Mammographic Density Estimation Research Articles

Pectoral muscle removal in mammogram images: A novel approach for improved accuracy and efficiency

PurposeAccurate pectoral muscle removal is critical in mammographic breast density estimation and many other computer-aided algorithms. We propose a novel approach to remove pectoral muscles form mediolateral oblique (MLO) view mammograms and compare accuracy and computational efficiency with existing method (Libra).MethodsA pectoral muscle identification pipeline was developed. The image is first binarized to enhance contrast and then the Canny algorithm was applied for edge detection. Robust interpolation is used to smooth out the pectoral muscle region. Accuracy and computational speed of pectoral muscle identification was assessed using 951 women (1,902 MLO mammograms) from the Joanne Knight Breast Health Cohort at Washington University School of Medicine.ResultsOur proposed algorithm exhibits lower mean error of 12.22% in comparison to Libra’s estimated error of 20.44%. This 40% gain in accuracy was statistically significant (p < 0.001). The computational time for the proposed algorithm is 5.4 times faster when compared to Libra (5.1 s for proposed vs. 27.7 s for Libra per mammogram).ConclusionWe present a novel approach for pectoral muscle removal in mammogram images that demonstrates significant improvement in accuracy and efficiency compared to existing method. Our findings have important implications for the development of computer-aided systems and other automated tools in this field.

Abstract 2600: Deep-LIBRA: An artificial intelligence approach for fully-automated assessment of breast density in digital mammography

Abstract Accurate estimation of mammographic breast density could aid in augmenting breast cancer risk assessment for women undergoing breast screening with full-field digital mammography (FFDM). Breast density can be estimated from FFDM and is most commonly assessed in the clinic by visual grading into one of the four categories defined by the American College of Radiology BI-RADS. However, BI-RADS density assessment is highly subjective and does not provide a quantitative, continuous measure of percent density (PD), which would allow for more refined risk stratification and assessment of density changes. Here, we introduce Deep-LIBRA, an artificial intelligence (AI) tool for fully-automated assessment of breast PD from FFDM images. Two key modules form the core of Deep-LIBRA: 1) an implementation of a modified U-Net architecture for breast segmentation and 2) a radiomic machine learning module that performs PD estimation within the segmented breast region. To develop and validate Deep-LIBRA, raw (i.e., "For Processing") FFDM images (Selenia Dimensions, Hologic Inc.) acquired at two breast cancer screening practices were retrospectively analyzed. For the breast segmentation module, we used a total of 12,100 FFDM studies from 2,200 individual women and a 90%-10% split-sample training-validation approach, using the Dice coefficient to evaluate the accuracy of Deep-LIBRA versus ground-truth manual breast segmentation. For the PD estimation module we used a total of 3,304 FFDM images from 1,652 individual women; manual PD scores obtained with the widely used Cumulus software were used as the "gold standard" in a three-fold cross-validation setting to assess the accuracy of Deep-LIBRA in PD estimation. PD estimates from Deep-LIBRA were also compared with breast density estimates from the commercially available Volpara software. Breast segmentation had a Dice coefficient of 95.31% when compared to ground-truth manual breast segmentation in the validation set. Deep-LIBRA average differences from ground-truth PD scores in the three cross-validation folds were 4.91%, 4.65%, and 4.22%, while Volpara had corresponding average differences of 6.20%, 6.01%, and 5.94%. Deep-LIBRA PD scores were also significantly different from Volpara PD (t-test p-value &amp;lt; 0.001 in all three folds). Preliminary evaluation results show that Deep-LIBRA is a promising AI approach for accurately assessing PD from FFDM images. Citation Format: Omid Haji Maghsoudi, Scott Christopher, Aimilia Gastounioti, Lauren Pantalone, Fang-Fang Wu, Eric A. Cohen, Winham Stacey, Emily F. Conant, Celine Vachon, Despina Kontos. Deep-LIBRA: An artificial intelligence approach for fully-automated assessment of breast density in digital mammography [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 2600.

Automated mammographic breast density estimation using a fully convolutional network.

The purpose of this study was to develop a fully automated algorithm for mammographic breast density estimation using deep learning. Our algorithm used a fully convolutional network, which is a deep learning framework for image segmentation, to segment both the breast and the dense fibroglandular areas on mammographic images. Using the segmented breast and dense areas, our algorithm computed the breast percent density (PD), which is the faction of dense area in a breast. Our dataset included full-field digital screening mammograms of 604 women, which included 1208 mediolateral oblique (MLO) and 1208 craniocaudal (CC) views. We allocated 455, 58, and 91 of 604 women and their exams into training, testing, and validation datasets, respectively. We established ground truth for the breast and the dense fibroglandular areas via manual segmentation and segmentation using a simple thresholding based on BI-RADS density assessments by radiologists, respectively. Using the mammograms and ground truth, we fine-tuned a pretrained deep learning network to train the network to segment both the breast and the fibroglandular areas. Using the validation dataset, we evaluated the performance of the proposed algorithm against radiologists' BI-RADS density assessments. Specifically, we conducted a correlation analysis between a BI-RADS density assessment of a given breast and its corresponding PD estimate by the proposed algorithm. In addition, we evaluated our algorithm in terms of its ability to classify the BI-RADS density using PD estimates, and its ability to provide consistent PD estimates for the left and the right breast and the MLO and CC views of the same women. To show the effectiveness of our algorithm, we compared the performance of our algorithm against a state of the art algorithm, laboratory for individualized breast radiodensity assessment (LIBRA). The PD estimated by our algorithm correlated well with BI-RADS density ratings by radiologists. Pearson's rho values of our algorithm for CC view, MLO view, and CC-MLO-averaged were 0.81, 0.79, and 0.85, respectively, while those of LIBRA were 0.58, 0.71, and 0.69, respectively. For CC view and CC-MLO averaged cases, the difference in rho values between the proposed algorithm and LIBRA showed statistical significance (P<0.006). In addition, our algorithm provided reliable PD estimates for the left and the right breast (Pearson's ρ>0.87) and for the MLO and CC views (Pearson's ρ=0.76). However, LIBRA showed a lower Pearson's rho value (0.66) for both the left and right breasts for the CC view. In addition, our algorithm showed an excellent ability to separate each sub BI-RADS breast density class (statistically significant, p-values = 0.0001 or less); only one comparison pair, density 1 and density 2 in the CC view, was not statistically significant (P=0.54). However, LIBRA failed to separate breasts in density 1 and 2 for both the CC and MLO views (P>0.64). We have developed a new deep learning based algorithm for breast density segmentation and estimation. We showed that the proposed algorithm correlated well with BI-RADS density assessments by radiologists and outperformed an existing state of the art algorithm.

Mammographic Ellipse Modelling for Risk Estimation

It has been shown that breast density and parenchymal patterns are significant indicators in mammographic risk assessment. In addition, studies have shown that the sensitivity of computer aided tools decreases significantly with increase in breast density. As such, mammographic density estimation and classification plays an important role in CAD systems. In this paper, we present the classification of mammographic images according to breast parenchymal structures through a multi-scale ellipse blob detection technique. Our classification is based on classifying the mammographic images of the MIAS dataset into high/low risk mammograms based on features extracted from a blob detection technique which is based on breast tissue structure. In addition, it evaluates the relation between the BIRADS classes and low/high risk mammograms. Results demonstrate the probability of estimating breast density using computer vision techniques to improve classification of mammographic images as low/high risk.

Mammographic Density Estimation with Automated Volumetric Breast Density Measurement

ObjectiveTo compare automated volumetric breast density measurement (VBDM) with radiologists' evaluations based on the Breast Imaging Reporting and Data System (BI-RADS), and to identify the factors associated with technical failure of VBDM.Materials and MethodsIn this study, 1129 women aged 19-82 years who underwent mammography from December 2011 to January 2012 were included. Breast density evaluations by radiologists based on BI-RADS and by VBDM (Volpara Version 1.5.1) were compared. The agreement in interpreting breast density between radiologists and VBDM was determined based on four density grades (D1, D2, D3, and D4) and a binary classification of fatty (D1-2) vs. dense (D3-4) breast using kappa statistics. The association between technical failure of VBDM and patient age, total breast volume, fibroglandular tissue volume, history of partial mastectomy, the frequency of mass > 3 cm, and breast density was analyzed.ResultsThe agreement between breast density evaluations by radiologists and VBDM was fair (k value = 0.26) when the four density grades (D1/D2/D3/D4) were used and moderate (k value = 0.47) for the binary classification (D1-2/D3-4). Twenty-seven women (2.4%) showed failure of VBDM. Small total breast volume, history of partial mastectomy, and high breast density were significantly associated with technical failure of VBDM (p = 0.001 to 0.015).ConclusionThere is fair or moderate agreement in breast density evaluation between radiologists and VBDM. Technical failure of VBDM may be related to small total breast volume, a history of partial mastectomy, and high breast density.

Breast cancer risk analysis based on a novel segmentation framework for digital mammograms.

The radiographic appearance of breast tissue has been established as a strong risk factor for breast cancer. Here we present a complete machine learning framework for automatic estimation of mammographic density (MD) and robust feature extraction for breast cancer risk analysis. Our framework is able to simultaneously classify the breast region, fatty tissue, pectoral muscle, glandular tissue and nipple region. Integral to our method is the extraction of measures of breast density (as the fraction of the breast area occupied by glandular tissue) and mammographic pattern. A novel aspect of the segmentation framework is that a probability map associated with the label mask is provided, which indicates the level of confidence of each pixel being classified as the current label. The Pearson correlation coefficient between the estimated MD value and the ground truth is 0.8012 (p-value < 0.0001). We demonstrate the capability of our methods to discriminate between women with and without cancer by analyzing the contralateral mammograms of 50 women with unilateral breast cancer, and 50 controls. Using MD we obtained an area under the ROC curve (AUC) of 0.61; however our texture-based measure of mammographic pattern significantly outperforms the MD discrimination with an AUC of 0.70.

Mammographic density estimation: one-to-one comparison between digital mammography and digital breast tomosynthesis using a fully automated software and magnetic resonance imaging calibration

Poster: "ECR 2013 / C-1055 / Mammographic density estimation: one-to-one comparison between digital mammography and digital breast tomosynthesis using a fully automated software and magnetic resonance imaging calibration" by: "A. Tagliafico, D. Astengo, F. Monetti, S. Airaldi, G. Tagliafico, B. Bingotti, M. Calabrese; Genoa/IT"

Mammographic density estimation: one-to-one comparison of digital mammography and digital breast tomosynthesis using fully automated software

To compare breast density on digital mammography and digital breast tomosynthesis using fully automated software. Following institutional approval and written informed consent from all participating women, both digital breast tomosynthesis (DBT) and full-field digital mammography (FFDM) were obtained. Breast percentage density was calculated with software on DBT and FFDM. Fifty consecutive patients (mean age, 51 years; range, 35-83 years) underwent both FFDM and DBT. Using a method based on the integral curve, breast density showed higher results on FFDM (68.1 ± 12.1 for FFDM and 51.9 ± 6.5 for DBT). FFDM overestimated breast density in 16.2% (P < 0.0001). Using a method based on maximum entropy thresholding, breast density showed higher results on FFDM (68.1 ± 12.1 for FFDM and 51.9 ± 6.5 for DBT). FFDM overestimated breast density in 11.4% (P < 0.0001). There was a good correlation among BI-RADS categories on a four-grade scale and the density evaluated with DBT and FFDM (r = 0.54, P < 0.01 and r = 0.44, P < 0.01). Breast density appeared to be significantly underestimated on digital breast tomosynthesis. Breast density is considered to be an independent risk factor for cancer Density can be assessed on full-field digital mammography and digital breast tomosynthesis Objective automated estimation of breast density eliminates subjectivity Automated estimation is more accurate than BI-RADS quantitative evaluation Breast density may be significantly underestimated on digital breast tomosynthesis.

Increased mammographic breast density in acromegaly: quantitative and qualitative assessment

Mammographic density is a strong independent risk factor for breast cancer, whose prevalence in acromegaly is still controversial. To compare breast density in premenopausal acromegalic patients and controls and to determine whether density correlated with disease duration, GH, and IGF1 levels. A prospective study involving 30 patients and 60 controls matched for age and body mass index. A quantitative computer-aided mammographic density estimation (MDEST) and a qualitative blind evaluation by two experienced radiologists using the breast imaging reporting and data system (BI-RADS) was performed. Totally, 60 (acromegaly) and 120 (controls) craniocaudal and mediolateral oblique mammograms were evaluated in both patients and controls. Breast density. Patients showed a significantly (P<0.01) increased mammographic breast density with both methods (MDEST: 0.33 ± 0.21% and BI-RADS category: 2.81 ± 0.78) in comparison with controls (MDEST: 0.26 ± 0.19% and BI-RADS category: 2.35 ± 0.61). The agreement between the two methods and inter-observer agreement between the two radiologists were excellent (k=0.63 and k=0.85). In patients grouped according to disease activity (17 controlled and 13 uncontrolled) and medical therapy (15 treated and 15 untreated), no differences were found. All these groups had significantly increased mammographic breast density compared with controls (P<0.01). A positive correlation was found between mammographic breast density, IGF1 values and disease duration (r=0.29 and r=0.39), whereas it was not found with GH (r=-0.02). Mammographic breast density in premenopausal acromegalic patients is significantly higher than controls and positively correlated with IGF1 and disease duration.

Sub-epithelial impedance: a new non-invasive method to measure mammographic density.

Abstract Abstract #4075 Background: Breast density is possibly the strongest risk factor for breast cancer after genetic predisposition. Estimation of mammographic density can be highly subjective. More accurate volumetric approaches to measure density using mammography are under development, but require x-ray exposure. Dense breast stroma has a lower electrical resistance, which becomes increased as the stroma is replaced with fat. We have previously reported that the impedance of breast stroma correlates with age, obesity and inversely with position in menstrual cycle. It is unknown whether mammographic density also correlates with stromal impedance. When a high frequency sine wave is used to interrogate the breast the impedance of the overlying skin and the breast epithelium are significantly reduced with most of the remaining impedance being due to the stromal tissue and is defined as the sub-epithelial impedance (Zsub).&amp;#x2028; Patients and Methods: With IRB and patient consent, electrical contact with ductal epithelium was established non-invasively, using a specially designed nipple sensor in 288 women. Measurements were made between the nipple sensor and skin surface electrodes placed in each of 4 quadrants of the breast. Zsub was measured at 60 KHz using a frequency response analyzer, and sine-wave correlation technique. Data were analyzed using a t-test, Mann-Whitney test, ANOVA, Pearson Moment Correlation, and stepwise regression as appropriate. In a subset of 112 women, in whom mammograms were available, a blinded review was performed to estimate the percent density averaged for two views of each breast.&amp;#x2028; Results: Zsub correlated with age of the patients (n = 288), (0.58 Correlation Coefficient (CC), p &amp;lt; 0.0001). Zsub was elevated in women who underwent a biopsy demonstrating breast cancer (n = 52) or proliferative breast disease (n = 27) with values of 211 ± 19Ω (median ± SEM), and 216 ± 18Ω respectively, compared with women with benign biopsies (n = 75) and values of 131 ± 15Ω (p &amp;lt; 0.01). In a subset of patients (n=112) Zsub was inversely correlated with mammographic density (-0.68 CC, p &amp;lt; 0.0001). Body mass index (BMI) increased with age (0.20 CC, p &amp;lt; 0.05), Zsub (0.48 CC, p &amp;lt; 0.00001) and inversely with mammographic density (-0.51 CC, p &amp;lt; 0.0001). Using stepwise regression, mammographic density was predicted by a linear combination of age, BMI and Zsub with a coefficient of 150.22 ± 11.98 (SE), Age -0.64 ± 0.20 (p &amp;lt; 0.002), BMI -1.43 ± 0.42 (p &amp;lt; 0.001) and Zsub -0.21 ± 0.04 (p &amp;lt; 0.001).&amp;#x2028; Discussion: Mammographic breast density decreases with increasing age and BMI because of replacement of stromal tissue with fat, which likely also affects the sub-epithelial electrical impedance of the breast. Breast density can be estimated non-invasively and objectively using measurements of Zsub. Increased impedance of the breast parenchyma may represent a novel biomarker of breast cancer risk associated with mammographic breast density. Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 4075.

Mammographic density estimation: Comparison among BI-RADS categories, a semi-automated software and a fully automated one

Although breast density is considered a strong predictor of breast cancer risk, its quantitative assessment is difficult. The aim of this study is to demonstrate that breast density assessment with a fully automated software is feasible and correlates with the semi-automated evaluation and the quantitative BI-RADS standards. A data set of 160 mammograms was evaluated by three blinded radiologists. Intra-observer (reader 1: k = 0.71; reader 2: k = 0.76; reader 3: k = 0.62) and inter-observer (reader 1 vs reader 2: k = 0.72; reader 2 vs reader 3: k = 0.80; reader 3 vs reader 1: k = 0.72) variability for the semi-automated software were good on a four-grade scale (D1/D2/D3/D4) and correlated with BI-RADS evaluation made by other two blinded radiologists ( r = 0.65, p < 0.01). Inter-observer (reader 1 vs reader 2: k = 0.85; reader 2 vs reader 3: k = 0.91; reader 3 vs reader 1: k = 0.85) variability for the semi-automated software was very good on a two-grade scale (D1–D2/D3–D4). The use of the fully automated software eliminated intra- and inter-observer differences, correlated with BI-RADS categories ( r = 0.62, p < 0.01) and can replace the semi-automated one (Bland–Altman statistics). Our study demonstrates that automated estimation of breast density is feasible and eliminates subjectivity. Furthermore both the semi-automated and the fully automated density estimation are more accurate than BI-RADS quantitative evaluation and could also be used in the daily clinical practice.

A Pilot Study of Compositional Analysis of the Breast and Estimation of Breast Mammographic Density Using Three-Dimensional T1-Weighted Magnetic Resonance Imaging

A method and computer tool to estimate percentage magnetic resonance (MR) imaging (MRI) breast density using three-dimensional T(1)-weighted MRI is introduced, and compared with mammographic percentage density [X-ray mammography (XRM)]. Ethical approval and informed consent were obtained. A method to assess MRI breast density as percentage volume occupied by water-containing tissue on three-dimensional T(1)-weighted MR images is described and applied in a pilot study to 138 subjects who were imaged by both MRI and XRM during the Magnetic Resonance Imaging in Breast Screening study. For comparison, percentage mammographic density was measured from matching XRMs as a ratio of dense to total projection areas scored visually using a 21-point score and measured by applying a two-dimensional interactive program (CUMULUS). The MRI and XRM percent methods were compared, including assessment of left-right and interreader consistency. Percent MRI density correlated strongly (r = 0.78; P < 0.0001) with percent mammographic density estimated using Cumulus. Comparison with visual assessment also showed a strong correlation. The mammographic methods overestimate density compared with MRI volumetric assessment by a factor approaching 2. MRI provides direct three-dimensional measurement of the proportion of water-based tissue in the breast. It correlates well with visual and computerized percent mammographic density measurements. This method may have direct application in women having breast cancer screening by breast MRI and may aid in determination of risk.

Estimation of mammographic density on an interval scale by transillumination breast spectroscopy

Transillumination breast spectroscopy (TiBS) uses nonionizing optical radiation to gain information about breast tissue morphological and structural properties. TiBS spectra are obtained from 232 women and compared to mammographic density (MD) quantified using Cumulus. The ability of TiBS to estimate MD is assessed using partial least-squares (PLS) regression methods, which requires TiBS spectra as input (X) and Cumulus MD as target (Y) data. Multiple PLS models are considered to determine the optimal processing technique(s) for the input (X) and target (Y) data. For each model, the association between TiBS estimated MD (Y) and Cumulus MD (Y) is established using Spearman's rank correlation and linear regression analysis. The model that best estimates MD has the fewest assumptions regarding target (Y) and spectral (X) processing. The Spearman's correlation coefficient between predicted MD and Cumulus MD for this model is 0.88, with a regression slope (beta) of 0.93 (95% CI 0.83-1.02) and an R(2) of 0.78. The approximation of individual MD was within 10% of Cumulus MD for the majority of women (80%), without stratification on age, body mass index (BMI), and menopausal status. TiBS provides an alternative to mammography assessed MD enabling frequent and earlier use of MD as a risk marker in preventive oncology.

Mammographic Density Measured with Quantitative Computer-aided Method: Comparison with Radiologists' Estimates and BI-RADS Categories

To retrospectively compare computer-aided mammographic density estimation (MDEST) with radiologist estimates of percentage density and Breast Imaging Reporting and Data System (BI-RADS) density classification. Institutional Review Board approval was obtained for this HIPAA-compliant study; patient informed consent requirements were waived. A fully automated MDEST computer program was used to measure breast density on digitized mammograms in 65 women (mean age, 53 years; range, 24-89 years). Pixel gray levels in detected breast borders were analyzed, and dense areas were segmented. Percentage density was calculated by dividing the number of dense pixels by the total number of pixels within the borders. Seven breast radiologists (five trained with MDEST, two not trained) prospectively assigned qualitative BI-RADS density categories and visually estimated percentage density on 260 mammograms. Qualitative BI-RADS assessments were compared with new quantitative BI-RADS standards. The reference standard density for this study was established by allowing the five trained radiologists to manipulate the MDEST gray-level thresholds, which segmented mammograms into dense and nondense areas. Statistical tests performed include Pearson correlation coefficients, Bland-Altman agreement method, kappa statistics, and unpaired t tests. There was a close correlation between the reference standard and radiologist-estimated density (R = 0.90-0.95) and MDEST density (R = 0.89). Untrained radiologists overestimated percentage density by an average of 37%, versus 6% for trained radiologists (P < .001). MDEST showed better agreement with the reference standard (average overestimate, 1%; range, -15% to +18%). MDEST correlated better with percentage density than with qualitative BI-RADS categories. There were large overlaps and ranges of percentage density in qualitative BI-RADS categories 2-4. Qualitative BI-RADS categories correlated poorly with new quantitative BI-RADS categories, and 16 (6%) of 260 views were erroneously classified by MDEST. MDEST compared favorably with radiologist estimates of percentage density and is more reproducible than radiologist estimates when qualitative BI-RADS density categories are used. Qualitative and quantitative BI-RADS density assessments differed markedly.

Computerized image analysis: estimation of breast density on mammograms.

An automated image analysis tool is being developed for the estimation of mammographic breast density. This tool may be useful for risk estimation or for monitoring breast density change in prevention or intervention programs. In this preliminary study, a data set of 4-view mammograms from 65 patients was used to evaluate our approach. Breast density analysis was performed on the digitized mammograms in three stages. First, the breast region was segmented from the surrounding background by an automated breast boundary-tracking algorithm. Second, an adaptive dynamic range compression technique was applied to the breast image to reduce the range of the gray level distribution in the low frequency background and to enhance the differences in the characteristic features of the gray level histogram for breasts of different densities. Third, rule-based classification was used to classify the breast images into four classes according to the characteristic features of their gray level histogram. For each image, a gray level threshold was automatically determined to segment the dense tissue from the breast region. The area of segmented dense tissue as a percentage of the breast area was then estimated. To evaluate the performance of the algorithm, the computer segmentation results were compared to manual segmentation with interactive thresholding by five radiologists. A "true" percent dense area for each mammogram was obtained by averaging the manually segmented areas of the radiologists. We found that the histograms of 6% (8 CC and 8 MLO views) of the breast regions were misclassified by the computer, resulting in poor segmentation of the dense region. For the images with correct classification, the correlation between the computer-estimated percent dense area and the "truth" was 0.94 and 0.91, respectively, for CC and MLO views, with a mean bias of less than 2%. The mean biases of the five radiologists' visual estimates for the same images ranged from 0.1% to 11%. The results demonstrate the feasibility of estimating mammographic breast density using computer vision techniques and its potential to improve the accuracy and reproducibility of breast density estimation in comparison with the subjective visual assessment by radiologists.