Mediolateral Oblique Mammograms Research Articles

Pectoral Muscle Segmentation for Digital Mammograms Based on Otsu Thresholding

The appearance of pectoral muscle in medio-lateral oblique (MLO) views of mammograms can increase the false positive in computer aided detection (CAD) of breast cancer detection. Pectoral muscle has to be identified and segmented from the breast region in a mammogram before further analysis. The main goal of this paper is to propose an accurate and efficient algorithm of pectoral muscle extraction on MLO mammograms. The proposed algorithm bases on the positional characteristic of pectoral muscle in a breast region to combine the iterative Otsu thresholding scheme and the mathematic morphological processing to find the rough border of the pectoral muscle. The multiple regression analysis is then employed on the rough border to obtain the accurate segmentation of the pectoral muscle. The presented algorithm is tested on the digital mammograms from the Mammogram Image Analysis Society (MIAS) database. The experimental results show that the pectoral muscle extracted by the presented algorithm approximately follows that extracted by an expert radiologist.

An Anatomically Oriented Breast Coordinate System for Mammogram Analysis

We have developed a breast coordinate system that is based on breast anatomy to register female breasts into a common coordinate frame in 2-D mediolateral (ML) or mediolateral oblique (MLO) view mammograms. The breasts are registered according to the location of the pectoral muscle and the nipple and the shape of the breast boundary because these are the most robust features independent of the breast size and shape. On the basis of these landmarks, we have constructed a nonlinear mapping between the parameter frame and the breast region in the mammogram. This mapping makes it possible to identify the corresponding positions and orientations among all of the ML or MLO mammograms, which facilitates an implicit use of the registration, i.e., no explicit image warping is needed. We additionally show how the coordinate transform can be used to extract Gaussian derivative features so that the feature positions and orientations are registered and extracted without nonlinearly deforming the images. We use the proposed breast coordinate transform in a cross-sectional breast cancer risk assessment study of 490 women, in which we attempt to learn breast cancer risk factors from mammograms that were taken prior to when the breast cancer became visible to a radiologist. The coordinate system provides both the relative position and orientation information on the breast region from which the features are derived. In addition, the coordinate system can be used in temporal studies to pinpoint anatomically equivalent locations between the mammograms of each woman and among the mammograms of all of the women in the study. The results of the cross-sectional study show that the classification into cancer and control groups can be improved by using the new coordinate system, compared to other systems evaluated. Comparisons were performed using the area-under-the-receiver-operating-characteristic-curve score. In general, the new coordinate system makes an accurate anatomical registration of breasts possible, which suggests its wide applicability wherever 2-D mammogram registration is required.

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.

Die diagnostische Bildqualität von Mammografien in der vertragsärztlichen Versorgung Deutschlands

PURPOSE: A total of 79 115 mammograms from statutory health insurance (SHI) physicians within German outpatient care were evaluated with respect to the diagnostic image quality. MATERIALS AND METHODS: Mammograms were randomly selected between 2006 and 2008 by the regional Associations of Statutory Health Insurance Physicians and submitted to regional boards of experts for external evaluation. The mammogram quality was evaluated using a 3-point scale (adequate, borderline, failure) and documented using a nationally standardized protocol. RESULTS: 87.6 % of the mammograms were classified as adequate, 11.0 % as borderline and 1.4 % as failure. Mediolateral oblique mammograms (mlo) had worse ratings than craniocaudal mammograms (cc). Main reasons for classifying the mammograms as borderline or failure were “inframammary fold not adequately visualized” (mlo), “pectoral muscle not in the correct angle or not to the level with the nipple” (mlo), “the nipple not in profile” (mlo, cc) and “breast not completely or not adequately visualized” (cc). CONCLUSION: The results show a good overall quality of mammograms in German outpatient medical care. Failures can be associated predominantly with incorrect positioning of the breast. More precisely defined quality criteria using objective measures are recommended, especially for craniocaudal mammograms (cc).

Using mammographic density to predict breast cancer risk: dense area or percentage dense area

IntroductionMammographic density (MD) is one of the strongest risk factors for breast cancer. It is not clear whether this association is best expressed in terms of absolute dense area or percentage dense area (PDA).MethodsWe measured MD, including nondense area (here a surrogate for weight), in the mediolateral oblique (MLO) mammogram using a computer-assisted thresholding technique for 634 cases and 1,880 age-matched controls from the Cambridge and Norwich Breast Screening programs. Conditional logistic regression was used to estimate the risk of breast cancer, and fits of the models were compared using likelihood ratio tests and the Bayesian information criteria (BIC). All P values were two-sided.ResultsSquare-root dense area was the best single predictor (for example, χ12 = 53.2 versus 44.4 for PDA). Addition of PDA and/or square-root nondense area did not improve the fit (both P > 0.3). Addition of nondense area improved the fit of the model with PDA (χ12 = 11.6; P < 0.001). According to the BIC, the PDA and nondense area model did not provide a better fit than the dense area alone model. The fitted values of the two models were highly correlated (r = 0.97). When a measure of body size is included with PDA, the predicted risk is almost identical to that from fitting dense area alone.ConclusionsAs a single parameter, dense area provides more information than PDA on breast cancer risk.

Die diagnostische Bildqualität von Mammografien in der vertragsärztlichen Versorgung Deutschlands

A total of 79115 mammograms from statutory health insurance (SHI) physicians within German outpatient care were evaluated with respect to the diagnostic image quality. Mammograms were randomly selected between 2006 and 2008 by the regional Associations of Statutory Health Insurance Physicians and submitted to regional boards of experts for external evaluation. The mammogram quality was evaluated using a 3-point scale (adequate, borderline, failure) and documented using a nationally standardized protocol. 87.6% of the mammograms were classified as adequate, 11.0% as borderline and 1.4% as failure. Mediolateral oblique mammograms (mlo) had worse ratings than craniocaudal mammograms (cc). Main reasons for classifying the mammograms as borderline or failure were "inframammary fold not adequately visualized" (mlo), "pectoral muscle not in the correct angle or not to the level with the nipple" (mlo), "the nipple not in profile" (mlo, cc) and "breast not completely or not adequately visualized" (cc). The results show a good overall quality of mammograms in German outpatient medical care. Failures can be associated predominantly with incorrect positioning of the breast. More precisely defined quality criteria using objective measures are recommended, especially for craniocaudal mammograms (cc).

Mammography image quality: Model for predicting compliance with posterior nipple line criterion

To develop a model using measurements of pectoral muscle width and length together with the acceptability of the posterior nipple line criteria (PNL) to predict the acceptability of the presentation of the pectoral muscle in the mediolateral oblique view of the breast. A total of 400 mediolateral oblique mammogram images were randomly selected from BreastScreen NSW South West, Australia. Measurements of length and width of the pectoral muscle and the acceptability of the pectoral muscle position relative to the PNL were recorded. Data analysis involved logistic regression and ROC analysis to test the predictors of width and length and the performance of the model. The model was then used to predict the outcome of acceptable or unacceptable PNL criterion for each case. The estimated odds ratio for an increase of 10mm was 1.98 (CI=1.68, 2.34) for the length predictor and 2.14 (CI=1.56, 2.93) for the width predictor. A cut off point of 0.6083 was derived from the training set and applied with the developed model to the test set. The area under the ROC curve was 0.9339 demonstrating an accurate model. This paper describes a model to predict the acceptability of the PNL criterion using the width and length of the pectoral muscle. This model could be used in the automated assessment of image quality which has the potential to enhance the consistency in mammographic image quality evaluation. Optimising image quality contributes to increased accuracy in radiological interpretation, which maximises the early detection of breast cancer and potentially reduces mortality rates.

Mammography: Correlation of pectoral muscle width and the length in the mediolateral oblique view of the breast

Purpose To investigate the relationship between the width and length of the pectoral muscle in the mediolateral oblique mammogram of the breast. Method Mammograms reviewed for this study were performed on women undergoing routine mammographic screening with BreastScreen NSW South West, Australia. The study included a total of 2800 randomly selected mediolateral oblique mammograms from examinations taken between July 2004 and September 2006. Measurements of the width and length of the pectoral muscle were recorded. Results No statistically significant difference was demonstrated between the mean values for right and left width ( p = 0.5293) or length ( p = 0.2079). Matched pair analysis demonstrated a statistically significant difference width, ( p = 0.0069) and length, ( p = 0.0369). No relationship could be demonstrated between the width and length of the pectoral muscle, ( p = 0.0701). Conclusion The inability to determine a relationship between the width and length of the pectoral muscle suggests that these criteria should be assessed independently of each other. This new understanding of the presentation of the pectoral muscle will enable clinically relevant redevelopment of current image evaluation systems to include statistically supported mean performance values. Further research is needed to explore the relationship of width and length to other key image evaluation criteria used in mammography image quality evaluation.

Predicting breast cancer risk using mammographic density measurements from both mammogram sides and views

Mammographic density is a strong risk factor for breast cancer. Which and how many x-rays are used for research, and how mammographic density is measured varies across studies. In this article, we compared three different measurements (absolute dense area, percent dense area and percent dense volume) from each of four mammograms [left, right, medio-lateral oblique (MLO) and cranio-caudal (CC) views] using three different methods of measurement [computer-assisted thresholding, visual assessment and standard mammogram form (SMF)] to investigate whether additional measurements and/or different methods of measurement provide more information in the prediction of breast cancer risk. Mammographic density was measured in all four mammograms from 318 cases and 899 age-matched controls combined from the Cambridge and Norwich Breast Screening Programmes. Measurements were averaged across various combinations of mammogram type and/or view. Conditional logistic regression was used to estimate odds ratios associated with increasing quintiles of each mammographic measure. Overall, there appeared to be no difference in the fit of the models using two or four mammograms compared to the models using just the contralateral MLO or CC mammogram (all P > 0.07) for all methods of measurement. Common practice of measuring just the contralateral MLO or CC mammogram for analysis in case-control studies investigating the association between mammographic density and breast cancer risk appears to be sufficient.

Influence of Using Manual or Automatic Breast Density Information in a Mass Detection CAD System

The goal of this article is to analyze and compare the performance of a developed mass computer-aided detection (CAD) system that takes breast density information into account when using manual or automatic breast density annotations in the training step. The advantages of considering this breast density information will be highlighted. The image database used in this article is 92 mediolateral oblique (MLO) and 92 craniocaudal (CC) mammograms obtained by a full-field digital mammographic unit. All mammograms contain at least one mass. The evaluation of the experiments is performed using free receiver operating characteristic analysis for evaluating the detection performance and pixel-based receiver operating characteristic analysis for evaluating the segmentation accuracy. In addition, the performance of the automatic breast density classifier is shown using confusion matrices. When the breast density information is not considered and at a specificity of two false positives per image, the sensitivity obtained by the CAD system is 0.747 for the CC views and 0.853 for the MLO views. Considering the breast density information, the sensitivity for CC and MLO mammograms increases to 0.800 and 0.893, respectively, using manual classification, and 0.827 and 0.907, respectively, using automatic estimation. The same trend is observed when evaluating the CAD segmentation accuracy for detected masses in terms of area under the curve values: without considering breast density, these are 0.920 +/- 0.057 and 0.917 +/- 0.072; using manual classification, 0.934 +/- 0.039 and 0.932 +/- 0.046; and using automatic estimation, 0.947 +/- 0.038 and 0.946 +/- 0.045 for CC and MLO views, respectively. The experiments showed improved results when breast density information was taken into account. Moreover, the results obtained when using automatic breast density estimation outperformed those based on the manual annotations provided by expert radiologists. In this sense, the experiments showed that breast density information can be beneficial for CAD systems, and this information can be estimated robustly by an automatic procedure, which reduces the inter- and intra-class variability of the radiologists.

Mammographic CAD: Correlation of regions in ipsilateral views - a pilot study

Background: Radiologists analyse both standard mammographic views of a breast to confirm the presence of abnormalities and reduce false-positives. However, at present no computer-aided diagnosis system uses ipsilateral mammograms to confirm the presence of suspicious features. &#x0D; Aim: The aim of this study was to develop image-processing algorithms that can be used to match a suspicious feature from one mammographic view to the same feature in another mammographic view of the same breast. This algorithm can be incorporated into a computer-aided diagnosis package to confirm the presence of suspicious features.&#x0D; Method: The algorithms were applied to 68 matched pairs of cranio-caudal and mediolateral-oblique mammograms. The results of this pilot study take the form of maps of similarity. A novel method of evaluating the similarity maps is presented, using the area under the receiver operating characteristic curve (AUC) and the contrast (C) between the area of the matched region and the background of the similarity map. &#x0D; Results and Conclusions: The first matching algorithm (using texture measures extracted from a grey-level co-occurrence matrix (GLCM) and a Euclidean distance similarity metric) achieved an average AUC=0.80±0.17 with an average C=0.46±0.26. The second algorithm (using GLCMs and a mutual information similarity metric) achieved an average AUC=0.77±0.25 with an average C=0.50±0.42. The latter algorithm also performed remarkably well with the matching of malignant masses and achieved an average AUC=0.96±0.05 with an average C=0.90±0.21.

Quality assurance of mammograms in the Norwegian Breast Cancer Screening Program

This study compared the proportion of mammograms classified as perfect, good, moderately good or inadequate by a radiographer specially trained for doing such a classification at a breast centre (local-PGMI radiographer) with the proportion similarly assessed by an expert-PGMI radiographer. The results were compared with the recommendations given in the quality assurance manual of the Norwegian Breast Cancer Screening Programme. The reasons for classifying the mammograms into other than perfect, such as good, moderately good or inadequate were investigated. The quality of the mammograms was measured by using the PGMI system, which is a quality-review model that classifies the images into the four categories according to positioning, compression, exposure, noise, artefacts, and movement. A total of 1280 mammograms from all 16 breast centres in the screening programme were classified. The distribution of perfect, good, moderately good, and inadequate mammograms differed between the local-PGMI radiographers and the expert radiographer, for both the cranio-caudal (CC) and mediolateral–oblique (MLO) mammograms ( P < 0.001 for both). The expert radiographer classified a higher proportion of both CC (28%) and MLO (14%) mammograms as inadequate than did the local-PGMI radiographers (7% and 3%, respectively; P < 0.001 for both). The guidelines recommend ≤3% of the mammograms to be inadequate. The reason given for the inadequate classifications by the expert radiographer was predominantly “parts of the breast missing” for both the CC and the MLO mammograms. There is room to improve the quality of the mammograms in the screening programme in Norway. Attention should be given to positioning and the use of standardized terms in the PGMI classification.

Abdominal Carcinomatosis Attributed to Metastatic Breast Carcinoma

A 41-year-old female suffered from epigastralgia and intermittent constipation for 10 months, and abdominal fullness and intermittent pain for 6 months, before seeking help. Double contrast barium study of the colon showed multiple indentations on the sigmoid, ascending, and proximal transverse portions with tethered adjacent mucosal outline as well as the presence of ascites compatible with peritoneal carcinomatosis. Mediolateral oblique mammogram showed a speculated mass with some intratumoral microcalcifications in the upper retroareolar portion of the right breast. Due to the persistent abdominal complaints, laparotomy was done. Breast lump biopsy was done simultaneously. On opening abdominal cavity, massive yellowish ascites was noted. Diffuse small nodules over omentum and mesentery retraction were found. Bilateral ovarian masses were also noted. Right oophorectomy and omentectomy were performed. Histologic findings and results of immunohistochemical stains were consistent with diagnosis of primary breast cancer with metastasis to ovary and omentum.

Volumetric breast density evaluation from ultrasound tomography images

Previous ultrasound tomography work conducted by our group showed a direct correlation between measured sound speed and physical density in vitro, and increased in vivo sound speed with increasing mammographic density, a known risk factor for breast cancer. Building on these empirical results, the purpose of this work was to explore a metric to quantify breast density using our ultrasound tomography sound speed images in a manner analogous to computer-assisted mammogram segmentation for breast density analysis. Therefore, volumetric ultrasound percent density (USPD) is determined by segmenting high sound speed areas from each tomogram using a k-means clustering routine, integrating these results over the entire volume of the breast, and dividing by whole-breast volume. First, a breast phantom comprised of fat inclusions embedded in fibroglandular tissue was scanned four times with both our ultrasound tomography clinical prototype (with 4 mm spatial resolution) and CT. The coronal transmission tomograms and CT images were analyzed using semiautomatic segmentation routines, and the integrated areas of the phantom's fat inclusions were compared between the four repeated scans. The average variability for inclusion segmentation was approximately 7% and approximately2%, respectively, and a close correlation was observed in the integrated areas between the two modalities. Next, a cohort of 93 patients was imaged, yielding volumetric coverage of the breast (45-75 sound speed tomograms/patient). The association of USPD with mammographic percent density (MPD) was evaluated using two measures: (1) qualitative, as determined by a radiologist's visual assessment using BI-RADS Criteria and (2) quantitative, via digitization and semiautomatic segmentation of craniocaudal and mediolateral oblique mammograms. A strong positive association between BI-RADS category and USPD was demonstrated [Spearman rho = 0.69 (p < 0.001)], with significant differences between all BI-RADS categories as assessed by one-way ANOVA and Scheffé posthoc analysis. Furthermore, comparing USPD to calculated mammographic density yielded moderate to strong positive associations for CC and MLO views (r2 = 0.75 and 0.59, respectively). These results support the hypothesis that utilizing USPD as an analogue to mammographic breast density is feasible, providing a nonionizing, whole-breast analysis.

Breast cancer risk factors and a novel measure of volumetric breast density: cross-sectional study

We conducted a cross-sectional study nested within a prospective cohort of breast cancer risk factors and two novel measures of breast density volume among 590 women who had attended Glasgow University (1948–1968), replied to a postal questionnaire (2001) and attended breast screening in Scotland (1989–2002). Volumetric breast density was estimated using a fully automated computer programme applied to digitised film-screen mammograms, from medio-lateral oblique mammograms at the first-screening visit. This measured the proportion of the breast volume composed of dense (non-fatty) tissue (Standard Mammogram Form (SMF)%) and the absolute volume of this tissue (SMF volume, cm3). Median age at first screening was 54.1 years (range: 40.0–71.5), median SMF volume 70.25 cm3 (interquartile range: 51.0–103.0) and mean SMF% 26.3%, s.d.=8.0% (range: 12.7–58.8%). Age-adjusted logistic regression models showed a positive relationship between age at last menstrual period and SMF%, odds ratio (OR) per year later: 1.05 (95% confidence interval: 1.01–1.08, P=0.004). Number of pregnancies was inversely related to SMF volume, OR per extra pregnancy: 0.78 (0.70–0.86, P<0.001). There was a suggestion of a quadratic relationship between birthweight and SMF%, with lowest risks in women born under 2.5 and over 4 kg. Body mass index (BMI) at university (median age 19) and in 2001 (median age 62) were positively related to SMF volume, OR per extra kg m−2 1.21 (1.15–1.28) and 1.17 (1.09–1.26), respectively, and inversely related to SMF%, OR per extra kg m−2 0.83 (0.79–0.88) and 0.82 (0.76–0.88), respectively, P<0.001. Standard Mammogram Form% and absolute SMF volume are related to several, but not all, breast cancer risk factors. In particular, the positive relationship between BMI and SMF volume suggests that volume of dense breast tissue will be a useful marker in breast cancer studies.

A new method for quantitative analysis of mammographic density

Women with mammographic percent density >50% have a approximately three-fold increased risk of developing breast cancer, potentially making them screening candidates for breast MRI scanning. The purpose of this work is to introduce a new method to quantify mammographic percent density (MPD), and to compare the results with the current standard of care for breast density assessment. Craniocaudal (CC) and mediolateral oblique (MLO) mammograms for 104 patients were digitized and analyzed using an interactive computer-assisted segmentation routine implemented for two purposes: (1) to segment the breast area from background and radiographic markers, and (2) to segment dense from fatty portions of the breast. Our technique was evaluated by comparing the results to qualitative estimates determined by a certified breast radiologist using the BI-RADS Categorical Assessment (1 (fatty) to 4 (dense) scale). Statistically significant correlations (two-tailed, p < 0.01) were observed between calculated MPD and BI-RADS for both CC (Spearman rho = 0.67) and MLO views (Spearman rho = 0.71). For the CC view, statistically significant differences were revealed between the mean MPD for each BI-RADS category except between fatty (BI-RADS 1) and scattered (BI-RADS 2). Finally, for the MLO views, statistically significant differences in the mean MPD between all BI-RADS categories were observed. Comparing the CC and MLO views revealed a strong positive correlation (Pearson r = 0.8) in calculated MPD. In addition, an evaluation of the reproducibility of our segmentation demonstrated the average standard deviation of MPD for a subsample of eight patients, measured five times, was 1.9% (range: 0.03%-9.9%). Eliminating one misassignment reduced the average standard deviation to 0.75% (range: 0.03%-3.16%). Further analysis of approximately 10% of the patient sample revealed strong agreement (ICC = 0.80-0.85) in the reliability of MPD estimates for both mammographic views. Overall, these results demonstrate the feasibility of utilizing our approach for quantitative breast density segmentation, which may be useful for detecting small changes in MPD introduced through chemoprevention, diet, or other interventions.

TH-E-L100J-07: Comparison of Mammographic Percent Density and Volumetric Percent Density Determined From Ultrasound Tomography Images

Purpose: Previous ultrasound tomography work conducted by our group showed a direct correlation between measured sound speed and physical density in vitro, and increased in vivo sound speed with increasing mammographic density, a known risk factor for breast cancer. Building on these empirical results, we investigated the use of volumetric ultrasound percent density (USPD) for breast density estimation. Method and Materials: A breast phantom comprised of fat inclusions embedded in fibroglandular tissue was scanned four times with both ultrasound tomography and CT. The coronal transmission tomograms and corresponding CT images were analyzed using a semi-automatic segmentation routine. Next, a cohort of ∼100 patients was imaged encompassing the entire breast volume (50–75 tomograms/patient). USPD was determined by segmenting high sound speed areas from each slice using a k-means clustering routine, integrating these results over the entire breast, and dividing by total breast area. USPD was evaluated using two mammographic density measures: (1) qualitative, as determined by a radiologist's visual assessment using BI-RADS Criteria and (2) quantitative, via digitization and semi-automatic segmentation of craniocaudal and medio-lateral oblique mammograms. Results: The integrated areas of the phantom's fat inclusions were compared using both transmission ultrasound and CT for four repeated scans. The average variability for inclusion segmentation was ∼2% and ∼10%, respectively, and a close correlation was observed in the integrated areas between the two modalities. A strong positive association between BI-RADS category and USPD was demonstrated. Furthermore, comparing USPD to calculated mammographic density yielded moderate to strong positive associations (Pearson r = 0.76–0.84) for MLO and CC views, respectively. Conclusion: These results support the hypothesis that utilizing USPD as an analogue to mammographic breast density is feasible. USPD has the potential to provide a non-ionizing, whole-breast analysis of breast density, which may better elucidate the relationship between breast density and breast cancer risk.

TH‐D‐M100J‐06: Evaluation of Breast Density Using Sound Speed Measurements

Purpose: Women with high mammographic breast density are at a 4‐ to 6‐fold increased risk of developing breast cancer compared to women with fatty breasts. The purpose of this investigation was to confirm preliminary acoustic velocity measurements for breast density assessment, and to evaluate them with the current standard of care. Method and Materials: A sample of ∼100 patients was imaged with our computed ultrasound tomography clinical prototype with each dataset comprised of 45–75 tomograms ranging from near the chest wall through the nipple region. Whole breast acoustic velocity was determined by creating image stacks and evaluating the sound speed frequency distribution. The acoustic measures of breast density were evaluated by comparing these results to two mammographic density measures: (1) qualitative, as determined by a certified radiologist using the BI‐RADS Categorical Assessment based on a 1 (fatty) to 4 (dense) scale, and (2) quantitative, via digitization and computerized analysis of archival mammograms. The former involved a radiologist's visual assessment of each mammogram, while the latter required digitizing cranio‐caudal (CC) and mediolateral oblique (MLO) mammograms and implementing semi‐automatic segmentation routines. Results: Approximately 140 m/s difference in acoustic velocity was found between the fatty and dense BI‐RADS Categories, with statistically significant differences in mean sound speed demonstrated between each category (ANOVA, α= 0.05). In addition, an increased sound speed was exhibited with increased BI‐RADS Category. Furthermore, moderately strong positive associations were demonstrated between mean acoustic velocity and quantitative measures of mammographic percent breast density for both MLO and CC views (r2 = 0.65–0.76), respectively. Conclusion: These results confirm the hypothesis that utilizing acoustical velocity can be used as an indicator of breast tissue density. Having an operator independent and non‐ionizing means of whole‐breast density evaluation has important clinical implications which include tracking temporal changes, identifying high‐risk populations, and evaluating treatment response.

Mammography image quality: Analysis of evaluation criteria using pectoral muscle presentation

Purpose To investigate the mammographic presentation of the pectoral muscle on the mediolateral oblique view of the breast and its relationship to current image evaluation criteria. Methods A retrospective review of 100 paired mediolateral oblique mammograms performed on women undergoing routine mammographic screening with BreastScreen NSW South West, Australia was undertaken. Measurements of length, width, contour type and inferior angle of the pectoral muscle and of the relationship of its inferior aspect to the posterior nipple line and the nipple level were recorded. Results A wide range of parameter values representing acceptable imaging were found to exist for each of the key imaging criteria. In 75.5% of images reviewed the inferior aspect of the pectoral muscle did not reach the level of the nipple, failing the quality criterion of the image evaluation systems used by the UK National Health Service Breast Screening Programme and the Australian BreastScreen Program. Conclusion It is recommended that aspects of current image evaluation criteria relating to the presentation of the pectoral muscle should be modified to include quantified ranges to reduce inherent subjectivity and variability in interpretation. The criterion for the inferior aspect of the pectoral muscle to reach the level of the nipple should be reviewed by the national accreditation bodies.

Segmentation of the fibro-glandular disc in mammograms using Gaussian mixture modelling.

The paper presents a technique for the segmentation of the fibro-glandular disc in mammograms based upon a statistical model of breast density. The density function of the model was represented by a mixture of up to four weighted Gaussians, each one corresponding to a specific density class in the breast. The parameters of the model and the number of tissue classes in the breast were determined using the expectation-maximisation algorithm and the minimum description length method. Grey-level statistics of the pectoral muscle were used to determine the tissue categories that are likely to represent the fibro-glandular disc. The method was applied to 84 medio-lateral oblique mammograms from the Mini-MIAS database. The results of the segmented fibro-glandular disc were assessed by a radiologist using the original and the segmented images, with reference to a ranking table categorising the results of segmentation as: 1: excellent; 2: good; 3: average; 4: poor; and 5: complete failure. Of the 84 cases analysed, 64.3% were rated as excellent, 16.7% were rated as good, 10.7% were rated as average, and 4.7% were rated as poor; only 3.6% of the cases were rated as a complete failure with regard to segmentation of the fibro-glandular disc.