 clinical FFDM cases from two high-risk datasets: BRCA1/2 gene-mutation carriers (53 cases) and unilateral cancer patients (75 cases), and a low-risk dataset as the control group (328 cases). All FFDM images (12-bit quantization and 100 micron pixel) were acquired with a GE Senographe 2000D system and were retrospectively collected under an IRB-approved, HIPAA-compliant protocol. Regions of interest of 256x256 pixels were selected from the central breast region behind the nipple in the craniocaudal projection. VGG19 pre-trained on the ImageNet dataset was used to classify the images either as high-risk or as low-risk subjects. The last fully-connected layer of pre-trained VGG19 was fine-tuned on FFDM images for breast cancer risk assessment. Performance was evaluated using the area under the receiver operating characteristic (ROC) curve (AUC) in the task of distinguishing between high-risk and low-risk subjects. AUC values of 0.84 (SE=0.05) and 0.72 (SE=0.06) were obtained in the task of distinguishing between the BRCA1/2 gene-mutation carriers and low-risk women and between unilateral cancer patients and low-risk women, respectively. Deep learning with CNNs appears to be able to extract parenchymal characteristics directly from FFDMs which are relevant to the task of distinguishing between cancer risk populations, and therefore has potential to aid clinicians in assessing mammographic parenchymal patterns for cancer risk assessment.
Open-source texture analysis software allows for the advancement of radiomics research. Variations in texture features, however, result from discrepancies in algorithm implementation. Anatomically matched regions of interest (ROIs) that captured normal breast parenchyma were placed in the magnetic resonance images (MRI) of 20 patients at two time points. Six first-order features and six gray-level co-occurrence matrix (GLCM) features were calculated for each ROI using four texture analysis packages. Features were extracted using package-specific default GLCM parameters and using GLCM parameters modified to yield the greatest consistency among packages. Relative change in the value of each feature between time points was calculated for each ROI. Distributions of relative feature value differences were compared across packages. Absolute agreement among feature values was quantified by the intra-class correlation coefficient. Among first-order features, significant differences were found for max, range, and mean, and only kurtosis showed poor agreement. All six second-order features showed significant differences using package-specific default GLCM parameters, and five second-order features showed poor agreement; with modified GLCM parameters, no significant differences among second-order features were found, and all second-order features showed poor agreement. While relative texture change discrepancies existed across packages, these differences were not significant when consistent parameters were used.
Extraction of high-dimensional quantitative data from medical images has become necessary in disease risk assessment,
diagnostics and prognostics. Radiomic workflows for mammography typically involve a single medical image for each
patient although medical images may exist for multiple imaging exams, especially in screening protocols. Our study
takes advantage of the availability of mammograms acquired over multiple years for the prediction of cancer onset. This
study included 841 images from 328 patients who developed subsequent mammographic abnormalities, which were
confirmed as either cancer (n=173) or non-cancer (n=155) through diagnostic core needle biopsy. Quantitative radiomic
analysis was conducted on antecedent FFDMs acquired a year or more prior to diagnostic biopsy. Analysis was limited
to the breast contralateral to that in which the abnormality arose. Novel metrics were used to identify robust radiomic
features. The most robust features were evaluated in the task of predicting future malignancies on a subset of 72 subjects
(23 cancer cases and 49 non-cancer controls) with mammograms over multiple years. Using linear discriminant analysis,
the robust radiomic features were merged into predictive signatures by: (i) using features from only the most recent
contralateral mammogram, (ii) change in feature values between mammograms, and (iii) ratio of feature values over
time, yielding AUCs of 0.57 (SE=0.07), 0.63 (SE=0.06), and 0.66 (SE=0.06), respectively. The AUCs for temporal
radiomics (ratio) statistically differed from chance, suggesting that changes in radiomics over time may be critical for
risk assessment. Overall, we found that our two-stage process of robustness assessment followed by performance
evaluation served well in our investigation on the role of temporal radiomics in risk assessment.
The robustness of radiomic texture analysis across different manufacturers of mammography imaging systems is investigated. We quantified feature robustness across mammography manufacturers using a dataset of 111 women who underwent consecutive screening mammography on both general electric and Hologic systems. In each mammogram, a square region of interest (ROI) directly behind the nipple was manually selected. Radiomic features describing parenchymal patterns were automatically extracted on each ROI. Feature comparisons were conducted between manufacturers (and breast densities) using newly developed robustness metrics descriptive of correlation, equivalence, and variability. By examining the distribution of these metric values, we propose the following selection criteria to guide feature evaluation in this dataset: (1) 0.80.5, and (4) p<0.05. Statistically significant correlation coefficients ranged from 0.13 to 0.68 in comparisons between the two mammographic systems tested. Features describing spatial patterns tended to exhibit high correlation coefficients, while intensity- and directionality-based features had comparatively poor correlation. Our proposed robustness metrics may be used to evaluate other datasets, for which different ranges of metric values may be appropriate.
KEYWORDS: Breast, Tissues, Image segmentation, Magnetic resonance imaging, Breast cancer, Medicine, Mammography, 3D image processing, 3D displays, Cancer, Fuzzy logic
Breast density is routinely assessed qualitatively in screening mammography. However, it is challenging to quantitatively determine a 3D density from a 2D image such as a mammogram. Furthermore, dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) is used more frequently in the screening of high-risk populations. The purpose of our study is to segment parenchyma and to quantitatively determine volumetric breast density on pre-contrast axial DCE-MRI images (i.e., non-contrast) using a semi-automated quantitative approach. In this study, we retroactively examined 3D DCE-MRI images taken for breast cancer screening of a high-risk population. We analyzed 66 cases with ages between 28 and 76 (mean 48.8, standard deviation 10.8). DCE-MRIs were obtained on a Philips 3.0 T scanner. Our semi-automated DCE-MRI algorithm includes: (a) segmentation of breast tissue from non-breast tissue using fuzzy cmeans clustering (b) separation of dense and fatty tissues using Otsu’s method, and (c) calculation of volumetric density as the ratio of dense voxels to total breast voxels. We examined the relationship between pre-contrast DCE-MRI density and clinical BI-RADS density obtained from radiology reports, and obtained a statistically significant correlation [Spearman ρ-value of 0.66 (p < 0.0001)]. Our method within precision medicine may be useful for monitoring high-risk populations.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.