Richard H. Moore

Enhanced imaging of microcalcifications in digital breast tomosynthesis through improved image‐reconstruction algorithms

PURPOSEnThe authors develop a practical, iterative algorithm for image-reconstruction in undersampled tomographic systems, such as digital breast tomosynthesis (DBT).nnnMETHODSnThe algorithm controls image regularity by minimizing the image total p variation (TpV), a function that reduces to the total variation when p = 1.0 or the image roughness when p = 2.0. Constraints on the image, such as image positivity and estimated projection-data tolerance, are enforced by projection onto convex sets. The fact that the tomographic system is undersampled translates to the mathematical property that many widely varied resultant volumes may correspond to a given data tolerance. Thus the application of image regularity serves two purposes: (1) Reduction in the number of resultant volumes out of those allowed by fixing the data tolerance, finding the minimum image TpV for fixed data tolerance, and (2) traditional regularization, sacrificing data fidelity for higher image regularity. The present algorithm allows for this dual role of image regularity in undersampled tomography.nnnRESULTSnThe proposed image-reconstruction algorithm is applied to three clinical DBT data sets. The DBT cases include one with microcalcifications and two with masses.nnnCONCLUSIONSnResults indicate that there may be a substantial advantage in using the present image-reconstruction algorithm for microcalcification imaging.

Combined Optical and X-ray Tomosynthesis Breast Imaging

PURPOSEnTo explore the optical and physiologic properties of normal and lesion-bearing breasts by using a combined optical and digital breast tomosynthesis (DBT) imaging system.nnnMATERIALS AND METHODSnInstitutional review board approval and patient informed consent were obtained for this HIPAA-compliant study. Combined optical and tomosynthesis imaging analysis was performed in 189 breasts from 125 subjects (mean age, 56 years ± 13 [standard deviation]), including 138 breasts with negative findings and 51 breasts with lesions. Three-dimensional (3D) maps of total hemoglobin concentration (Hb(T)), oxygen saturation (So(2)), and tissue reduced scattering coefficients were interpreted by using the coregistered DBT images. Paired and unpaired t tests were performed between various tissue types to identify significant differences.nnnRESULTSnThe estimated average bulk Hb(T) from 138 normal breasts was 19.2 μmol/L. The corresponding mean So(2) was 0.73, within the range of values in the literature. A linear correlation (R = 0.57, P < .0001) was found between Hb(T) and the fibroglandular volume fraction derived from the 3D DBT scans. Optical reconstructions of normal breasts revealed structures corresponding to chest-wall muscle, fibroglandular, and adipose tissues in the Hb(T), So(2), and scattering images. In 26 malignant tumors of 0.6-2.5 cm in size, Hb(T) was significantly greater than that in the fibroglandular tissue of the same breast (P = .0062). Solid benign lesions (n = 17) and cysts (n = 8) had significantly lower Hb(T) contrast than did the malignant lesions (P = .025 and P = .0033, respectively).nnnCONCLUSIONnThe optical and DBT images were structurally consistent. The malignant tumors and benign lesions demonstrated different Hb(T) and scattering contrasts, which can potentially be exploited to reduce the false-positive rate of conventional mammography and unnecessary biopsies.

Combined Optical Imaging and Mammography of the Healthy Breast: Optical Contrast Derived From Breast Structure and Compression

In this paper, we report new progress in developing the instrument and software platform of a combined X-ray mammography/diffuse optical breast imaging system. Particularly, we focus on system validation using a series of balloon phantom experiments and the optical image analysis of 49 healthy patients. Using the finite-element method for forward modeling and a regularized Gauss-Newton method for parameter reconstruction, we recovered the inclusions inside the phantom and the hemoglobin images of the human breasts. An enhanced coupling coefficient estimation scheme was also incorporated to improve the accuracy and robustness of the reconstructions. The recovered average total hemoglobin concentration (HbT) and oxygen saturation (SO₂) from 68 breast measurements are 16.2 mum and 71%, respectively, where the HbT presents a linear trend with breast density. The low HbT value compared to literature is likely due to the associated mammographic compression. From the spatially co-registered optical/X-ray images, we can identify the chest-wall muscle, fatty tissue, and fibroglandular regions with an average HbT of 20.1plusmn6.1 &nbsp;mum for fibroglandular tissue, 15.4plusmn5.0&nbsp;mum for adipose, and 22.2plusmn7.3&nbsp;mum for muscle tissue. The differences between fibroglandular tissue and the orresponding adipose tissue are significant. At the same time, we recognize that the optical images are influenced, to a certain extent, by mammographical compression. The optical images from a subset of patients show composite features from both tissue structure and pressure distribution. We present mechanical simulations which further confirm this hypothesis.

Mammographic Breast Density and Race

OBJECTIVEnWomen with increased mammographic breast density are known to be at higher risk of developing breast cancer. Reports of differences in breast density by race have implied that genetic and environmental factors may in part determine breast density. We first compared breast density among white, African American, and Asian women and then correlated breast density and race with age, body mass index (BMI), and breast or cup size.nnnMATERIALS AND METHODSnA retrospective review of data collected from 15,292 women was conducted. A stepwise multiple regression for an ordered response (breast density) was used to test for a relationship between race or ethnicity and breast density. We then determined whether differences in breast density by race might be caused by differences among races and ethnic groups in the age at imaging and BMI. We informally assessed the strength of the contribution of each term by means of the incremental change in the percent concordance. We also compared models using bra and cup sizes and age with models using BMI and age to try to determine whether the effects of breast size are local or systemic.nnnRESULTSnWe did not find evidence that mammographic breast density differences exist across racial groups (p < 0.0001) other than those associated with BMI and age at screening. Ignoring age and BMI, breast density depends on race for all comparisons (p < 0.0001). To generalize, we found that breast density appears to be greater in Asian women and least in African American women. However, when controlling for BMI and age, breast density differences by race disappeared in all groups except Asians (p < 0.0001). In all racial groups, bra and cup size in addition to age correlated with breast density after controlling for BMI (p < 0.0001). Except in Asian women, in women of any racial group, age and any of the following parameters accounted for all of the breast density differences: BMI, bra size, and cup size.nnnCONCLUSIONnAlthough breast density is associated with breast cancer risk, our results indicate that innate mammographic breast density differences across racial groups do not explain the risk differences known for the development of breast cancer. Age and BMI or age, bra size, and cup size can account for the reported density differences except among Asians. There may be no innate racial differences in breast density beyond those associated with racial differences in age and body habitus.

Computer‐aided detection of masses in digital tomosynthesis mammography: Comparison of three approaches

The authors are developing a computer-aided detection (CAD) system for masses on digital breast tomosynthesis mammograms (DBT). Three approaches were evaluated in this study. In the first approach, mass candidate identification and feature analysis are performed in the reconstructed three-dimensional (3D) DBT volume. A mass likelihood score is estimated for each mass candidate using a linear discriminant analysis (LDA) classifier. Mass detection is determined by a decision threshold applied to the mass likelihood score. A free response receiver operating characteristic (FROC) curve that describes the detection sensitivity as a function of the number of false positives (FPs) per breast is generated by varying the decision threshold over a range. In the second approach, prescreening of mass candidate and feature analysis are first performed on the individual two-dimensional (2D) projection view (PV) images. A mass likelihood score is estimated for each mass candidate using an LDA classifier trained for the 2D features. The mass likelihood images derived from the PVs are backprojected to the breast volume to estimate the 3D spatial distribution of the mass likelihood scores. The FROC curve for mass detection can again be generated by varying the decision threshold on the 3D mass likelihood scores merged by backprojection. In the third approach, the mass likelihood scores estimated by the 3D and 2D approaches, described above, at the corresponding 3D location are combined and evaluated using FROC analysis. A data set of 100 DBT cases acquired with a GE prototype system at the Breast Imaging Laboratory in the Massachusetts General Hospital was used for comparison of the three approaches. The LDA classifiers with stepwise feature selection were designed with leave-one-case-out resampling. In FROC analysis, the CAD system for detection in the DBT volume alone achieved test sensitivities of 80% and 90% at average FP rates of 1.94 and 3.40 per breast, respectively. With the 2D detection approach, the FP rates were 2.86 and 4.05 per breast, respectively, at the corresponding sensitivities. In comparison, the average FP rates of the system combining the 3D and 2D information were 1.23 and 2.04 per breast, respectively, at 80% and 90% sensitivities. The difference in the detection performances between the 2D and the 3D approach, and that between the 3D and the combined approach were both statistically significant (p = 0.02 and 0.01, respectively) as estimated by alternative FROC analysis. The combined system is a promising approach to improving automated mass detection on DBTs.

Spatio-temporal imaging of the hemoglobin in the compressed breast with diffuse optical tomography

We develop algorithms for imaging the time-varying optical absorption within the breast given diffuse optical tomographic data collected over a time span that is long compared to the dynamics of the medium. Multispectral measurements allow for the determination of the time-varying total hemoglobin concentration and of oxygen saturation. To facilitate the image reconstruction, we decompose the hemodynamics in time into a linear combination of spatio-temporal basis functions, the coefficients of which are estimated using all of the data simultaneously, making use of a Newton-based nonlinear optimization algorithm. The solution of the extremely large least-squares problem which arises in computing the Newton update is obtained iteratively using the LSQR algorithm. A Laplacian spatial regularization operator is applied, and, in addition, we make use of temporal regularization which tends to encourage similarity between the images of the spatio-temporal coefficients. Results are shown for an extensive simulation, in which we are able to image and quantify localized changes in both total hemoglobin concentration and oxygen saturation. Finally, a breast compression study has been performed for a normal breast cancer screening subject, using an instrument which allows for highly accurate co-registration of multispectral diffuse optical measurements with an x-ray tomosynthesis image of the breast. We are able to quantify the global return of blood to the breast following compression, and, in addition, localized changes are observed which correspond to the glandular region of the breast.

Dynamic functional and mechanical response of breast tissue to compression

Physiological tissue dynamics following breast compression offer new contrast mechanisms for evaluating breast health and disease with near infrared spectroscopy. We monitored the total hemoglobin concentration and hemoglobin oxygen saturation in 28 healthy female volunteers subject to repeated fractional mammographic compression. The compression induces a reduction in blood flow, in turn causing a reduction in hemoglobin oxygen saturation. At the same time, a two phase tissue viscoelastic relaxation results in a reduction and redistribution of pressure within the tissue and correspondingly modulates the tissue total hemoglobin concentration and oxygen saturation. We observed a strong correlation between the relaxing pressure and changes in the total hemoglobin concentration bearing evidence of the involvement of different vascular compartments. Consequently, we have developed a model that enables us to disentangle these effects and obtain robust estimates of the tissue oxygen consumption and blood flow. We obtain estimates of 1.9+/-1.3 micromol/100 mL/min for OC and 2.8+/-1.7 mL/100 mL/min for blood flow, consistent with other published values.

Compression-induced changes in the physiological state of the breast as observed through frequency domain photon migration measurements

We use optical spectroscopy to characterize the influence of mammographic-like compression on the physiology of the breast. We note a reduction in total hemoglobin content, tissue oxygen saturation, and optical scattering under compression. We also note a hyperemic effect during repeated compression cycles. By modeling the time course of the tissue oxygen saturation, we are able to obtain estimates for the volumetric blood flow (1.64+/-0.6 mL/100 mL/min) and the oxygen consumption (1.97+/-0.6 micromol/100 mL/min) of compressed breast tissue. These values are comparable to estimates obtained from previously published positron emission tomography (PET) measurements. We conclude that compression-induced changes in breast physiological properties are significant and should be accounted for when performing optical breast imaging. Additionally, the dynamic characteristics of the changes in breast physiological parameters, together with the ability to probe the tissue metabolic state, may prove useful for breast cancer detection.

Detecting Nonpalpable Recurrent Breast Cancer: The Role of Routine Mammographic Screening of Transverse Rectus Abdominis Myocutaneous Flap Reconstructions

PURPOSEnTo perform a retrospective cohort study to determine the rates of recall and cancer detection and then to develop a decision analytic model to evaluate the effectiveness of routine screening of transverse rectus abdominis myocutaneous (TRAM) flap reconstructions.nnnMATERIALS AND METHODSnThis study was approved by the institutional review board, and the methods comply with HIPAA regulations. A retrospective search of the institutional mammographic results database was done to identify bilateral screening mammographic examinations obtained from January 1, 1999, through July 15, 2005. The search included the term TRAM; the recall and cancer detetion rates were then detected. Subsequently, a decision analytic model was constructed to evaluate a hypothetical cohort of women with TRAM flap reconstructions.nnnRESULTSnOf 554 mammograms (265 TRAM flap reconstructions), 546 (98.6%) had negative results (Breast Imaging Reporting and Data System category 1 or 2). Eight (1.4%) had positive test results (Breast Imaging Reporting and Data System category 0, 3, 4, or 5). All suspicious lesions underwent biopsy and had benign pathologic results. No interval breast cancers were identified. The detection rate for nonpalpable recurrent breast cancer was 0% (exact 95% confidence interval: 0.0%, 1.4%). According to decision analysis, screening would help detect an estimated 12 additional recurrent cancers per 1000 women screened, providing an additional 1.6 days of life expectancy for the screened cohort. Under base-case conditions, screening of TRAM flap reconstructions is less effective than screening asymptomatic women in their 40s. Sensitivity analysis revealed that a benefit equivalent to that of screening asymptomatic women in their 40s was achievable under conditions related to estimates of screening effectiveness and cancer detection rate.nnnCONCLUSIONnRoutine screening mammography of TRAM flap reconstructions has a very low detection rate for nonpalpable recurrent breast cancer. Decision analysis indicates that screening such women is less effective than screening asymptomatic women in their 40s for primary breast cancer.

Characterization of masses in digital breast tomosynthesis: Comparison of machine learning in projection views and reconstructed slices

PURPOSEnIn digital breast tomosynthesis (DBT), quasi-three-dimensional (3D) structural information is reconstructed from a small number of 2D projection view (PV) mammograms acquired over a limited angular range. The authors developed preliminary computer-aided diagnosis (CADx) methods for classification of malignant and benign masses and compared the effectiveness of analyzing lesion characteristics in the reconstructed DBT slices and in the PVs.nnnMETHODSnA data set of MLO view DBT of 99 patients containing 107 masses (56 malignant and 51 benign) was collected at the Massachusetts General Hospital with IRB approval. The DBTs were obtained with a GE prototype system which acquired 11 PVs over a 50 degree arc. The authors reconstructed the DBTs at 1 mm slice interval using a simultaneous algebraic reconstruction technique. The region of interest (ROI) containing the mass was marked by a radiologist in the DBT volume and the corresponding ROIs on the PVs were derived based on the imaging geometry. The subsequent processes were fully automated. For classification of masses using the DBT-slice approach, the mass on each slice was segmented by an active contour model initialized with adaptive k-means clustering. A spiculation likelihood map was generated by analysis of the gradient directions around the mass margin and spiculation features were extracted from the map. The rubber band straightening transform (RBST) was applied to a band of pixels around the segmented mass boundary. The RBST image was enhanced by Sobel filtering in the horizontal and vertical directions, from which run-length statistics texture features were extracted. Morphological features including those from the normalized radial length were designed to describe the mass shape. A feature space composed of the spiculation features, texture features, and morphological features extracted from the central slice alone and seven feature spaces obtained by averaging the corresponding features from three to 19 slices centered at the central slice were compared. For classification of masses using the PV approach, a feature extraction process similar to that described above for the DBT approach was performed on the ROIs from the individual PVs. Six feature spaces obtained from the central PV alone and by averaging the corresponding features from three to 11 PVs were formed. In each feature space for either the DBT-slice or the PV approach, a linear discriminant analysis classifier with stepwise feature selection was trained and tested using a two-loop leave-one-case-out resampling procedure. Simplex optimization was used to guide feature selection automatically within the training set in each leave-one-case-out cycle. The performance of the classifiers was evaluated by the area (Az) under the receiver operating characteristic curve.nnnRESULTSnThe test Az values from the DBT-slice approach ranged from 0.87 +/- 0.03 to 0.93 +/- 0.02, while those from the PV approach ranged from 0.78 +/- 0.04 to 0.84 +/- 0.04. The highest test Az of 0.93 +/- 0.02 from the nine-DBT-slice feature space was significantly (p = 0.006) better than the highest test Az of 0.84 +/- 0.04 from the nine-PV feature space.nnnCONCLUSIONnThe features of breast lesions extracted from the DBT slices consistently provided higher classification accuracy than those extracted from the PV images.