A quantitative description of the percentage of breast density measurement using full-field digital mammography

Reproductive Factors Related to Childbearing and a Novel Automated Mammographic Measure, V

BACKGROUND

We investigated the associations between several reproductive factors related to childbearing and the variation (V) measure (a novel, objective, single summary measure of breast image intensity) by menopausal status.

METHODS

Our study included 3,814 cancer-free women within the Nurses' Health Study (NHS) and NHSII cohorts. The data on reproductive variables and covariates were obtained from biennial questionnaires closest to the mammogram date. V-measures were obtained from mammographic images using a previously developed algorithm capturing the standard deviation of pixel values. We used multivariate linear regression to examine the associations of parity, age at first birth, time between menarche and first birth, time since last pregnancy, and lifetime breastfeeding duration with V-measure, adjusting for breast cancer risk factors, including the percentage of mammographic density (PMD). We further examined whether these associations were statistically accounted for (mediated) by PMD.

RESULTS

Among premenopausal women, none of the reproductive factors were associated with V. Among postmenopausal women, inverse associations of parity and positive associations of age at first birth with V were mediated by PMD (percent mediated: nulliparity: 66.7%, P < 0.0001; parity: 50.5%, P < 0.01; age at first birth 76.1%, P < 0.001) and were no longer significant in PMD-adjusted models. Lifetime duration of breastfeeding was positively associated with V [>36 vs. 0 ≤1 months β = 0.29; 95% confidence interval (CI) 0.07; 0.52, Ptrend < 0.01], independent of PMD.

CONCLUSIONS

Parity, age at first birth, and breastfeeding were associated with postmenopausal V.

IMPACT

This study highlights associations of reproductive factors with mammographic image intensity.

Breast Density Analysis Using Digital Breast Tomosynthesis

We evaluated an automated percentage of breast density (BD) technique (PDa) with digital breast tomosynthesis (DBT) data. The approach is based on the wavelet expansion followed by analyzing signal dependent noise. Several measures were investigated as risk factors: normalized volumetric; total dense volume; average of the DBT slices (slice-mean); a two-dimensional (2D) metric applied to the synthetic images; and the mean and standard deviations of the pixel values. Volumetric measures were derived theoretically, and PDa was modeled as a function of compressed breast thickness. An alternative method for constructing synthetic 2D mammograms was investigated using the volume results. A matched case-control study (n = 426 pairs) was analyzed. Conditional logistic regression modeling, controlling body mass index and ethnicity, was used to estimate odds ratios (ORs) for each measure with 95% confidence intervals provided parenthetically. There were several significant findings: volumetric measure [OR = 1.43 (1.18, 1.72)], which produced an identical OR as the slice-mean measure as predicted; [OR =1.44 (1.18, 1.75)] when applied to the synthetic images; and mean of the pixel values (volume or 2D synthetic) [ORs ~ 1.31 (1.09, 1.57)]. PDa was modeled as 2nd degree polynomial (concave-down): its maximum value occurred at 0.41×(compressed breast thickness), which was similar across case-control groups, and was significant from this position [OR = 1.47 (1.21, 1.78)]. A standardized 2D synthetic image was produced, where each pixel value represents the percentage of BD above its location. The significant findings indicate the validity of the technique. Derivations supported by empirical analyses produced a new synthetic 2D standardized image technique. Ancillary to the objectives, the results provide evidence for understanding the percentage of BD measure applied to 2D mammograms. Notwithstanding the findings, the study design provides a template for investigating other measures such as texture.

Improving the repeatability of deep learning models with Monte Carlo dropout

The integration of artificial intelligence into clinical workflows requires reliable and robust models. Repeatability is a key attribute of model robustness. Ideal repeatable models output predictions without variation during independent tests carried out under similar conditions. However, slight variations, though not ideal, may be unavoidable and acceptable in practice. During model development and evaluation, much attention is given to classification performance while model repeatability is rarely assessed, leading to the development of models that are unusable in clinical practice. In this work, we evaluate the repeatability of four model types (binary classification, multi-class classification, ordinal classification, and regression) on images that were acquired from the same patient during the same visit. We study the each model's performance on four medical image classification tasks from public and private datasets: knee osteoarthritis, cervical cancer screening, breast density estimation, and retinopathy of prematurity. Repeatability is measured and compared on ResNet and DenseNet architectures. Moreover, we assess the impact of sampling Monte Carlo dropout predictions at test time on classification performance and repeatability. Leveraging Monte Carlo predictions significantly increases repeatability, in particular at the class boundaries, for all tasks on the binary, multi-class, and ordinal models leading to an average reduction of the 95% limits of agreement by 16% points and of the class disagreement rate by 7% points. The classification accuracy improves in most settings along with the repeatability. Our results suggest that beyond about 20 Monte Carlo iterations, there is no further gain in repeatability. In addition to the higher test-retest agreement, Monte Carlo predictions are better calibrated which leads to output probabilities reflecting more accurately the true likelihood of being correctly classified.

A genome-wide association study of mammographic texture variation

Background

Breast parenchymal texture features, including grayscale variation (V), capture the patterns of texture variation on a mammogram and are associated with breast cancer risk, independent of mammographic density (MD). However, our knowledge on the genetic basis of these texture features is limited.

Methods

We conducted a genome-wide association study of V in 7040 European-ancestry women. V assessments were generated from digitized film mammograms. We used linear regression to test the single-nucleotide polymorphism (SNP)-phenotype associations adjusting for age, body mass index (BMI), MD phenotypes, and the top four genetic principal components. We further calculated genetic correlations and performed SNP-set tests of V with MD, breast cancer risk, and other breast cancer risk factors.

Results

We identified three genome-wide significant loci associated with V: rs138141444 (6q24.1) in ECT2L, rs79670367 (8q24.22) in LINC01591, and rs113174754 (12q22) near PGAM1P5. 6q24.1 and 8q24.22 have not previously been associated with MD phenotypes or breast cancer risk, while 12q22 is a known locus for both MD and breast cancer risk. Among known MD and breast cancer risk SNPs, we identified four variants that were associated with V at the Bonferroni-corrected thresholds accounting for the number of SNPs tested: rs335189 (5q23.2) in PRDM6, rs13256025 (8p21.2) in EBF2, rs11836164 (12p12.1) near SSPN, and rs17817449 (16q12.2) in FTO. We observed significant genetic correlations between V and mammographic dense area (r_g = 0.79, P = 5.91 × 10⁻⁵), percent density (r_g = 0.73, P = 1.00 × 10⁻⁴), and adult BMI (r_g =  − 0.36, P = 3.88 × 10⁻⁷). Additional significant relationships were observed for non-dense area (z =  − 4.14, P = 3.42 × 10⁻⁵), estrogen receptor-positive breast cancer (z = 3.41, P = 6.41 × 10⁻⁴), and childhood body fatness (z =  − 4.91, P = 9.05 × 10⁻⁷) from the SNP-set tests.

Conclusions

These findings provide new insights into the genetic basis of mammographic texture variation and their associations with MD, breast cancer risk, and other breast cancer risk factors.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13058-022-01570-8.

Associations of Oral Contraceptives with Mammographic Breast Density in Premenopausal Women

BACKGROUND

We investigated the associations of oral contraceptives (OC) with percent breast density (PD), absolute dense area (DA), nondense area (NDA), and a novel image intensity variation (V) measure in premenopausal women.

METHODS

This study included 1,233 controls from a nested case-control study within Nurses' Health Study II cohort. Information on OCs was collected in 1989 and updated biennially. OC use was defined from the questionnaire closest to the mammogram date. PD, DA, and NDA were measured from digitized film mammograms using a computer-assisted thresholding technique; the V measure was obtained with a previously developed algorithm measuring the SD of pixel values in the eroded breast region. Generalized linear regression was used to assess associations between OCs and density measures (square root-transformed PD, DA, and NDA, and -untransformed V).

RESULTS

OC use was not associated with PD [current vs. never: β = -0.06; 95% confidence interval (CI), -0.37-0.24; past vs. never: β = 0.10; 95% CI, -0.09-0.29], DA (current vs. never: β = -0.20; 95% CI -0.59-0.18; past vs. never: β = 0.13; 95% CI, -0.12-0.39), and NDA (current vs. never: β = -0.19; 95% CI, -0.56-0.18; past vs. never: β = -0.01; 95% CI, -0.28-0.25). Women with younger age at initiation had significantly greater V-measure (<20 years vs. never: β = 26.88; 95% CI, 3.18-50.58; 20-24 years vs. never: β = 20.23; 95% CI, -4.24-44.71; 25-29 years vs. never: β = 2.61; 95% CI -29.00-34.23; ≥30 years vs. never: β = 0.28; 95% CI, -34.16-34.72, P _trend = 0.03).

CONCLUSIONS

Our findings suggest that an earlier age at first OC use was associated with significantly greater V.

IMPACT

These findings could guide decisions about the age for OC initiation.

Automated percent mammographic density, mammographic texture variation, and risk of breast cancer: a nested case-control study

Percent mammographic density (PMD) is a strong breast cancer risk factor, however, other mammographic features, such as V, the standard deviation (SD) of pixel intensity, may be associated with risk. We assessed whether PMD, automated PMD (APD), and V, yielded independent associations with breast cancer risk. We included 1900 breast cancer cases and 3921 matched controls from the Nurses' Health Study (NHS) and the NHSII. Using digitized film mammograms, we estimated PMD using a computer-assisted thresholding technique. APD and V were determined using an automated computer algorithm. We used logistic regression to generate odds ratios (ORs) and 95% confidence intervals (CIs). Median time from mammogram to diagnosis was 4.1 years (interquartile range: 1.6-6.8 years). PMD (OR per SD:1.52, 95% CI: 1.42, 1.63), APD (OR per SD:1.32, 95% CI: 1.24, 1.41), and V (OR per SD:1.32, 95% CI: 1.24, 1.40) were positively associated with breast cancer risk. Associations for APD were attenuated but remained statistically significant after mutual adjustment for PMD or V. Women in the highest quartile of both APD and V (OR vs Q1/Q1: 2.49, 95% CI: 2.02, 3.06), or PMD and V (OR vs Q1/Q1: 3.57, 95% CI: 2.79, 4.58) had increased breast cancer risk. An automated method of PMD assessment is feasible and yields similar, but somewhat weaker, estimates to a manual measure. PMD, APD and V are each independently, positively associated with breast cancer risk. Women with dense breasts and greater texture variation are at the highest relative risk of breast cancer.

Empirically-derived synthetic populations to mitigate small sample sizes

Limited sample sizes can lead to spurious modeling findings in biomedical research. The objective of this work is to present a new method to generate synthetic populations (SPs) from limited samples using matched case-control data (n = 180 pairs), considered as two separate limited samples. SPs were generated with multivariate kernel density estimations (KDEs) with unconstrained bandwidth matrices. We included four continuous variables and one categorical variable for each individual. Bandwidth matrices were determined with Differential Evolution (DE) optimization by covariance comparisons. Four synthetic samples (n = 180) were derived from their respective SPs. Similarity between observed samples with synthetic samples was compared assuming their empirical probability density functions (EPDFs) were similar. EPDFs were compared with the maximum mean discrepancy (MMD) test statistic based on the Kernel Two-Sample Test. To evaluate similarity within a modeling context, EPDFs derived from the Principal Component Analysis (PCA) scores and residuals were summarized with the distance to the model in X-space (DModX) as additional comparisons. Four SPs were generated from each sample. The probability of selecting a replicate when randomly constructing synthetic samples (n = 180) was infinitesimally small. MMD tests indicated that the observed sample EPDFs were similar to the respective synthetic EPDFs. For the samples, PCA scores and residuals did not deviate significantly when compared with their respective synthetic samples. The feasibility of this approach was demonstrated by producing synthetic data at the individual level, statistically similar to the observed samples. The methodology coupled KDE with DE optimization and deployed novel similarity metrics derived from PCA. This approach could be used to generate larger-sized synthetic samples. To develop this approach into a research tool for data exploration purposes, additional evaluation with increased dimensionality is required. Moreover, given a fully specified population, the degree to which individuals can be discarded while synthesizing the respective population accurately will be investigated. When these objectives are addressed, comparisons with other techniques such as bootstrapping will be required for a complete evaluation.

Calibrated Breast Density Measurements

RATIONALE AND OBJECTIVES

Mammographic density is an important risk factor for breast cancer, but translation to the clinic requires assurance that prior work based on mammography is applicable to current technologies. The purpose of this work is to evaluate whether a calibration methodology developed previously produces breast density metrics predictive of breast cancer risk when applied to a case-control study.

MATERIALS AND METHODS

A matched case control study (n = 319 pairs) was used to evaluate two calibrated measures of breast density. Two-dimensional mammograms were acquired from six Hologic mammography units: three conventional Selenia two-dimensional full-field digital mammography systems and three Dimensions digital breast tomosynthesis systems. We evaluated the capability of two calibrated breast density measures to quantify breast cancer risk: the mean (PGm) and standard deviation (PGsd) of the calibrated pixels. Matching variables included age, hormone replacement therapy usage/duration, screening history, and mammography unit. Calibrated measures were compared to the percentage of breast density (PD) determined with the operator-assisted Cumulus method. Conditional logistic regression was used to generate odds ratios (ORs) from continuous and quartile (Q) models with 95% confidence intervals. The area under the receiver operating characteristic curve (Az) was also used as a comparison metric. Both univariate models and models adjusted for body mass index and ethnicity were evaluated.

RESULTS

In adjusted models, both PGsd and PD were statistically significantly associated with breast cancer with similar Az of 0.61-0.62. The corresponding ORs and confidence intervals were also similar. For PGsd, the OR was 1.34 (1.09, 1.66) for the continuous measure and 1.83 (1.11, 3.02), 2.19 (1.28, 3.73), and 2.20 (1.26, 3.85) for Q2-Q4. For PD, the OR was 1.43 (1.16, 1.76) for the continuous measure and 0.84 (0.52, 1.38), 1.96 (1.19, 3.23), and 2.27 (1.29, 4.00) for Q2-Q4. The results for PGm were slightly attenuated and not statistically significant. The OR was 1.22 (0.99, 1.51) with Az = 0.60 for the continuous measure and 1.24 (0.78, 1.97), 0.98 (0.60, 1.61), and 1.26, (0.77, 2.07) for Q2-Q4 with Az = 0.60.

CONCLUSION

The calibrated PGsd measure provided significant associations with breast cancer comparable to those given by PD. The calibrated PGm performed slightly worse. These findings indicate that the calibration approach developed previously replicates under more general conditions.

Spatial Correlation and Breast Cancer Risk

We present a novel method for evaluating local spatial correlation structure in two-dimensional (2D) mammograms and evaluate its capability for risk prediction as one possible application. Two matched case-control studies were analyzed. Study 1 included women (N = 588 pairs) with mammograms acquired with either Hologic Selenia full field digital mammography (FFDM) units or Hologic Dimensions digital breast tomosynthesis units. Study 2 included women (N =180 pairs) with mammograms acquired with a General Electric Senographe 2000D FFDM unit. Matching variables included age, HRT usage/duration, screening history, and mammography unit. Local autocorrelation functions were determined with Fourier analysis and compared with a template defined as a 2D double-sided exponential function with one spatial extent parameter: n = 4, 12, 24, 50, 74, 100, and 124, where (n+1)×(n+1) is the area of the local spatial extent measured in pixels. The difference between the local correlation and template was gauged within an adjustable parameter kernel and summarized, producing two measures: the mean (m_n+1)_, and standard deviation (s_n+1). Both adjustable parameters were varied in Study 1. Select measures that produced significant associations with breast cancer were translated to Study 2. Breast cancer associations were evaluated with conditional logistic regression, adjusted for body mass index and ethnicity. Odds ratios (ORs) were estimated as per standard deviation increment with 95% confidence intervals (CIs). Two measures were selected for breast cancer association analysis in Study 1: m₇₅ and s₂₅. Both measures revealed significant associations with breast cancer: OR = 1.45 (1.23, 1.66) for m₇₅ and OR = 1.30 (1.14, 1.49) for s_25. When translating to Study 2, these measures also revealed significant associations: OR = 1.49 (1.12, 1.96) for m₇₅ and OR = 1.34 (1.06, 1.69) for s₂₅. Novel correlation metrics presented in this work produced significant associations with breast cancer risk. This approach is general and may have applications beyond mammography.

Technical challenges in generalizing calibration techniques for breast density measurements

PURPOSE

We are developing a calibration methodology for full-field digital mammography (FFDM). Calibration compensates for image acquisition technique influences on the pixel representation, ideally producing improved inter-image breast density estimates. This approach relies on establishing references with rigid breast tissue-equivalent phantoms (BTEs) and requires an accurate estimate of the compressed breast thickness because the system readout is nominal. There is also an attenuation mismatch between adipose breast tissue and the adipose BTE that was noted in our previous work. It is referred to as the "attenuation anomaly" and addressed in this report. The objectives are to evaluate methods to correct for the compressed breast thickness and compensate for the attenuation anomaly.

METHODS

Thickness correction surfaces were established with a deformable phantom (DP) using both image and physical measurements for three direct x-ray conversion FFDM units. The Cumulative Sum serial quality control procedure was established to ensure the thickness correction measurements were stable over time by imaging and calibrating DPs biweekly in lieu of physical measurements. The attenuation anomaly was addressed by evaluating adipose image regions coupled with an optimization technique to adjust the adipose calibration data. We compared calibration consistency across matched left and right cranial caudal (CC) mammographic views (n = 199) with and without corrections using Bland-Altman plots. These plots were complemented by comparing the right and left breast calibrated average (μ_a ) and population distribution mean (m_a ) with 95% confidence intervals and difference distribution variances with the F-test for uncorrected and corrected data.

RESULTS

Thickness correction surfaces were well approximated as tilted planes and were dependent upon compression force. A correction was developed for the attenuation anomaly. All paddles (large and small paddles for all units) exhibited similar tilt as a function of force. Without correction, m_a  = 0.92 (-1.77, 3.62) was not significantly different from zero with many negative μ_a samples. The thickness correction produced a significant shift in the μ_a distribution in the positive direction with m_a  = 13.99 (11.17, 16.80) and reduced the difference distribution variance significantly (P < 0.0001). Applying both corrections in tandem gave m_a  = 22.83 (20.32, 25.34), representing another significant positive shift in comparison with the thickness correction in isolation. Thickness corrections were stable over approximately a 2-year timeframe for all units.

CONCLUSION

These correction techniques are valid approaches for addressing technical problems with calibration that relies on reference phantoms. The efficacy of the calibration methodology will require validation with clinical endpoints in future studies.

Generalized breast density metrics

Mammograms represent data that can inform future risk of breast cancer. Data from two case-control study populations were analyzed. Population 1 included women (N  =  180 age matched case-control pairs) with mammograms acquired with one indirect x-ray conversion mammography unit. Population 2 included women (N  =  319 age matched case-control pairs) with mammograms acquired from 6 direct x-ray conversion units. The Fourier domain was decomposed into n concentric rings (radial spatial frequency bands). The power in each ring was summarized giving a set of measures. We investigated images in raw, for presentation (processed) and calibrated representations and made comparison with the percentage of breast density (BD) determined with the operator assisted Cumulus method. Breast cancer associations were evaluated with conditional logistic regression, adjusted for body mass index and ethnicity. Odds ratios (ORs), per standard deviation increase derived from the respective breast density distributions and 95% confidence intervals (CIs) were estimated. A measure from a lower radial frequency ring, corresponding 0.083-0.166 cycles mm^-1 and BD had significant associations with risk in both populations. In Population 1, the Fourier measure produced significant associations in each representation: OR  =  1.76 (1.33, 2.32) for raw; OR  =  1.43 (1.09, 1.87) for processed; and OR  =  1.68 (1.26, 2.25) for calibrated. BD also provided significant associations in Population 1: OR  =  1.72 (1.27, 2.33). In Population 2, the Fourier measure produced significant associations for each representation as well: OR  =  1.47 (1.19, 1.80) for raw; OR  =  1.38 (1.15, 1.67) for processed; and OR  =  1.42 (1.15, 1.75) for calibrated. BD provided significant associations in Population 2: OR  =  1.43 (1.17, 1.76). Other coincident spectral regions were also predictive of case-control status. In sum, generalized breast density measures were significantly associated with breast cancer in both FFDM technologies.

The role of matricellular proteins and tissue stiffness in breast cancer: a systematic review

Malignancies consist not only of cancerous and nonmalignant cells, but also of additional elements, as extracellular matrix. The aim of this review is to summarize meta-analyses, describing breast tissue stiffness and risk of breast carcinoma (BC) assessing the potential relationship between matricellular proteins (MPs) and survival. A systematic computer-based search of published articles, according to PRISMA statement, was conducted through Ovid interface. Mammographic density and tissue stiffness are associated with the risk of BC development, suggesting that MPs may influence BC prognosis. No definitive conclusions are available and additional researches are required to definitively clarify the role of each MP, mammographic density and stiffness in BC development and the mechanisms involved in the onset of this malignancy.

Breast parenchymal patterns in processed versus raw digital mammograms: A large population study toward assessing differences in quantitative measures across image representations

PURPOSE

With raw digital mammograms (DMs), which retain the relationship with x-ray attenuation of the breast tissue, not being routinely available, processed DMs are often the only viable means to acquire imaging measures. The authors investigate differences in quantitative measures of breast density and parenchymal texture, shown to have value in breast cancer risk assessment, between the two DM representations.

METHODS

The authors report data from 8458 pairs of bilateral raw ("FOR PROCESSING") and processed ("FOR PRESENTATION") DMs acquired from 4278 women undergoing routine screening evaluation, collected with DM units from two different vendors. Breast dense tissue area and percent density (PD), as well as a range of quantitative descriptors of breast parenchymal texture (statistical, co-occurrence, run-length, and structural descriptors), were measured using previously validated, fully automated software. Feature measurements were compared using matched-pairs Wilcoxon signed-ranks test, correlation (r), and linear-mixed-effects (LME) models, where potential interactions with woman- and system-specific factors were also assessed. The authors also compared texture feature correlations with the established risk factors of the Gail lifetime risk score (r_G) and breast PD (r_PD), and evaluated the within woman intraclass feature correlation (ICC), a measure of bilateral breast-tissue symmetry, in raw versus processed images.

RESULTS

All density measures and most of the texture features were strongly (r ≥ 0.6) or moderately (0.4 ≤ r < 0.6) correlated between raw and processed images. However, measurements were significantly different between the two imaging formats (Wilcoxon signed-ranks test, p_w < 0.05). The association between measurements varied across features and vendors, and was substantially modified by woman- and system-specific image acquisition factors, such as age, BMI, and mAs/kVp, respectively. The strongest correlation, combined with minimal LME-model interactions, was observed for structural texture features. Overall, texture measures from either image representation were weakly associated with Gail lifetime risk (-0.2 ≤ r_G ≤ 0.2), weakly to moderately associated with breast PD (-0.6 ≤ r_PD ≤ 0.6), and had overall strong bilateral symmetry (ICC ≥ 0.6).

CONCLUSIONS

Differences in measures from processed versus raw DM depend highly on the feature, the DM vendor, and image acquisition settings, where structural features appear to be more robust across the different DM settings. The reported findings may serve as a reference in the design of future large-scale studies on mammographic features and breast cancer risk assessment involving multiple DM representations.

Parenchymal texture analysis in digital mammography: A fully automated pipeline for breast cancer risk assessment

PURPOSE

Mammographic percent density (PD%) is known to be a strong risk factor for breast cancer. Recent studies also suggest that parenchymal texture features, which are more granular descriptors of the parenchymal pattern, can provide additional information about breast cancer risk. To date, most studies have measured mammographic texture within selected regions of interest (ROIs) in the breast, which cannot adequately capture the complexity of the parenchymal pattern throughout the whole breast. To better characterize patterns of the parenchymal tissue, the authors have developed a fully automated software pipeline based on a novel lattice-based strategy to extract a range of parenchymal texture features from the entire breast region.

METHODS

Digital mammograms from 106 cases with 318 age-matched controls were retrospectively analyzed. The lattice-based approach is based on a regular grid virtually overlaid on each mammographic image. Texture features are computed from the intersection (i.e., lattice) points of the grid lines within the breast, using a local window centered at each lattice point. Using this strategy, a range of statistical (gray-level histogram, co-occurrence, and run-length) and structural (edge-enhancing, local binary pattern, and fractal dimension) features are extracted. To cover the entire breast, the size of the local window for feature extraction is set equal to the lattice grid spacing and optimized experimentally by evaluating different windows sizes. The association between their lattice-based texture features and breast cancer was evaluated using logistic regression with leave-one-out cross validation and further compared to that of breast PD% and commonly used single-ROI texture features extracted from the retroareolar or the central breast region. Classification performance was evaluated using the area under the curve (AUC) of the receiver operating characteristic (ROC). DeLong's test was used to compare the different ROCs in terms of AUC performance.

RESULTS

The average univariate performance of the lattice-based features is higher when extracted from smaller than larger window sizes. While not every individual texture feature is superior to breast PD% (AUC: 0.59, STD: 0.03), their combination in multivariate analysis has significantly better performance (AUC: 0.85, STD: 0.02, p < 0.001). The lattice-based texture features also outperform the single-ROI texture features when extracted from the retroareolar or the central breast region (AUC: 0.60-0.74, STD: 0.03). Adding breast PD% does not make a significant performance improvement to the lattice-based texture features or the single-ROI features (p > 0.05).

CONCLUSIONS

The proposed lattice-based strategy for mammographic texture analysis enables to characterize the parenchymal pattern over the entire breast. As such, these features provide richer information compared to currently used descriptors and may ultimately improve breast cancer risk assessment. Larger studies are warranted to validate these findings and also compare to standard demographic and reproductive risk factors.

Comparison of mammographic density estimation by Volpara software with radiologists' visual assessment: analysis of clinical-radiologic factors affecting discrepancy between them

BACKGROUND

Volumetric breast density analysis is useful for quantitative mammographic assessment. However, there are few studies about clinical-radiologic factors contributing to discrepancies in the visual assessment by radiologists.

PURPOSE

To compare automated volumetric breast density measurement with BI-RADS breast density category by radiologists' visual assessments and to evaluate the clinical-radiologic factors affecting disagreement between two estimations.

MATERIAL AND METHODS

From February 2011 to September 2012, 860 patients (mean age, 54.7 ± 10.2 years) who had undergone digital mammography including fully automated volumetric breast density analysis, were enrolled. The agreement in breast density assessments between two radiologists, and between an experienced radiologist and the automated software were evaluated using a weighted kappa (k) value. Clinical-radiologic factors contributing to disagreement between the results obtained by a radiologist and the automated software were evaluated using univariate and multivariate analysis.

RESULTS

Breast density assessments obtained by two different radiologists were in good agreement (weighted k statistics 0.835%; 95% confidence interval [CI], 0.8098-0.8608); breast density assessments obtained by an experienced radiologist versus automated software were in moderate agreement (weighted k statistics 0.799%; 95% CI, 0.7708-0.8263). Univariate analysis identified a difference in bilateral breast density and patient age as two factors that significantly contributed to disagreement between the two approaches (P = 0.0002, P = 0.019). Multivariate analysis only identified a difference in bilateral breast density as a contributing factor.

CONCLUSION

The automated volumetric breast density measurement showed good agreement with radiologists' assessment. The difference in bilateral breast density affected the disagreement between results from visual assessment and automated software.

A Review on Automatic Mammographic Density and Parenchymal Segmentation

Breast cancer is the most frequently diagnosed cancer in women. However, the exact cause(s) of breast cancer still remains unknown. Early detection, precise identification of women at risk, and application of appropriate disease prevention measures are by far the most effective way to tackle breast cancer. There are more than 70 common genetic susceptibility factors included in the current non-image-based risk prediction models (e.g., the Gail and the Tyrer-Cuzick models). Image-based risk factors, such as mammographic densities and parenchymal patterns, have been established as biomarkers but have not been fully incorporated in the risk prediction models used for risk stratification in screening and/or measuring responsiveness to preventive approaches. Within computer aided mammography, automatic mammographic tissue segmentation methods have been developed for estimation of breast tissue composition to facilitate mammographic risk assessment. This paper presents a comprehensive review of automatic mammographic tissue segmentation methodologies developed over the past two decades and the evidence for risk assessment/density classification using segmentation. The aim of this review is to analyse how engineering advances have progressed and the impact automatic mammographic tissue segmentation has in a clinical environment, as well as to understand the current research gaps with respect to the incorporation of image-based risk factors in non-image-based risk prediction models.

A population-based tissue probability map-driven level set method for fully automated mammographic density estimations

PURPOSE

A major challenge when distinguishing glandular tissues on mammograms, especially for area-based estimations, lies in determining a boundary on a hazy transition zone from adipose to glandular tissues. This stems from the nature of mammography, which is a projection of superimposed tissues consisting of different structures. In this paper, the authors present a novel segmentation scheme which incorporates the learned prior knowledge of experts into a level set framework for fully automated mammographic density estimations.

METHODS

The authors modeled the learned knowledge as a population-based tissue probability map (PTPM) that was designed to capture the classification of experts' visual systems. The PTPM was constructed using an image database of a selected population consisting of 297 cases. Three mammogram experts extracted regions for dense and fatty tissues on digital mammograms, which was an independent subset used to create a tissue probability map for each ROI based on its local statistics. This tissue class probability was taken as a prior in the Bayesian formulation and was incorporated into a level set framework as an additional term to control the evolution and followed the energy surface designed to reflect experts' knowledge as well as the regional statistics inside and outside of the evolving contour.

RESULTS

A subset of 100 digital mammograms, which was not used in constructing the PTPM, was used to validate the performance. The energy was minimized when the initial contour reached the boundary of the dense and fatty tissues, as defined by experts. The correlation coefficient between mammographic density measurements made by experts and measurements by the proposed method was 0.93, while that with the conventional level set was 0.47.

CONCLUSIONS

The proposed method showed a marked improvement over the conventional level set method in terms of accuracy and reliability. This result suggests that the proposed method successfully incorporated the learned knowledge of the experts' visual systems and has potential to be used as an automated and quantitative tool for estimations of mammographic breast density levels.

Calibrated breast density methods for full field digital mammography: a system for serial quality control and inter-system generalization

PURPOSE

The authors are developing a system for calibrated breast density measurements using full field digital mammography (FFDM). Breast tissue equivalent (BTE) phantom images are used to establish baseline (BL) calibration curves at time zero. For a given FFDM unit, the full BL dataset is comprised of approximately 160 phantom images, acquired prior to calibrating prospective patient mammograms. BL curves are monitored serially to ensure they produce accurate calibration and require updating when calibration accuracy degrades beyond an acceptable tolerance, rather than acquiring full BL datasets repeatedly. BL updating is a special case of generalizing calibration datasets across FFDM units, referred to as cross-calibration. Serial monitoring, BL updating, and cross-calibration techniques were developed and evaluated.

METHODS

BL curves were established for three Hologic Selenia FFDM units at time zero. In addition, one set of serial phantom images, comprised of equal proportions of adipose and fibroglandular BTE materials (50/50 compositions) of a fixed height, was acquired biweekly and monitored with the cumulative sum (Cusum) technique. These 50/50 composition images were used to update the BL curves when the calibration accuracy degraded beyond a preset tolerance of ±4 standardized units. A second set of serial images, comprised of a wide-range of BTE compositions, was acquired biweekly to evaluate serial monitoring, BL updating, and cross-calibration techniques.

RESULTS

Calibration accuracy can degrade serially and is a function of acquisition technique and phantom height. The authors demonstrated that all heights could be monitored simultaneously while acquiring images of a 50/50 phantom with a fixed height for each acquisition technique biweekly, translating into approximately 16 image acquisitions biweekly per FFDM unit. The same serial images are sufficient for serial monitoring, BL updating, and cross-calibration. Serial calibration accuracy was maintained within ±4 standardized unit variation from the ideal when applying BL updating. BL updating is a special case of cross-calibration; the BL dataset of unit 1 can be converted to the BL dataset for another similar unit (i.e., unit 2) at any given time point using the 16 serial monitoring 50/50 phantom images of unit 2 (or vice versa) acquired near this time point while maintaining the ±4 standardized unit tolerance.

CONCLUSIONS

A methodology for monitoring and maintaining serial calibration accuracy for breast density measurements was evaluated. Calibration datasets for a given unit can be translated forward in time with minimal phantom imaging effort. Similarly, cross-calibration is a method for generalizing calibration datasets across similar units without additional phantom imaging. This methodology will require further evaluation with mammograms for complete validation.

Area and volumetric density estimation in processed full-field digital mammograms for risk assessment of breast cancer

Introduction

Mammographic density, the white radiolucent part of a mammogram, is a marker of breast cancer risk and mammographic sensitivity. There are several means of measuring mammographic density, among which are area-based and volumetric-based approaches. Current volumetric methods use only unprocessed, raw mammograms, which is a problematic restriction since such raw mammograms are normally not stored. We describe fully automated methods for measuring both area and volumetric mammographic density from processed images.

Methods

The data set used in this study comprises raw and processed images of the same view from 1462 women. We developed two algorithms for processed images, an automated area-based approach (CASAM-Area) and a volumetric-based approach (CASAM-Vol). The latter method was based on training a random forest prediction model with image statistical features as predictors, against a volumetric measure, Volpara, for corresponding raw images. We contrast the three methods, CASAM-Area, CASAM-Vol and Volpara directly and in terms of association with breast cancer risk and a known genetic variant for mammographic density and breast cancer, rs10995190 in the gene ZNF365. Associations with breast cancer risk were evaluated using images from 47 breast cancer cases and 1011 control subjects. The genetic association analysis was based on 1011 control subjects.

Results

All three measures of mammographic density were associated with breast cancer risk and rs10995190 (p<0.025 for breast cancer risk and p<1×10⁻⁶ for rs10995190). After adjusting for one of the measures there remained little or no evidence of residual association with the remaining density measures (p>0.10 for risk, p>0.03 for rs10995190).

Conclusions

Our results show that it is possible to obtain reliable automated measures of volumetric and area mammographic density from processed digital images. Area and volumetric measures of density on processed digital images performed similar in terms of risk and genetic association.

Breast Imaging Reporting and Data System (BI-RADS) breast composition descriptors: automated measurement development for full field digital mammography

PURPOSE

The Breast Imaging Reporting and Data System (BI-RADS) breast composition descriptors are used for standardized mammographic reporting and are assessed visually. This reporting is clinically relevant because breast composition can impact mammographic sensitivity and is a breast cancer risk factor. New techniques are presented and evaluated for generating automated BI-RADS breast composition descriptors using both raw and calibrated full field digital mammography (FFDM) image data.

METHODS

A matched case-control dataset with FFDM images was used to develop three automated measures for the BI-RADS breast composition descriptors. Histograms of each calibrated mammogram in the percent glandular (pg) representation were processed to create the new BR(pg) measure. Two previously validated measures of breast density derived from calibrated and raw mammograms were converted to the new BR(vc) and BR(vr) measures, respectively. These three measures were compared with the radiologist-reported BI-RADS compositions assessments from the patient records. The authors used two optimization strategies with differential evolution to create these measures: method-1 used breast cancer status; and method-2 matched the reported BI-RADS descriptors. Weighted kappa (κ) analysis was used to assess the agreement between the new measures and the reported measures. Each measure's association with breast cancer was evaluated with odds ratios (ORs) adjusted for body mass index, breast area, and menopausal status. ORs were estimated as per unit increase with 95% confidence intervals.

RESULTS

The three BI-RADS measures generated by method-1 had κ between 0.25-0.34. These measures were significantly associated with breast cancer status in the adjusted models: (a) OR = 1.87 (1.34, 2.59) for BR(pg); (b) OR = 1.93 (1.36, 2.74) for BR(vc); and (c) OR = 1.37 (1.05, 1.80) for BR(vr). The measures generated by method-2 had κ between 0.42-0.45. Two of these measures were significantly associated with breast cancer status in the adjusted models: (a) OR = 1.95 (1.24, 3.09) for BR(pg); (b) OR = 1.42 (0.87, 2.32) for BR(vc); and (c) OR = 2.13 (1.22, 3.72) for BR(vr). The radiologist-reported measures from the patient records showed a similar association, OR = 1.49 (0.99, 2.24), although only borderline statistically significant.

CONCLUSIONS

A general framework was developed and validated for converting calibrated mammograms and continuous measures of breast density to fully automated approximations for the BI-RADS breast composition descriptors. The techniques are general and suitable for a broad range of clinical and research applications.

Automated Percentage of Breast Density Measurements for Full-field Digital Mammography Applications

RATIONALE AND OBJECTIVES

Increased mammographic breast density is a significant risk factor for breast cancer. A reproducible, accurate, and automated breast density measurement is required for full-field digital mammography (FFDM) to support clinical applications. We evaluated a novel automated percentage of breast density measure (PDa) and made comparisons with the standard operator-assisted measure (PD) using FFDM data.

METHODS

We used a nested breast cancer case-control study matched on age, year of mammogram and diagnosis with images acquired from a specific direct x-ray conversion FFDM technology. PDa was applied to the raw and clinical display (or processed) representation images. We evaluated the transformation (pixel mapping) of the raw image, giving a third representation (raw-transformed), to improve the PDa performance using differential evolution optimization. We applied PD to the raw and clinical display images as a standard for measurement comparison. Conditional logistic regression was used to estimate the odd ratios (ORs) for breast cancer with 95% confidence intervals (CI) for all measurements; analyses were adjusted for body mass index. PDa operates by evaluating signal-dependent noise (SDN), captured as local signal variation. Therefore, we characterized the SDN relationship to understand the PDa performance as a function of data representation and investigated a variation analysis of the transformation.

RESULTS

The associations of the quartiles of operator-assisted PD with breast cancer were similar for the raw (OR: 1.00 [ref.]; 1.59 [95% CI, 0.93-2.70]; 1.70 [95% CI, 0.95-3.04]; 2.04 [95% CI, 1.13-3.67]) and clinical display (OR: 1.00 [ref.]; 1.31 [95% CI, 0.79-2.18]; 1.14 [95% CI, 0.65-1.98]; 1.95 [95% CI, 1.09-3.47]) images. PDa could not be assessed on the raw images without preprocessing. However, PDa had similar associations with breast cancer when assessed on 1) raw-transformed (OR: 1.00 [ref.]; 1.27 [95% CI, 0.74-2.19]; 1.86 [95% CI, 1.05-3.28]; 3.00 [95% CI, 1.67-5.38]) and 2) clinical display (OR: 1.00 [ref.]; 1.79 [95% CI, 1.04-3.11]; 1.61 [95% CI, 0.90-2.88]; 2.94 [95% CI, 1.66-5.19]) images. The SDN analysis showed that a nonlinear relationship between the mammographic signal and its variation (ie, the biomarker for the breast density) is required for PDa. Although variability in the transform influenced the respective PDa distribution, it did not affect the measurement's association with breast cancer.

CONCLUSIONS

PDa assessed on either raw-transformed or clinical display images is a valid automated breast density measurement for a specific FFDM technology and compares well against PD. Further work is required for measurement generalization.

Mammographic density phenotypes and risk of breast cancer: a meta-analysis

BACKGROUND

Fibroglandular breast tissue appears dense on mammogram, whereas fat appears nondense. It is unclear whether absolute or percentage dense area more strongly predicts breast cancer risk and whether absolute nondense area is independently associated with risk.

METHODS

We conducted a meta-analysis of 13 case-control studies providing results from logistic regressions for associations between one standard deviation (SD) increments in mammographic density phenotypes and breast cancer risk. We used random-effects models to calculate pooled odds ratios and 95% confidence intervals (CIs). All tests were two-sided with P less than .05 considered to be statistically significant.

RESULTS

Among premenopausal women (n = 1776 case patients; n = 2834 control subjects), summary odds ratios were 1.37 (95% CI = 1.29 to 1.47) for absolute dense area, 0.78 (95% CI = 0.71 to 0.86) for absolute nondense area, and 1.52 (95% CI = 1.39 to 1.66) for percentage dense area when pooling estimates adjusted for age, body mass index, and parity. Corresponding odds ratios among postmenopausal women (n = 6643 case patients; n = 11187 control subjects) were 1.38 (95% CI = 1.31 to 1.44), 0.79 (95% CI = 0.73 to 0.85), and 1.53 (95% CI = 1.44 to 1.64). After additional adjustment for absolute dense area, associations between absolute nondense area and breast cancer became attenuated or null in several studies and summary odds ratios became 0.82 (95% CI = 0.71 to 0.94; P heterogeneity = .02) for premenopausal and 0.85 (95% CI = 0.75 to 0.96; P heterogeneity < .01) for postmenopausal women.

CONCLUSIONS

The results suggest that percentage dense area is a stronger breast cancer risk factor than absolute dense area. Absolute nondense area was inversely associated with breast cancer risk, but it is unclear whether the association is independent of absolute dense area.

Mammographic density estimation with automated volumetric breast density measurement

OBJECTIVE

To 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 METHODS

In 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.

RESULTS

The 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).

CONCLUSION

There 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.

Mammographic texture resemblance generalizes as an independent risk factor for breast cancer

Introduction

Breast density has been established as a major risk factor for breast cancer. We have previously demonstrated that mammographic texture resemblance (MTR), recognizing the local texture patterns of the mammogram, is also a risk factor for breast cancer, independent of percent breast density. We examine if these findings generalize to another population.

Methods

Texture patterns were recorded in digitalized pre-diagnosis (3.7 years) film mammograms of a nested case–control study within the Dutch screening program (S1) comprising of 245 breast cancers and 250 matched controls. The patterns were recognized in the same study using cross-validation to form resemblance scores associated with breast cancer. Texture patterns from S1 were examined in an independent nested case–control study within the Mayo Mammography Health Study cohort (S2) of 226 cases and 442 matched controls: mammograms on average 8.5 years prior to diagnosis, risk factor information and percent mammographic density (PD) estimated using Cumulus were available. MTR scores estimated from S1, S2 and S1 + S2 (the latter two as cross-validations) were evaluated in S2. MTR scores were analyzed as both quartiles and continuously for association with breast cancer using odds ratios (OR) and adjusting for known risk factors including age, body mass index (BMI), and hormone usage.

Results

The mean ages of S1 and S2 were 58.0 ± 5.7 years and 55.2 ± 10.5 years, respectively. The MTR scores on S1 showed significant capability to discriminate cancers from controls (area under the operator characteristics curve (AUC) = 0.63 ± 0.02, P <0.001), which persisted after adjustment for PD. S2 showed an AUC of 0.63, 0.61, and 0.60 based on PD, MTR scores trained on S2, and MTR scores trained on S1, respectively. When adjusted for PD, MTR scores of S2 trained on S1 showed an association with breast cancer for the highest quartile alone: OR in quartiles of controls as reference; 1.04 (0.59 to 1.81); 0.95 (0.52 to 1.74); 1.84 (1.10 to 3.07) respectively. The combined continuous model with both PD and MTR scores based on S1 had an AUC of 0.66 ± 0.03.

Conclusions

The local texture patterns associated with breast cancer risk in S1 were also an independent risk factor in S2. Additional textures identified in S2 did not significantly improve risk segregation. Hence, the textural patterns that indicated elevated risk persisted under differences in X-ray technology, population demographics, follow-up time and geography.

Quantitative analysis for breast density estimation in low dose chest CT scans

A computational method was developed for the measurement of breast density using chest computed tomography (CT) images and the correlation between that and mammographic density. Sixty-nine asymptomatic Asian women (138 breasts) were studied. With the marked lung area and pectoralis muscle line in a template slice, demons algorithm was applied to the consecutive CT slices for automatically generating the defined breast area. The breast area was then analyzed using fuzzy c-mean clustering to separate fibroglandular tissue from fat tissues. The fibroglandular clusters obtained from all CT slices were summed then divided by the summation of the total breast area to calculate the percent density for CT. The results were compared with the density estimated from mammographic images. For CT breast density, the coefficient of variations of intraoperator and interoperator measurement were 3.00 % (0.59 %-8.52 %) and 3.09 % (0.20 %-6.98 %), respectively. Breast density measured from CT (22 ± 0.6 %) was lower than that of mammography (34 ± 1.9 %) with Pearson correlation coefficient of r=0.88. The results suggested that breast density measured from chest CT images correlated well with that from mammography. Reproducible 3D information on breast density can be obtained with the proposed CT-based quantification methods.

A comparison of calibration data from full field digital mammography units for breast density measurements

Background

Breast density is a significant breast cancer risk factor measured from mammograms. The most appropriate method for measuring breast density for risk applications is still under investigation. Calibration standardizes mammograms to account for acquisition technique differences prior to making breast density measurements. We evaluated whether a calibration methodology developed for an indirect x-ray conversion full field digital mammography (FFDM) technology applies to direct x-ray conversion FFDM systems.

Methods

Breast tissue equivalent (BTE) phantom images were used to establish calibration datasets for three similar direct x-ray conversion FFDM systems. The calibration dataset for each unit is a function of the target/filter combination, x-ray tube voltage, current × time (mAs), phantom height, and two detector fields of view (FOVs). Methods were investigated to reduce the amount of calibration data by restricting the height, mAs, and FOV sampling. Calibration accuracy was evaluated with mixture phantoms. We also compared both intra- and inter-system calibration characteristics and accuracy.

Results

Calibration methods developed previously apply to direct x-ray conversion systems with modification. Calibration accuracy was largely within the acceptable range of ± 4 standardized units from the ideal value over the entire acquisition parameter space for the direct conversion units. Acceptable calibration accuracy was maintained with a cubic-spline height interpolation, representing a modification to previous work. Calibration data is unit specific, can be acquired with the large FOV, and requires a minimum of one reference mAs sample. The mAs sampling, calibration accuracy, and the necessity for machine specific calibration data are common characteristics and in agreement with our previous work.

Conclusion

The generality of our calibration approach was established under ideal conditions. Evaluation with patient data using breast cancer status as the endpoint is required to demonstrate that the approach produces a breast density measure associated with breast cancer.

Reader variability in breast density estimation from full-field digital mammograms: the effect of image postprocessing on relative and absolute measures

RATIONALE AND OBJECTIVES

Mammographic breast density, a strong risk factor for breast cancer, may be measured as either a relative percentage of dense (ie, radiopaque) breast tissue or as an absolute area from either raw (ie, "for processing") or vendor postprocessed (ie, "for presentation") digital mammograms. Given the increasing interest in the incorporation of mammographic density in breast cancer risk assessment, the purpose of this study is to determine the inherent reader variability in breast density assessment from raw and vendor-processed digital mammograms, because inconsistent estimates could to lead to misclassification of an individual woman's risk for breast cancer.

MATERIALS AND METHODS

Bilateral, mediolateral-oblique view, raw, and processed digital mammograms of 81 women were retrospectively collected for this study (N = 324 images). Mammographic percent density and absolute dense tissue area estimates for each image were obtained from two radiologists using a validated, interactive software tool.

RESULTS

The variability of interreader agreement was not found to be affected by the image presentation style (ie, raw or processed, F-test: P > .5). Interreader estimates of relative and absolute breast density are strongly correlated (Pearson r > 0.84, P < .001) but systematically different (t-test, P < .001) between the two readers.

CONCLUSION

Our results show that mammographic density may be assessed with equal reliability from either raw or vendor postprocessed images. Furthermore, our results suggest that the primary source of density variability comes from the subjectivity of the individual reader in assessing the absolute amount of dense tissue present in the breast, indicating the need to use standardized tools to mitigate this effect.

Comparison of percent density from raw and processed full-field digital mammography data

Introduction

Mammographic density has been established as a strong risk factor for breast cancer, primarily using digitized film mammograms. Full-field digital mammography (FFDM) is replacing film mammography, has different properties than film, and provides both raw and processed clinical display representation images. We evaluated and compared FFDM raw and processed breast density measures and their associations with breast cancer.

Methods

A case-control study of 180 cases and 180 controls matched by age, postmenopausal hormone use, and screening history was conducted. Mammograms were acquired from a General Electric Senographe 2000D FFDM unit. Percent density (PD) was assessed for each FFDM representation using the operator-assisted Cumulus method. Reproducibility within image type (n = 80) was assessed using Lin's concordance correlation coefficient (r_c). Correlation of PD between image representations (n = 360) was evaluated using Pearson's correlation coefficient (r) on the continuous measures and the weighted kappa statistic (κ) for quartiles. Conditional logistic regression was used to estimate odds ratios (ORs) for the PD and breast cancer associations for both image representations with 95% confidence intervals. The area under the receiver operating characteristic curve (AUC) was used to assess the discriminatory accuracy.

Results

Percent density from the two representations provided similar intra-reader reproducibility (r_c= 0.92 for raw and r_c= 0.87 for processed images) and was correlated (r = 0.82 and κ = 0.64). When controlling for body mass index, the associations of quartiles of PD with breast cancer and discriminatory accuracy were similar for the raw (OR: 1.0 (ref.), 2.6 (1.2 to 5.4), 3.1 (1.4 to 6.8), 4.7 (2.1 to 10.6); AUC = 0.63) and processed representations (OR: 1.0 (ref.), 2.2 (1.1 to 4.1), 2.2 (1.1 to 4.4), 3.1 (1.5 to 6.6); AUC = 0.64).

Conclusions

Percent density measured with an operator-assisted method from raw and processed FFDM images is reproducible and correlated. Both percent density measures provide similar associations with breast cancer.

Estimation of breast percent density in raw and processed full field digital mammography images via adaptive fuzzy c-means clustering and support vector machine segmentation

PURPOSE

The amount of fibroglandular tissue content in the breast as estimated mammographically, commonly referred to as breast percent density (PD%), is one of the most significant risk factors for developing breast cancer. Approaches to quantify breast density commonly focus on either semiautomated methods or visual assessment, both of which are highly subjective. Furthermore, most studies published to date investigating computer-aided assessment of breast PD% have been performed using digitized screen-film mammograms, while digital mammography is increasingly replacing screen-film mammography in breast cancer screening protocols. Digital mammography imaging generates two types of images for analysis, raw (i.e., "FOR PROCESSING") and vendor postprocessed (i.e., "FOR PRESENTATION"), of which postprocessed images are commonly used in clinical practice. Development of an algorithm which effectively estimates breast PD% in both raw and postprocessed digital mammography images would be beneficial in terms of direct clinical application and retrospective analysis.

METHODS

This work proposes a new algorithm for fully automated quantification of breast PD% based on adaptive multiclass fuzzy c-means (FCM) clustering and support vector machine (SVM) classification, optimized for the imaging characteristics of both raw and processed digital mammography images as well as for individual patient and image characteristics. Our algorithm first delineates the breast region within the mammogram via an automated thresholding scheme to identify background air followed by a straight line Hough transform to extract the pectoral muscle region. The algorithm then applies adaptive FCM clustering based on an optimal number of clusters derived from image properties of the specific mammogram to subdivide the breast into regions of similar gray-level intensity. Finally, a SVM classifier is trained to identify which clusters within the breast tissue are likely fibroglandular, which are then aggregated into a final dense tissue segmentation that is used to compute breast PD%. Our method is validated on a group of 81 women for whom bilateral, mediolateral oblique, raw and processed screening digital mammograms were available, and agreement is assessed with both continuous and categorical density estimates made by a trained breast-imaging radiologist.

RESULTS

Strong association between algorithm-estimated and radiologist-provided breast PD% was detected for both raw (r = 0.82, p < 0.001) and processed (r = 0.85, p < 0.001) digital mammograms on a per-breast basis. Stronger agreement was found when overall breast density was assessed on a per-woman basis for both raw (r = 0.85, p < 0.001) and processed (0.89, p < 0.001) mammograms. Strong agreement between categorical density estimates was also seen (weighted Cohen's κ ≥ 0.79). Repeated measures analysis of variance demonstrated no statistically significant differences between the PD% estimates (p > 0.1) due to either presentation of the image (raw vs processed) or method of PD% assessment (radiologist vs algorithm).

CONCLUSIONS

The proposed fully automated algorithm was successful in estimating breast percent density from both raw and processed digital mammographic images. Accurate assessment of a woman's breast density is critical in order for the estimate to be incorporated into risk assessment models. These results show promise for the clinical application of the algorithm in quantifying breast density in a repeatable manner, both at time of imaging as well as in retrospective studies.

A novel automated mammographic density measure and breast cancer risk

BACKGROUND

Mammographic breast density is a strong breast cancer risk factor but is not used in the clinical setting, partly because of a lack of standardization and automation. We developed an automated and objective measurement of the grayscale value variation within a mammogram, evaluated its association with breast cancer, and compared its performance with that of percent density (PD).

METHODS

Three clinic-based studies were included: a case-cohort study of 217 breast cancer case subjects and 2094 non-case subjects and two case-control studies comprising 928 case subjects and 1039 control subjects and 246 case subjects and 516 control subjects, respectively. Percent density was estimated from digitized mammograms using the computer-assisted Cumulus thresholding program, and variation was estimated from an automated algorithm. We estimated hazards ratios (HRs), odds ratios (ORs), the area under the receiver operating characteristic curve (AUC), and 95% confidence intervals (CIs) using Cox proportional hazards models for the cohort and logistic regression for case-control studies, with adjustment for age and body mass index. We performed a meta-analysis using random study effects to obtain pooled estimates of the associations between the two mammographic measures and breast cancer. All statistical tests were two-sided.

RESULTS

The variation measure was statistically significantly associated with the risk of breast cancer in all three studies (highest vs lowest quartile: HR = 2.0 [95% CI = 1.3 to 3.1]; OR = 2.7 [95% CI = 2.1 to 3.6]; OR = 2.4 [95% CI = 1.4 to 3.9]; [corrected] all P (trend) < .001). [corrected]. The risk estimates and AUCs for the variation measure were similar to [corrected] those for percent density (AUCs for variation = 0.60-0.62 and [corrected] AUCs for percent density = 0.61-0.65). [corrected]. A meta-analysis of the three studies demonstrated similar associations [corrected] between variation and breast cancer (highest vs lowest quartile: RR = 1.8, 95% CI = 1.4 to 2.3) and [corrected] percent density and breast cancer (highest vs lowest quartile: RR = 2.3, 95% CI = 1.9 to 2.9).

CONCLUSION

The association between the automated variation measure and the risk of breast cancer is at least as strong as that for percent density. Efforts to further evaluate and translate the variation measure to the clinical setting are warranted.

Full field digital mammography and breast density: comparison of calibrated and noncalibrated measurements

RATIONALE AND OBJECTIVES

Mammographic breast density is an important and widely accepted risk factor for breast cancer. A statement about breast density in the mammographic report is becoming a requirement in many States. However, there is significant inter-observer variation between radiologists in their interpretation of breast density. A properly designed automated system could provide benefits in maintaining consistency and reproducibility. We have developed a new automated and calibrated measure of breast density using full field digital mammography (FFDM). This new measure assesses spatial variation within a mammogram and produced significant associations with breast cancer in a small study. The costs of this automation are delays from advanced image and data analyses before the study can be processed. We evaluated this new calibrated variation measure using a larger dataset than previously. We also explored the possibility of developing an automated measure from unprocessed (raw data) mammograms as an approximation for this calibrated breast density measure.

MATERIALS AND METHODS

A case-control study comprised of 160 cases and 160 controls matched by age, screening history, and hormone replacement therapy was used to compare the calibrated variation measure of breast density with three variants of a noncalibrated measure of spatial variation. The operator-assisted percentage of breast density measure (PD) was used as a standard reference for comparison. Odds ratio (OR) quartile analysis was used to compare these measures. Linear regression analysis was applied to assess the calibration's impact on the raw pixel distribution.

RESULTS

All breast density measures showed significant breast cancer associations. The calibrated spatial variation measure produced the strongest associations (OR: 1.0 [ref.], 4.6, 4.3, 7.4). The associations for PD were diminished in comparison (OR: 1.0 [ref.], 2.7, 2.9, 5.2). Two additional non-calibrated measures restricted in region size also showed significant associations (OR: 1.0 [ref.], 2.9, 4.4, 5.4), and (OR: 1.0 [ref.], 3.5, 3.1, 4.9). Regression analyses indicated the raw image mean is influenced by the calibration more so than its standard deviation.

CONCLUSION

Breast density measures can be automated. The associated calibration produced risk information not retrievable from the raw data representation. Although the calibrated measure produced the stronger association, the non-calibrated measures may offer an alternative to PD and other operator based methods after further evaluation, because they can be implemented automatically with a simple processing algorithm.

Mammographic density and breast cancer risk: current understanding and future prospects

Variations in percent mammographic density (PMD) reflect variations in the amounts of collagen and number of epithelial and non-epithelial cells in the breast. Extensive PMD is associated with a markedly increased risk of invasive breast cancer. The PMD phenotype is important in the context of breast cancer prevention because extensive PMD is common in the population, is strongly associated with risk of the disease, and, unlike most breast cancer risk factors, can be changed. Work now in progress makes it likely that measurement of PMD will be improved in the near future and that understanding of the genetics and biological basis of the association of PMD with breast cancer risk will also improve. Future prospects for the application of PMD include mammographic screening, risk prediction in individuals, breast cancer prevention research, and clinical decision making.

Calibrated measures for breast density estimation

RATIONALE AND OBJECTIVES

Breast density is a significant breast cancer risk factor measured from mammograms. Evidence suggests that the spatial variation in mammograms may also be associated with risk. We investigated the variation in calibrated mammograms as a breast cancer risk factor and explored its relationship with other measures of breast density using full field digital mammography (FFDM).

MATERIALS AND METHODS

A matched case-control analysis was used to assess a spatial variation breast density measure in calibrated FFDM images, normalized for the image acquisition technique variation. Three measures of breast density were compared between cases and controls: (a) the calibrated average measure, (b) the calibrated variation measure, and (c) the standard percentage of breast density (PD) measure derived from operator-assisted labeling. Linear correlation and statistical relationships between these three breast density measures were also investigated.

RESULTS

Risk estimates associated with the lowest to highest quartiles for the calibrated variation measure were greater in magnitude (odds ratios: 1.0 [ref.], 3.5, 6.3, and 11.3) than the corresponding risk estimates for quartiles of the standard PD measure (odds ratios: 1.0 [ref.], 2.3, 5.6, and 6.5) and the calibrated average measure (odds ratios: 1.0 [ref.], 2.4, 2.3, and 4.4). The three breast density measures were highly correlated, showed an inverse relationship with breast area, and related by a mixed distribution relationship.

CONCLUSION

The three measures of breast density capture different attributes of the same data field. These preliminary findings indicate the variation measure is a viable automated method for assessing breast density. Insights gained by this work may be used to develop a standard for measuring breast density.