Science

Abstract

Frequency-domain ultrasound waveform tomography is a promising method for the visualization and characterization of breast disease. It has previously been shown to accurately reconstruct the sound speed distributions of breasts of varying densities. The reconstructed images show detailed morphological and quantitative information that can help differentiate different types of breast disease including benign and malignant lesions. The attenuation properties of an ex vivo phantom have also been assessed. However, the reconstruction algorithms assumed a 2D geometry while the actual data acquisition process was not. Although clinically useful sound speed images can be reconstructed assuming this mismatched geometry, artifacts from the reconstruction process exist within the reconstructed images. This is especially true for registration across different modalities and when the 2D assumption is violated. For example, this happens when a patient’s breast is rapidly sloping. It is also true for attenuation imaging where energy lost or gained out of the plane gets transformed into artifacts within the image space. In this paper, we will briefly review ultrasound waveform tomography techniques, give motivation for pursuing the 3D method, discuss the 3D reconstruction algorithm, present the results of 3D forward modeling, show the mismatch that is induced by the violation of 3D modeling via numerical simulations, and present a 3D inversion of a numerical phantom.

Abstract

Mammographic percent density (MPD) is an independent risk factor for developing breast cancer, but its inclusion in clinical risk models provides only modest improvements in individualized risk prediction, and MPD is not typically assessed in younger women because of ionizing radiation concerns. Previous studies have shown that tissue sound speed, derived from whole breast ultrasound tomography (UST), a non-ionizing modality, is a potential surrogate marker of breast density, but prior to this study, sound speed has not been directly linked to breast cancer risk. To that end, we explored the relation of sound speed and MPD with breast cancer risk in a case-control study, including 61 cases with recent breast cancer diagnoses and a comparison group of 165 women, frequency matched to cases on age, race, and menopausal status, and with a recent negative mammogram and no personal history of breast cancer. Multivariable odds ratios (ORs) and 95% confidence intervals (CIs) were estimated for the relation of quartiles of MPD and sound speed with breast cancer risk adjusted for matching factors. Elevated MPD was associated with increased breast cancer risk, although the trend did not reach statistical significance (OR per quartile = 1.27, 95% CI: 0.95, 1.70; p_trend = 0.10). In contrast, elevated sound speed was significantly associated with breast cancer risk in a dose–response fashion (OR per quartile = 1.83, 95% CI: 1.32, 2.54; p_trend = 0.0003). The OR trend for sound speed was statistically significantly different from that observed for MPD (p = 0.005). These findings suggest that whole breast sound speed may be more strongly associated with breast cancer risk than MPD and offer future opportunities for refining the magnitude and precision of risk associations in larger, population-based studies, including women younger than usual screening ages.

Abstract

Objectives: Some US states have mandated that women be informed when they have dense breasts; however, little is known about how general knowledge about breast density (BD) affects related health decision-making. We examined the effects of BD information and imaging technology information on 138 African–American (AA) and European–American (EA) women’s intentions to discuss breast cancer screening with their physicians.

Methods: Women were randomly assigned to receive BD information and/or imaging technology information via 2 by 2 factorial design, and completed planned behavior measures (e.g., attitudes, intentions) related to BC screening.

Results: Attitudes mediated the effects of BD information, and the mediation was stronger for AA women compared to EA women. Effects were more robust for BD information compared to imaging technology information. Results of moderator analyses revealed suppressor effects of injunctive norms that were moderated by imaging technology information.

Conclusion: Information about BD favorably influences women’s intentions to engage in relevant breast health behaviors. Stronger attitude mediated-effects for AA women suggest greater scrutiny of BD information.

Practice implications: Since BD information may influence women’s intentions to discuss BC screening, strategies to effectively present BD information to AA women should be investigated given the likelihood of their increased scrutiny of BD information.

Abstract

Little is known about women’s knowledge of breast density or between-race differences in this knowledge. In the current study, we examined knowledge of breast density and awareness of its role as a breast cancer risk factor among women who had previously taken part in a breast imaging study. Seventy-seven women (54.5 % Black) returned a survey assessing perceptions and accuracy of breast density knowledge, knowledge of one’s own breast density, and breast cancer risk awareness. White women had greater perceived knowledge of breast density compared to Black women; however, differences in the accuracy of definitions of breast density were due to education. Black women were less likely to know how dense their own breasts were. Black and White women both lacked awareness that having dense breast increased breast cancer risk. The results highlight the need to disseminate information regarding breast density to women, while ensuring that the information is equally accessible to both Black and White women.

Abstract

Breast density, as assessed by mammography, reflects breast tissue composition. Breast epithelium and stroma attenuate x-rays more than fat and thus appear light on mammograms while fat appears dark. In this review, we provide an overview of selected areas of current knowledge about the relationship between breast density and susceptibility to breast cancer. We review the evidence that breast density is a risk factor for breast cancer, the histological and other risk factors that are associ­ated with variations in breast density, and the biological plausibility of the associations with risk of breast cancer. We also dis­cuss the potential for improved risk prediction that might be achieved by using alternative breast imaging methods, such as magnetic resonance or ultrasound. After adjustment for other risk factors, breast density is consistently associated with breast cancer risk, more strongly than most other risk factors for this disease, and extensive breast density may account for a substan­tial fraction of breast cancer. Breast density is associated with risk of all of the proliferative lesions that are thought to be pre­cursors of breast cancer. Studies of twins have shown that breast density is a highly heritable quantitative trait. Associations between breast density and variations in breast histology, risk of proliferative breast lesions, and risk of breast cancer may be the result of exposures of breast tissue to both mitogens and mutagens. Characterization of breast density by mammography has several limitations, and the uses of breast density in risk prediction and breast cancer prevention may be improved by other methods of imaging, such as magnetic resonance or ultrasound tomography.

Abstract

Women with high mammographic breast density are at 4- to 6-fold increased risk of developing breast cancer compared to women with fatty breasts. However, current breast density estimations rely on mammography, which cannot provide accurate volumetric breast representation. Therefore, we explored two techniques of breast density evaluation via ultrasound tomography. A sample of 93 patients was imaged with our clinical prototype; each dataset contained 45-75 tomograms ranging from near the chest wall through the nipple. Whole breast acoustic velocity was determined by creating image stacks and evaluating the sound speed frequency distribution. Ultrasound percent density (USPD) was determined by segmenting high sound speed areas from each tomogram using k-means clustering, integrating over the entire breast, and dividing by total breast area. Both techniques were independently evaluated using two mammographic density measures: (1) qualitative, determined by a radiologist’s visual assessment using BI-RADS Categories, and (2) quantitative, via semi-automatic segmentation to calculate mammographic percent density (MPD) for craniocaudal and medio-lateral oblique mammograms. ~140 m/s difference in acoustic velocity was observed between fatty and dense BI- RADS Categories. Increased sound speed was found with increased BI-RADS Category and quantitative MPD. Furthermore, strong positive associations between USPD, BI-RADS Category, and calculated MPD were observed. These results confirm that utilizing sound speed, both for whole-breast evaluation and segmenting locally, can be implemented to evaluate breast density.

Abstract

The Delphinus Pivotal Retrospective Reader Study (DMT-2019.002) was an analysis of radiologist image interpretation performance utilizing prospectively collected patient data obtained from the Discover SoftVue Prospective Case Collection (PCC) (DMT-2015.001). The patient data utilized in the reader study was collected from a total of six PCC Registry clinical sites across the U.S. The retrospective analysis performed was an observational case-controlled, multi-reader, multi-case Receiver Operating Characteristic (ROC) study involving 32 Readers who were MQSA qualified radiologists with experience in breast image interpretation. The cases were comprised of bilateral full-field digital mammography [FFDM] and SoftVue (SV) screening imaging acquired from the same patient during the same screening interval. There were one hundred and forty (140) cases sampled for the reader study from a pool of asymptomatic female volunteers with BI-RADS c or d breast density. The reader study provided pivotal data on AUC improvement, sensitivity, and specificity of SoftVue as a screening tool in comparison with digital mammography.

Objective To assess the feasibility of using tissue sound speed as a quantitative marker of breast density.

Methods: This study was carried out under an Institutional Review Board–approved protocol (written consent required). Imaging data were selected retrospectively based on the availability of US tomography (UST) exams, screening mammograms with volumetric breast density data, patient age of 18 to 80 years, and weight less than 300 lbs. Sound speed images from the UST exams were used to measure the volume of dense tissue, the volume averaged sound speed (VASS), and the percent of high sound speed tissue (PHSST). The mammographic breast density and volume of dense tissue were estimated with three-dimensional (3D) software. Differences in volumes were assessed with paired t-tests. Spearman correlation coefficients were calculated to determine the strength of the correlations between the mammographic and UST assessments of breast density.

Results: A total of 100 UST and 3D mammographic data sets met the selection criteria. The resulting measurements showed that UST measured a more than 2-fold larger volume of dense tissue compared to mammography. The differences were statistically significant (P < 0.001). A strong correlation of r_S = 0.85 (95% CI: 0.79–0.90) between 3D mammographic breast density (BD) and the VASS was noted. This correlation is significantly stronger than those reported in previous two-dimensional studies (r_S = 0.85 vs r_S = 0.71). A similar correlation was found for PHSST and mammographic BD with r_S = 0.86 (95% CI: 0.80–0.90).

Conclusions: The strong correlations between UST parameters and 3D mammographic BD suggest that breast sound speed should be further studied as a potential new marker for inclusion in clinical risk models.

Abstract

A population of 165 women with negative mammographic screens also received an ultrasound tomography (UST) examination at the Karmanos Cancer Institute in Detroit, MI. Standard statistical techniques were employed to measure the associations between the various mammographic- and UST-related density measures and various participant characteristics such as age, weight and height. The mammographic percent density (MPD) was found to have similar strength associations with UST mean sound speed (Spearman coefficient, r_s = 0.722, p < 0.001) and UST median sound speed (r_s = 0.737, p < 0.001). Both were stronger than the associations between MPD with two separate measures of UST percent density, a k-means (r_s = 0.568, p < 0.001) or a threshold (r_s = 0.715, p < 0.001) measure. Segmentation of the UST sound speed images into dense and non-dense volumes showed weak to moderate associations with the mammographically equivalent measures. Relationships were found to be inversely and weakly associated between age and the UST mean sound speed (r_s = −0.239, p = 0.002), UST median sound speed (r_s = −0.226, p = 0.004) and MPD (r_s = −0.204, p = 0.008). Relationships were found to be inversely and moderately associated between body mass index (BMI) and the UST mean sound speed (r_s = −0.429, p < 0.001), UST median sound speed (r_s = −0.447, p < 0.001) and MPD (r_s = −0.489, p < 0.001). The results confirm and strengthen findings presented in previous work indicating that UST sound speed imaging yields viable markers of breast density in a manner consistent with mammography, the current clinical standard. These results lay the groundwork for further studies to assess the role of sound speed imaging in risk prediction.

Authors: Carri K Glide-Hurst, Neb Duric, Peter Littrup

Abstract

Authors: Carri Glide, Nebojsa Duric, Peter Littrup

Abstract
Women with high mammographic breast density have a four- to fivefold increased risk of developing breast cancer compared to women with fatty breasts. Many preventative strategies have attempted to correlate changes in breast density with response to interventions including drugs and diet. The purpose of this work is to investigate the feasibility of assessing breast density with acoustic velocity measurements with ultrasound tomography, and to compare the results with existing measures of mammographic breast density. An anthropomorphic breast tissue phantom was first imaged with our computed ultrasound tomography clinical prototype. Strong positive correlations were observed between sound speed and material density, and sound speed and computed tomography number (Pearson correlation urn:x-wiley:0094-2405:media:mp8408:mp8408-math-0001 and 0.91, respectively). A cohort of 48 women was then imaged. 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 estimates determined by a certified radiologist using the BI-RADS Categorical Assessment based on a 1 (fatty) to 4 (dense) scale, and (2) quantitative measurements via digitization and computerized analysis of archival mammograms. A one-way analysis of variance showed that a significant difference existed between the mean values of sound speed according to BI-RADS category, while post hoc analyses using the Scheffé criterion for significance indicated that BI-RADS 4 (dense) patients had a significantly higher sound speed than BI-RADS 1, 2, and 3 at an alpha level of 0.05. Using quantitative measures of breast density, a direct correlation between the mean acoustic velocity and calculated mammographic percent breast density was demonstrated with correlation coefficients ranging from 0.75 to 0.89. The results presented here support the hypothesis that sound speed can be used as an indicator of breast tissue density. Noninvasive, nonionizing monitoring of dietary and chemoprevention interventions that affect breast density are now possible.

Abstract

Objective: To analyze the preferred tissue locations of common breast masses in relation to anatomic quadrants and the fat-glandular interface (FGI) using ultrasound tomography (UST).

Methods: Ultrasound tomography scanning was performed in 206 consecutive women with 298 mammographically and/or sonographically visible, benign and malignant breast masses following written informed consent to participate in an 8-site multicenter, Institutional Review Board-approved cohort study. Mass locations were categorized by their anatomic breast quadrant and the FGI, which was defined by UST as the high-contrast circumferential junction of fat and fibroglandular tissue on coronal sound speed imaging. Quantitative UST mass comparisons were done for each tumor and peritumoral region using mean sound speed and percentage of fibroglandular tissue. Chi-squared and analysis of variance tests were used to assess differences.

Results: Cancers were noted at the FGI in 95% (74/78) compared to 51% (98/194) of fibroadenomas and cysts combined (P < 0.001). No intra-quadrant differences between cancer and benign masses were noted for tumor location by anatomic quadrants (P = 0.66). Quantitative peritumoral sound speed properties showed that cancers were surrounded by lower mean sound speeds (1477 m/s) and percent fibroglandular tissue (47%), compared to fibroadenomas (1496 m/s; 65.3%) and cysts (1518 m/s; 84%) (P < 0.001; P < 0.001, respectively).

Conclusion: Breast cancers form adjacent to fat and UST localized the vast majority to the FGI, while cysts were most often completely surrounded by dense tissue. These observations were supported by quantitative peritumoral analyses of sound speed values for fat and fibroglandular tissue.

Abstract

Recent analyses of tumor location by breast magnetic resonance (MR) imaging have shown that breast cancers were significantly more likely to be located at or near the fat-gland interface (FGI). However, these analyses did not include benign masses or coronal evaluations. Ultrasound tomography (UST) is a novel multi-sequence tomographic form of sonography with anatomic and physiologic information that can improve the sensitivity of breast cancer detection, particularly in women with dense breasts. The purpose of this study was to use the native coronal imaging plane of UST SoftVue (Delphinus Medical Technologies, Inc.) to map the locations of cancers, fibroadenomas and cysts, relative to the FGI, in order to facilitate the detection and characterization of breast masses in future UST screening.

Abstract

Objective: To analyze the preferred tissue locations of common breast masses in relation to anatomic quadrants and the fat-glandular interface (FGI) using ultrasound tomography (UST).

Methods: Ultrasound tomography scanning was performed in 206 consecutive women with 298 mammographically and/or sonographically visible, benign and malignant breast masses following written informed consent to participate in an 8-site multicenter, Institutional Review Board-approved cohort study. Mass locations were categorized by their anatomic breast quadrant and the FGI, which was defined by UST as the high-contrast circumferential junction of fat and fibroglandular tissue on coronal sound speed imaging. Quantitative UST mass comparisons were done for each tumor and peritumoral region using mean sound speed and percentage of fibroglandular tissue. Chi-squared and analysis of variance tests were used to assess differences.

Results: Cancers were noted at the FGI in 95% (74/78) compared to 51% (98/194) of fibroadenomas and cysts combined (P < 0.001). No intra-quadrant differences between cancer and benign masses were noted for tumor location by anatomic quadrants (P = 0.66). Quantitative peritumoral sound speed properties showed that cancers were surrounded by lower mean sound speeds (1477 m/s) and percent fibroglandular tissue (47%), compared to fibroadenomas (1496 m/s; 65.3%) and cysts (1518 m/s; 84%) (P < 0.001; P < 0.001, respectively).

Conclusion: Breast cancers form adjacent to fat and UST localized the vast majority to the FGI, while cysts were most often completely surrounded by dense tissue. These observations were supported by quantitative peritumoral analyses of sound speed values for fat and fibroglandular tissue.

Abstract

Recent analyses of tumor location by breast magnetic resonance (MR) imaging have shown that breast cancers were significantly more likely to be located at or near the fat-gland interface (FGI). However, these analyses did not include benign masses or coronal evaluations. Ultrasound tomography (UST) is a novel multi-sequence tomographic form of sonography with anatomic and physiologic information that can improve the sensitivity of breast cancer detection, particularly in women with dense breasts. The purpose of this study was to use the native coronal imaging plane of UST SoftVue (Delphinus Medical Technologies, Inc.) to map the locations of cancers, fibroadenomas and cysts, relative to the FGI, in order to facilitate the detection and characterization of breast masses in future UST screening.

Abstract

Mammography is not sufficiently effective for women with dense breast tissue – women who are at much higher risk for developing breast cancer. Consequently, many breast cancers go undetected at their treatable stage. Improved cancer detection and characterization for women with dense breast tissue is urgently needed. Our clinical study has shown that ultrasound tomography (UST) is an emerging technique that moves beyond B-mode imaging by its through transmission capabilities. Transmission ultrasound provides additional tissue parameters such as sound speed, attenuation, and through-transmission rendered tissue stiffness information. For women with dense breasts, these parameters can be used to assist in detecting malignant masses within glandular or fatty tissue and differentiating malignant and benign masses. This paper focuses on the use of waveform ultrasound sound speed imaging and tissue stiffness information generated using through-transmission data to characterize different breast tissues and breast masses. In-vivo examples will be given to assess its effectiveness.

Abstract

Both mammography and standard ultrasound (US) rely upon subjective criteria within the breast imaging reporting and data system (BI-RADS) to provide more uniform interpretation outcomes, as well as differentiation and risk stratification of associated abnormalities. In addition, the technical performance and professional interpretation of both tests suffer from machine and operator dependence. Breast MR has become the new gold standard for screening of high-risk women but has cost and access limitations in extending screening to the entire population. We have been developing a new technique for breast imaging that is based on ultrasound tomography which quantifies tissue characteristics while also producing 3-D images of breast anatomy. Results are presented from clinical studies that utilize this method. Informed consent was obtained from all patients, prospectively recruited in an IRB-approved protocol following HIPAA guidelines. Images were produced by tomographic algorithms for reflection, sound speed and attenuation. All images were reviewed by a board-certified radiologist who has more than 20 years of experience in breast imaging and US-technology development. In the first phase of the study, UST images were compared to multi-modal imaging to determine the appearance of lesions and breast parenchyma. In the second phase of the study, correlative comparisons with MR breast imaging were used to establish basic operational capabilities of the UST system including the identification and characterization of parenchymal patterns. Our study demonstrated a high degree of correlation of breast tissue structures relative to fat subtracted contrast enhanced MRI. With a scan duration of ~ 1-3 minutes, no significant motion artifacts were observed.

Abstract

To evaluate the diagnostic value of Ultrasound Computer Tomography (USCT), the imaging results have to be correlated with conventional breast imaging techniques. This is challenging due to different patient positioning in the modalities with nonlinear deformations of the breast tissue. We have developed a patient-specific image registration method, which simulates different breast positionings in both X-ray mammography and Magnetic Resonance Imaging (MRI) through biomechanical modelling. An average registration error below 5 and 17 mm for MRI to USCT and USCT to mammography registration, respectively, allowed us to evaluate the diagnostic performance of USCT. It was shown that regions of high sound speed corresponded well with the tumour position indicated from the MRI contrast kinetic map. Moreover, the quantitative analysis of sound speed and attenuation values with respect to the segmented mammograms revealed that sound speed gives a better distinction between breast tissue, whereas their combined information further improves the classification. Although the results are based on a preliminary study, the promising outcome points that the registration could assist radiologists in comparing the USCT with both MRI and X-ray mammography.

Abstract

Ultrasound tomography (UST) generates several different imaging modalities. This includes reflection, sound speed, and attenuation images. The images visualize different types of breast diseases or tissues. Typically, a radiologist views the images to determine a diagnosis for the patient. However, a learning algorithm could be trained to predict diagnosis based on the features contained within the image. Thus, our objective is to create classifier models which map features in images to labels.

Abstract

Ultrasound computed tomography (USCT) holds great promise for breast cancer screening. Waveform inversion-based image reconstruction methods account for higher order diffraction effects and can produce high-resolution USCT images, but are computationally demanding. Recently, a source encoding technique has been combined with stochastic gradient descent (SGD) to greatly reduce image reconstruction times. However, this method bundles the stochastic data fidelity term with the deterministic regularization term. This limitation can be overcome by replacing SGD with a structured optimization method, such as the regularized dual averaging method, that exploits knowledge of the composition of the cost function. In this paper, the dual averaging method is combined with source encoding techniques to improve the effectiveness of regularization while maintaining the reduced reconstruction times afforded by source encoding. It is demonstrated that each iteration can be decomposed into a gradient descent step based on the data fidelity term and a proximal update step corresponding to the regularization term. Furthermore, the regularization term is never explicitly differentiated, allowing nonsmooth regularization penalties to be naturally incorporated. The wave equation is solved by the use of a time-domain method. The effectiveness of this approach is demonstrated through computer simulation and experimental studies. The results suggest that the dual averaging method can produce images with less noise and comparable resolution to those obtained by the use of SGD.

Abstract

Frequency-domain ultrasound waveform tomography is a promising method for the visualization and characterization of breast disease. It has previously been shown to accurately reconstruct the sound speed distributions of breasts of varying densities. The reconstructed images show detailed morphological and quantitative information that can help differentiate different types of breast disease including benign and malignant lesions. The attenuation properties of an ex vivo phantom have also been assessed. However, the reconstruction algorithms assumed a 2D geometry while the actual data acquisition process was not. Although clinically useful sound speed images can be reconstructed assuming this mismatched geometry, artifacts from the reconstruction process exist within the reconstructed images. This is especially true for registration across different modalities and when the 2D assumption is violated. For example, this happens when a patient’s breast is rapidly sloping. It is also true for attenuation imaging where energy lost or gained out of the plane gets transformed into artifacts within the image space. In this paper, we will briefly review ultrasound waveform tomography techniques, give motivation for pursuing the 3D method, discuss the 3D reconstruction algorithm, present the results of 3D forward modeling, show the mismatch that is induced by the violation of 3D modeling via numerical simulations, and present a 3D inversion of a numerical phantom.

Abstract

A population of 165 women with negative mammographic screens also received an ultrasound tomography (UST) examination at the Karmanos Cancer Institute in Detroit, MI. Standard statistical techniques were employed to measure the associations between the various mammographic- and UST-related density measures and various participant characteristics such as age, weight and height. The mammographic percent density (MPD) was found to have similar strength associations with UST mean sound speed (Spearman coefficient, r_s = 0.722, p < 0.001) and UST median sound speed (r_s = 0.737, p < 0.001). Both were stronger than the associations between MPD with two separate measures of UST percent density, a k-means (r_s = 0.568, p < 0.001) or a threshold (r_s = 0.715, p < 0.001) measure. Segmentation of the UST sound speed images into dense and non-dense volumes showed weak to moderate associations with the mammographically equivalent measures. Relationships were found to be inversely and weakly associated between age and the UST mean sound speed (r_s = −0.239, p = 0.002), UST median sound speed (r_s = −0.226, p = 0.004) and MPD (r_s = −0.204, p = 0.008). Relationships were found to be inversely and moderately associated between body mass index (BMI) and the UST mean sound speed (r_s = −0.429, p < 0.001), UST median sound speed (r_s = −0.447, p < 0.001) and MPD (r_s = −0.489, p < 0.001). The results confirm and strengthen findings presented in previous work indicating that UST sound speed imaging yields viable markers of breast density in a manner consistent with mammography, the current clinical standard. These results lay the groundwork for further studies to assess the role of sound speed imaging in risk prediction.

Abstract

Ultrasound waveform tomography techniques have shown promising results for the visualization and characterization of breast disease. By using frequency-domain waveform tomography techniques and a gradient descent algorithm, we have previously reconstructed the sound speed distributions of breasts of varying densities with different types of breast disease including benign and malignant lesions. By allowing the sound speed to have an imaginary component, we can model the intrinsic attenuation of a medium. We can similarly recover the imaginary component of the velocity and thus the attenuation. In this paper, we will briefly review ultrasound waveform tomography techniques, discuss attenuation and its relations to the imaginary component of the sound speed, and provide both numerical and ex vivo examples of waveform tomography attenuation reconstructions.

Abstract

Conventional ultrasonography reconstruction techniques, such as B-mode, are based on a simple wave propagation model derived from a high frequency approximation. Therefore, to minimize model mismatch, the central frequency of the input pulse is typically chosen between 3 and 15 megahertz. Despite the increase in theoretical resolution, operating at higher frequencies comes at the cost of lower signal-to-noise ratio. This ultimately degrades the image contrast and overall quality at higher imaging depths. To address this issue, we investigate a reflection imaging technique, known as reverse time migration, which uses a more accurate propagation model for reconstruction. We present preliminary phantom results obtained using data acquired with a breast imaging ultrasound tomography prototype. The original reconstructions are filtered to remove low-wavenumber artifacts that arise due to the inclusion of the direct arrivals. We demonstrate the advantage of using an accurate sound speed model in the reverse time migration process. We also explain how the increase in computational complexity can be mitigated using a frequency domain approach and a parallel computing platform.

Abstract

Application of the frequency domain acoustic wave equation on data acquired from ultrasound tomography scans is shown to yield high resolution sound speed images on the order of the wavelength of the highest reconstructed frequency. Using a signal bandwidth of 0.4–1 MHz and an average sound speed of 1500 m s⁻¹, the resolution is approximately 1.5 mm. The quantitative sound speed values and morphology provided by these images have the potential to inform diagnosis and classification of breast disease. In this study, we present the formalism, practical application, and in vivo results of waveform tomography applied to breast data gathered by two different ultrasound tomography scanners that utilize ring transducers. The formalism includes a review of frequency domain modeling of the wave equation using finite difference operators as well as a review of the gradient descent method for the iterative reconstruction scheme. It is shown that the practical application of waveform tomography requires an accurate starting model, careful data processing, and a method to gradually incorporate higher frequency information into the sound speed reconstruction. Following these steps resulted in high resolution quantitative sound speed images of the breast. These images show marked improvement relative to commonly used ray tomography reconstruction methods. The robustness of the method is demonstrated by obtaining similar results from two different ultrasound tomography devices. We also compare our method to MRI to demonstrate concordant findings. The clinical data used in this work was obtained from a HIPAA compliant clinical study (IRB 040912M1F).

Abstract

Ultrasound computed tomography (USCT) holds great promise for improving the detection and management of breast cancer. Because they are based on the acoustic wave equation, waveform inversion-based reconstruction methods can produce images that possess improved spatial resolution properties over those produced by ray-based methods. However, waveform inversion methods are computationally demanding and have not been applied widely in USCT breast imaging. In this work, source encoding concepts are employed to develop an accelerated USCT reconstruction method that circumvents the large computational burden of conventional waveform inversion methods. This method, referred to as the waveform inversion with source encoding (WISE) method, encodes the measurement data using a random encoding vector and determines an estimate of the speed-of-sound distribution by solving a stochastic optimization problem by use of a stochastic gradient descent algorithm. Computer-simulation studies are conducted to demonstrate the use of the WISE method. Using a single graphics processing unit card, each iteration can be completed within 25 seconds for a 128 × 128 mm² reconstruction region. The results suggest that the WISE method maintains the high spatial resolution of waveform inversion methods while significantly reducing the computational burden.

Abstract

Ultrasound tomography is a promising modality for breast imaging. Many current ultrasound tomography imaging algorithms are based on ray theory and assume a homogeneous background which is inaccurate for complex heterogeneous regions. They fail when the size of lesions approaches the wavelength of ultrasound used. Therefore, to accurately image small lesions, wave theory must be used in ultrasound imaging algorithms to properly handle the heterogeneous nature of breast tissue and the diffraction effects that it induces. Using frequency-domain ultrasound waveform tomography, we present sound speed reconstructions of both a tissue-mimicking breast phantom and in vivo data sets. Significant improvements in contrast and resolution are made upon the previous ray based methods. Where it might have been difficult to differentiate a high sound speed tumor from bulk breast parenchyma using ray based methods, waveform tomography improves the shape and margins of a tumor to help more accurately differentiate it from the bulk breast tissue. Waveform tomography sound speed imaging might improve the ability of finding lesions in very dense tissues, a difficult environment for mammography. By comparing the sound speed images produced by waveform tomography to MRI, we see that the complex structures in waveform tomography are consistent with those in MRI. The robustness of the method is established by reconstructing data acquired by two different ultrasound tomography prototypes.

Abstract

A number of clinical trials have shown that screening ultrasound, supplemental to mammography, detects additional cancers in women with dense breasts. However, labor intensity, operator dependence and high recall rates have limited adoption. This paper describes the use of ultrasound tomography for whole-breast tissue stiffness measurements as a first step toward addressing the issue of high recall rates. The validation of the technique using an anthropomorphic phantom is described. In-vivo applications are demonstrated on 13 breast masses, indicating that lesion stiffness correlates with lesion type as expected. Comparison of lesion stiffness measurements with standard elastography was available for 11 masses and showed a strong correlation between the 2 measures. It is concluded that ultrasound tomography can map out the 3 dimensional distribution of tissue stiffness over the whole breast. Such a capability is well suited for screening where additional characterization may improve the specificity of screening ultrasound, thereby lowering barriers to acceptance.

Abstract

Ultrasound computed tomography (USCT) holds great promise for improving the detection and management of breast cancer. Because they are based on the acoustic wave equation, waveform inversion-based reconstruction methods can produce images that possess improved spatial resolution properties over those produced by ray-based methods. However, waveform inversion methods are computationally demanding and have not been applied widely in USCT breast imaging. In this work, source encoding concepts are employed to develop an accelerated USCT reconstruction method that circumvents the large computational burden of conventional waveform inversion methods. This method, referred to as the waveform inversion with source encoding (WISE) method, encodes the measurement data using a random encoding vector and determines an estimate of the sound speed distribution by solving a stochastic optimization problem by use of a stochastic gradient descent algorithm. Both computer simulation and experimental phantom studies are conducted to demonstrate the use of the WISE method. The results suggest that the WISE method maintains the high spatial resolution of waveform inversion methods while significantly reducing the computational burden.

Abstract

Ultrasound Computer Tomography (USCT) is a promising breast imaging modality under development. Comparison to a standard method like mammography is essential for further development. Due to significant differences in image dimensionality and compression state of the breast, correlating USCT images and X-ray mammograms is challenging. In this paper we present a 2D/3D registration method to improve the spatial correspondence and allow direct comparison of the images. It is based on biomechanical modeling of the breast and simulation of the mammographic compression. We investigate the effect of including patient-specific material parameters estimated automatically from USCT images. The method was systematically evaluated using numerical phantoms and in-vivo data. The average registration accuracy using the automated registration was 11.9 mm. Based on the registered images a method for analysis of the diagnostic value of the USCT images was developed and initially applied to analyze sound speed and attenuation images based on X-ray mammograms as ground truth. Combining sound speed and attenuation allows differentiating lesions from surrounding tissue. Overlaying this information on mammograms, combines quantitative and morphological information for multimodal diagnosis.

Abstract

Ultrasound tomography is an emerging modality for breast imaging. However, most current ultrasonic tomography imaging algorithms, historically hindered by the limited memory and processor speed of computers, are based on ray theory and assume a homogeneous background which is inaccurate for complex heterogeneous regions. Therefore, wave theory, which accounts for diffraction effects, must be used in ultrasonic imaging algorithms to properly handle the heterogeneous nature of breast tissue in order to accurately image small lesions. However, application of waveform tomography to medical imaging has been limited by extreme computational cost and convergence. By taking advantage of the computational architecture of Graphic Processing Units (GPUs), the intensive processing burden of waveform tomography can be greatly alleviated. In this study, using breast imaging methods, we implement a frequency domain waveform tomography algorithm on GPUs with the goal of producing high-accuracy and high-resolution breast images on clinically relevant time scales. We present some simulation results and assess the resolution and accuracy of our waveform tomography algorithms based on the simulation data.

Abstract

We describe the clinical performance of SoftVue, a breast imaging device based on the principles of ultrasound tomography. Participants were enrolled in an IRB-approved study at Wayne State University, Detroit, MI. The main research findings indicate that SoftVue is able to image the whole uncompressed breast up to cup size H. Masses can be imaged in even the densest breasts with the ability to discern margins and mass shapes. Additionally, it is demonstrated that multi-focal disease can also be imaged. The system was also tested in its research mode for additional imaging capabilities. These tests demonstrated the potential for generating tissue stiffness information for the entire breast using through-transmission data. This research capability differentiates SoftVue from the other whole breast systems on the market. It is also shown that MRI-like images can be generated using alternative processing of the echo data. Ongoing research is focused on validating and quantifying these findings in a larger sample of study participants and quantifying SoftVue's ability to differentiate benign masses from cancer.

Abstract

Ultrasound tomography (UST) employs sound waves to produce three-dimensional images of breast tissue and precisely measures the attenuation of sound speed secondary to breast tissue composition. High breast density is a strong breast cancer risk factor and sound speed is directly proportional to breast density. UST provides a quantitative measure of breast density based on three-dimensional imaging without compression, thereby overcoming the shortcomings of many other imaging modalities. The quantitative nature of the UST breast density measures are tied to an external standard, so sound speed measurement in breast tissue should be independent of specific hardware. The work presented here compares breast sound speed measurement obtained with two different UST devices. The Computerized Ultrasound Risk Evaluation (CURE) system located at the Karmanos Cancer Institute in Detroit, Michigan was recently replaced with the SoftVue ultrasound tomographic device. Ongoing clinical trials have used images generated from both sets of hardware, so maintaining consistency in sound speed measurements is important. During an overlap period when both systems were in the same exam room, a total of 12 patients had one or both of their breasts imaged on both systems on the same day. There were 22 sound speed scans analyzed from each system and the average breast sound speeds were compared. Images were either reconstructed using saved raw data (for both CURE and SoftVue) or were created during the image acquisition (saved in DICOM format for SoftVue scans only). The sound speed measurements from each system were strongly and positively correlated with each other. The average difference in sound speed between the two sets of data was on the order of 1-2 m/s and this result was not statistically significant. The only sets of images that showed a statistical difference were the DICOM images created during the SoftVue scan compared to the SoftVue images reconstructed from the raw data. However, the discrepancy between the sound speed values could be easily handled by uniformly increasing the DICOM sound speed by approximately 0.5 m/s. These results suggest that there is no fundamental difference in sound speed measurement for the two systems and support combining data generated with these instruments in future studies.

Abstract

The purpose of this paper is to describe the technical and clinical performance of SoftVue, a breast imaging device based on the principles of ultrasound tomography. We report on initial results from data acquired from 30 participants in a recently undertaken clinical study. Initial results suggest that SoftVue can accurately image the full range of breast anatomy, including both benign lesions and cancer. Ongoing research is focused on assessing SoftVue's ability to differentiate benign masses from cancer.

Abstract

Ultrasound tomography (UST) is a breast imaging modality that is radiation free, operator independent, and does not involve breast compression. In the UST system under consideration, the breast is surrounded by a transducer ring that moves along the coronal axis from the chest wall to the nipple region. The deployment of UST in a clinical setting is technically challenging from three major standpoints: acquisition speed, storage capability, and computational requirements. Data acquisition must be fast to maximize patient throughput and minimize image registration artifacts. Unlike traditional ultrasound, hundreds of gigabytes of data must be acquired, buffered, and processed to image various characteristics of breast tissues such as sound speed, attenuation, and reflectivity. The tomographic image reconstruction methods are non-linear, iterative algorithms with significant computational complexity. Moreover, the scanner hosting the acquisition and reconstruction components must satisfy stringent cost, power, and size requirements. For decades, the above constraints have hindered the practicality of UST in a clinical scenario. We describe the design of a UST system that addresses the relevant clinical requirements as a means to demonstrate the feasibility of UST deployment in a clinical setting.

Abstract

Ultrasound Computer Tomography (USCT) is a promising 3D modality for early breast cancer detection, which is expected to provide quantitative imaging. The aim of this paper is to evaluate the quantitative diagnostic value of the USCT images, i.e. sound speed and attenuation images, using X-ray mammograms as ground truth. For this purpose we applied our 2D/3D registration method, which is based on biomechanical modeling of the breast. Mammograms were segmented into fatty, glandular and tumorous tissue. For each tissue, the average sound speed and attenuation in the corresponding USCT images was calculated. Tumorous tissue could be separated from fatty and glandular tissue using a fixed absolute sound speed threshold in all regarded datasets. By combining sound speed and attenuation, the separation between fatty and glandular tissue could be improved. By overlaying sound speed and attenuation information on the mammogram, quantitative and morphological information can be combined for multimodal diagnosis. This may benefit early breast cancer detection in future.

Abstract

Ultrasound Computer Tomography (USCT) is an upcoming modality for early breast cancer diagnosis, which provides morphological as well as quantitative imaging. In order to compare USCT images to the standard modality X-ray mammography, a 2D/3D registration has to be applied. To analyze the relevance of sound speed images as a quantitative imaging method for tissue characterization, the aim of this paper is to quantify sound speed values of different types of tissue using mammograms as ground truth. Mammograms are segmented and classified into fat, glandular and tumorous tissue. For each tissue class the average sound speed in the registered sound speed image is calculated. The mean absolute sound speed was 1457 m/s for regions segmented as fatty tissue, 1470 m/s for glandular and 1509 m/s for tumorous tissue. For all ten datasets, the sound speed in tumorous tissue was significantly higher than in glandular and fatty tissue. Also glandular and fatty tissue could be separated easily by the absolute sound speed values. By color-coding sound speed, quantitative information from USCT and morphological information from X-ray mammography are fused for combined diagnosis. We believe this method will help radiologists in gaining experience in the reading of USCT images. The combination of diagnostic information is likely to be beneficial to early breast cancer detection.

Abstract

For women with dense breast tissue, who are at much higher risk for developing breast cancer, the performance of mammography is at its worst. Consequently, many early cancers go undetected when they are the most treatable. Improved cancer detection for women with dense breasts would decrease the proportion of breast cancers diagnosed at later stages, which would significantly lower the mortality rate. The emergence of whole breast ultrasound provides good performance for women with dense breast tissue, and may eliminate the current trade-off between the cost effectiveness of mammography and the imaging performance of more expensive systems such as magnetic resonance imaging. We report on the performance of SoftVue, a whole breast ultrasound imaging system, based on the principles of ultrasound tomography. SoftVue was developed by Delphinus Medical Technologies and builds on an early prototype developed at the Karmanos Cancer Institute. We present results from preliminary testing of the SoftVue system, performed both in the lab and in the clinic. These tests aimed to validate the expected improvements in image performance. Initial qualitative analyses showed major improvements in image quality, thereby validating the new imaging system design. Specifically, SoftVue’s imaging performance was consistent across all breast density categories and had much better resolution and contrast. The implications of these results for clinical breast imaging are discussed and future work is described.

Abstract

Ultrasound Computer Tomography is an upcoming imaging modality for early breast cancer detection. For evaluation of the method, comparison with the standard method X-ray mammography is of strongest interest. To overcome the significant differences in dimensionality and compression state of the breast, in earlier work a registration method based on biomechanical modeling of the breast was proposed. However only homogeneous models could be applied, i.e. inner structures of the breast were neglected. In this work we extend the biomechanical modeling of the breast by estimating patient-specific tissue parameters automatically from the speed of sound volume. Two heterogeneous models are proposed modeling a quadratic and an exponential relationship between speed of sound and tissue stiffness. The models were evaluated using phantom images and clinical data. The size of all lesions is better preserved using heterogeneous models, especially using an exponential relationship. The presented approach yields promising results and gives a physical justification to our registration method. It can be considered as a first step towards a realistic modeling of the breast.

Abstract

Previous studies concluded that the resolution limitation of travel-time tomography is the width of the first Fresnel zone. However, we believe that the resolution of ray tomography cannot simply be limited to the first Fresnel zone and is affected by many factors. In this study, we investigate a variety of factors that affect the resolving power of travel time tomography. These factors include accuracy of picked travel time, ray coverage (data density) and data signal-to-noise ratio (SNR). We also investigate to what extent that bent-ray travel-time tomography is capable of resolving anomalous objects smaller than the first Fresnel zone based on numerical simulations. We have shown that bent-ray travel-time tomography resolvability and detectability of small objects is better than the first Fresnel zone.

Abstract

Ultrasound tomography (UST) is being developed to address the limitations of mammography in breast cancer detection. Central to the success of UST is the possibility of obtaining high-resolution images of tissue mechanical properties across the whole breast. A recent paper [Huthwaite and Simonetti, J. Acoust. Soc. Am. 130, 1721–1734 (2011)] made use of a numerical phantom to demonstrate that sufficient image resolution can be obtained by simply treating refraction and diffraction effects in consecutive steps through the combination of ray-based time of flight and diffraction tomography. This letter presents the first experimental demonstration of the method using phantom and invivo data from a cancer patient.

Abstract

Imaging with acoustic waves has made great advances in recent decades. In opposing limits of wavelength, acoustics have played a major role in geophysical applications on the one hand and in medical ultrasound imaging on the other. In contrast to X-rays, acoustic waves interact strongly with materials through which they propagate, through processes such as refraction, reflection and diffraction. The interactions can be very strong in heterogeneous media such as human tissue. Tomographic reconstructions of acoustic data therefore require much more sophisticated modeling of acoustic wave propagation often involving highly non-linear inversions. These factors have impeded progress in this otherwise promising methodology. The advancement of computing power and the rise of high-throughput data acquisition hardware have made acoustic tomography (AT) feasible in recent years. The objective of this paper is to relate these developments to practical applications of AT, particularly in the area of medical imaging. Today, a number of laboratory groups are collecting data with AT prototypes and some projects have become commercial ventures. This paper reviews the status of AT imaging, articularly in the area of breast cancer detection, where some of the most recent advances have taken place. It is shown that parallel developments in AT methodologies have given rise to exciting new possibilities for acoustic tomography, at all wavelengths, with potential applications in areas as diverse as seismic exploration, non-destructive testing and cancer detection.

Abstract

Breast cancer is the most common cancer among women. The established screening method to detect breast cancer in an early state is X-ray mammography. However, X-ray frequently provides limited contrast of tumors located within glandular tissue. A new imaging approach is Ultrasound Computer Tomography generating three dimensional volumes of the breast. Three different images are available: reflectivity, attenuation and speed of sound. The correlation of USCT volumes with X-ray mammograms is of interest for evaluation of the new imaging modality as well as for a multimodal diagnosis. Yet, both modalities differ in image dimensionality, patient positioning and deformation state of the breast. In earlier work we proposed a methodology based on Finite Element Method to register speed of sound images with the according mammogram. In this work, we enhanced the methodology to register all three image types provided by USCT. Furthermore, the methodology is now completely automated using image similarity measures to estimate rotations in datasets. A fusion methodology is proposed which combines the information of the three USCT image types with the X-ray mammogram via semitransparent overlay images. The evaluation was done using 13 datasets from a clinical study. The registration accuracy was measured by the displacement of the center of a lesion marked in both modalities. Using the automated rotation estimation, a mean displacement of 10.4 mm was achieved. Due to the clinically relevant registration accuracy, the methodology provides a basis for evaluation of the new imaging device USCT as well as for multimodal diagnosis.

Abstract

Conventional sonography, which performs well in dense breast tissue and is comfortable and radiation-free, is not practical for screening because of its operator dependence and the time needed to scan the whole breast. While magnetic resonance imaging (MRI) can significantly improve on these limitations, it is also not practical because it has long been prohibitively expensive for routine use. There is therefore a need for an alternative breast imaging method that obviates the constraints of these standard imaging modalities. The lack of such an alternative is a barrier to dramatically impacting mortality (about 45,000 women in the US per year) and morbidity from breast cancer because, currently, there is a trade-off between the cost effectiveness of mammography and sonography on the one hand and the imaging accuracy of MRI on the other. This paper presents a progress report on our long term goal to eliminate this trade-off and thereby improve breast cancer survival rates and decrease unnecessary biopsies through the introduction of safe, cost-effective, operatorindependent sonography that can rival MRI in accuracy. The objective of the study described in this paper was to design and build an improved ultrasound tomography (UST) scanner in support of our goals. To that end, we report on a design that builds on our current research prototype. The design of the new scanner is based on a comparison of the capabilities of our existing prototype and the performance needed for clinical efficacy. The performance gap was quantified by using clinical studies to establish the baseline performance of the research prototype, and using known MRI capabilities to establish the required performance. Simulation software was used to determine the basic operating characteristics of an improved scanner that would provide the necessary performance. Design elements focused on transducer geometry, which in turn drove the data acquisition system and the image reconstruction engine specifications. The feasibility of UST established by our earlier work and that of other groups, forms the rationale for developing a UST system that has the potential to become a practical, low-cost device for breast cancer screening and diagnosis.

Abstract

A multi-grid tomographic inversion approach that uses variable grid sizes in both forward modeling and inverse process is proposed and tested on breast phantom data and breast ultrasound data. In iterative tomographic inversion, fine scale features are more sensitive to starting model than coarse scale features. The proposed multi-grid algorithm starts from coarse grids for both forward modeling and inverse process and gradually proceeds to fine grids, which can effectively suppress artifacts related to over iteration of fine scale features. Since the computational complexity of inverse problems increases with number of grid points in both forward model and inverse model, the proposed algorithm greatly reduces the computational cost in contrast to standard fixed-grid approaches. Both in vitro and in vivo results indicate that the proposed multi-grid methods result in significant improvement in the inverted sound speed and attenuation images compared to fixed-grid methods.

Abstract

Accurate time delay estimation is critical for a wide range of remote sensing applications. We propose a technique that exploits the redundancy between absolute and relative time delays in transducer arrays as a means to reduce the level of noise present in the measurements. We formalize the problem of interest and present two different strategies to solve it. The first strategy is optimal in the mean square sense but requires a quadratic programming solver. The second approach is based on a sub-optimal iterative denoising technique. The effectiveness of our approach is demonstrated in the context of travel time tomographic imaging using numerical and physical breast mimicking phantoms as well as patient data.

Abstract

Breast ultrasound tomography is a rapidly developing imaging modality that has the potential to impact breast cancer screening and diagnosis. Double difference (DD) tomography utilizes more accurate differential time-of-flight (ToF) data to reconstruct the sound speed structure of the breast. It can produce more precise and better resolution sound speed images than standard tomography that uses absolute ToF data. We apply DD tomography to phantom data and excised mouse mammary glands data. DD tomograms demonstrate sharper sound speed contrast than the standard tomograms.

Abstract

Conventional ultrasound techniques use beam-formed, constant sound speed ray models for fast image reconstruction. However, these techniques are inadequate for the emerging new field of ultrasound tomography (UST). We present a new technique for the reconstruction of reflection images from UST data. We have extended the planar Kirchhoff migration method used in geophysics, and combined it with sound speed and attenuation data obtained from the transmission signals to create reflection ultrasound images that are corrected for refractive and attenuative effects. The resulting technique was applied to in-vivo breast data obtained with an experimental prototype. The results indicate that sound speed and attenuation corrections lead to considerable improvements in image quality, particularly in dense tissues where the refractive and scattering effects are the greatest.

Abstract

We report on the use of ultrasound tomography (UST) to characterize breast cancer and study the local and distant tumor environments. We have imaged the tumor and its environment in 3 cases of breast cancer using a UST prototype and its associated image reconstruction algorithms. After generating images of reflection, sound speed and attenuation, the images were fused in combinations that allowed visualization and characterization of the interior of the tumor as well as the tissue immediate to the tumor and beyond. The reflection UST images demonstrated the presence of spiculation, and architectural distortion, indicators of both local tumor invasion and distant involvement with surrounding tissues. Furthermore, the sound speed images showed halos of elevated sound speed surrounding the tumors, indicating a local environment characterized by stiff tissues. The combination of sound speed and attenuation images revealed that the tumor interiors were the stiffest tissues in the region studied. These features and characteristics are commensurate with the known biomechanical properties of cancer and may be manifestations of the desmoplastic process that is associated with tumor invasion. We propose that UST imaging may prove to be a valuable tool for characterizing cancers and studying the tumor invasion process.

Abstract

Accurate calibration is a requirement of many array signal processing techniques. We investigate the calibration of a transducer array using time delays. We derive a strategy based on the mean square error criterion and discuss how time delays that are not available can be interpolated from existing ones. The proposed method is made robust to noise and model mismatch by means of a novel iterative technique for distance matrix denoising. The convergence of the method is proved. Finally, the accuracy of the proposed calibration algorithm is assessed both in simulated scenarios and using experimental data obtained from an ultrasound scanner designed for breast cancer detection.

Abstract

Purpose: To explore the feasibility of improving cross-sectional reflection imaging of the breast using refractive and attenuation corrections derived from ultrasound tomography data.
Methods: The authors have adapted the planar Kirchhoff migration method, commonly used in geophysics to reconstruct reflection images, for use in ultrasound tomography imaging of the breast. Furthermore, the authors extended this method to allow for refractive and attenuative corrections. Using clinical data obtained with a breast imaging prototype, the authors applied this method to generate cross-sectional reflection images of the breast that were corrected using known distributions of sound speed and attenuation obtained from the same data.
Results: A comparison of images reconstructed with and without the corrections showed varying degrees of improvement. The sound speed correction resulted in sharpening of detail, while the attenuation correction reduced the central darkening caused by path length dependent losses. The improvements appeared to be greatest when dense tissue was involved and the least for fatty tissue. These results are consistent with the expectation that denser tissues lead to both greater refractive effects and greater attenuation.
Conclusions: Although conventional ultrasound techniques use time-gain control to correct for attenuation gradients, these corrections lead to artifacts because the true attenuation distribution is not known. The use of constant sound speed leads to additional artifacts that arise from not knowing the sound speed distribution. The authors show that in the context of ultrasound tomography, it is possible to construct reflection images of the breast that correct for inhomogeneous distributions of both sound speed and attenuation.

Abstract

We propose a new method for reconstruction of breast images from measurements obtained by ultrasound tomography (UT) scanners. Our solution for this inverse problem is based on sparse image representation in an overcomplete dictionary that is adapted to the properties of UT images. This dictionary is learned from high resolution MRI breast scans using an unsupervised dictionary learning method described in Ref. [1]. The proposed dictionary-based regularization method significantly improves the quality of reconstructed breast UT images. It outperforms the wavelet-based reconstruction and the l 2 +lowpass minimization algorithm, on both numerical and in vivo data. Our results demonstrate that the use of the learned dictionary improves the image accuracy for up to 4 dB with the exact measurement matrix and for 3.5 dB with the estimated measurement matrix over the wavelet-based reconstruction under the same setup.

Abstract

Our laboratory has focused on the development of ultrasound tomography (UST) for breast imaging. To that end we have been developing and testing a clinical prototype in the Karmanos Cancer Institute's (KCI) breast center. The development of our prototype has been guided by clinical feedback from data accumulated from over 300 patients recruited over the last 4 years. Our techniques generate whole breast reflection images as well as images of the acoustic parameters of sound speed and attenuation. The combination of these images reveals major breast anatomy, including fat, parenchyma, fibrous stroma and masses. Fusion imaging, utilizing thresholding, is shown to visualize mass characterization and facilitates separation of cancer from benign masses. These results indicate that operator-independent whole-breast imaging and the detection and characterization of cancerous breast masses are feasible using acoustic tomography techniques. Analyses of the prototype images suggests that we can detect the variety of mass attributes noted by current ultrasound-BIRADS criteria, such as mass shape, acoustic mass properties and architecture of the tumor environment. These attributes help quantify current BIRADS criteria (e.g. "shadowing" or high attenuation) and provide greater possibilities for defining a unique signature of cancer. The potential for UST to detect and characterize breast masses was quantified using UST measurements of 86 masses from the most recent cohort of patients imaged with the latest version of our prototype. Our preliminary results suggest that the development of a formal predictive model, in support of larger future trials, is warranted.

Abstract

We present preliminary results obtained using a time domain wave-based reconstruction algorithm for an ultrasound transmission tomography scanner with a circular geometry. While a comprehensive description of this type of algorithm has already been given elsewhere, the focus of this work is on some practical issues arising with this approach. In fact, wave-based reconstruction methods suffer from two major drawbacks which limit their application in a practical setting: convergence is difficult to obtain and the computational cost is prohibitive. We address the first problem by appropriate initialization using a ray-based reconstruction. Then, the complexity of the method is reduced by means of an efficient parallel implementation on graphical processing units (GPU). We provide a mathematical derivation of the wave-based method under consideration, describe some details of our implementation and present simulation results obtained with a numerical phantom designed for a breast cancer detection application. The source code of our GPU implementation is freely available on the web at www.usense.org.

Abstract

Breast cancer is the most common type of cancer among women in Europe and North America. The established screening method to detect breast cancer is X-ray mammography, although X-ray frequently provides poor contrast for tumors located within glandular tissue. A new imaging approach is Ultrasound Tomography generating three-dimensional speed of sound images. This paper describes a method to evaluate the clinical applicability of three-dimensional speed of sound images by automatically registering the images with the corresponding X-ray mammograms. The challenge is that X-ray mammograms show two-dimensional projections of a deformed breast whereas speed of sound images render a three-dimensional undeformed breast in prone position. This conflict requires estimating the relation between deformed and undeformed breast and applying the deformation to the three-dimensional speed of sound image. The deformation is simulated based on a biomechanical model using the finite element method. After simulation of the compression, the contours of the X-ray mammogram and the projected speed of sound image overlap congruently. The quality of the matching process was evaluated by measuring the overlap of a lesion marked in both modalities. Using four test datasets, the evaluation of the registration resulted in an average tumor overlap of 97%. The developed registration provides a basis for systematic evaluation of the new modality of three-dimensional speed of sound images, e.g. allows a greater understanding of tumor depiction in these new images.

Abstract

Tomography of complex three-dimensional objects with diffractive waves remains an open challenge due to the large number of scattering measurements required to obtain a stable solution to the inverse problem of reconstructing an image of the object from a set of independent scattering experiments. Here, this problem is addressed with a multiscale approach that is demonstrated experimentally using ultrasonic waves and which leads to high resolution images comparable to x-ray computerized tomography but without the limitations associated with ionizing radiation.

Abstract

Ultrasound is commonly used as an adjunct to mammography for diagnostic evaluation of suspicions arising from breast cancer screening. As an alternative to conventional sonography that uses hand-held transducers, toroidal array probes that encircle the breast immersed in a water bath have been investigated for ultrasound tomography. In this paper, two sets of experiments performed with a prototype ultrasound scanner on a phantom and a human breast in vivo are used to investigate the effects of diffraction and coherence in ultrasound tomography. Reconstructions obtained with transmission diffraction tomography (TDT) are compared with conventional reflection imaging and computerized ultrasound tomography showing a substantial improvement. The in vivo tests demonstrate that TDT can image the complex boundary of a cancer mass and suggest that it can reveal the anatomy of milk ducts and Cooper's ligaments.

Abstract

We discuss a bent-ray ultrasound tomography algorithm with total-variation (TV) regularization. We have applied this algorithm to 61 in vivo breast datasets collected with our in-house clinical prototype for imaging sound-speed distributions in the breast. Our analysis showed that TV regularization could preserve sharper lesion edges than the classic Tikhonov regularization. Furthermore, the image quality of our TV bent-ray sound-speed tomograms was superior to that of the straight-ray counterparts for all types of breasts within BI-RADS density categories 1 through 4. Our analysis showed that the improvements for average sharpness (in the unit of (m · s)⁻¹) of lesion edges in our TV bent-ray tomograms are between 2.1 to 3.4-fold compared with the straight ray tomograms. Reconstructed sound-speed tomograms illustrated that our algorithm could successfully image fatty and glandular tissues within the breast. We calculated the mean sound-speed values for fatty tissue and breast parenchyma as 1422 ± 9 m/s (mean ± SD) and1487 ± 21 m/s, respectively. Based on 32 lesions in a cohort of 61 patients, we also found that the mean sound-speed for malignant breast lesions (1548 ± 17 m/s) was higher, on average, than that of benign ones (1513 ± 27 m/s) (one-sided p < 0.001). These results suggest that, clinically, sound-speed tomograms can be used to assess breast density (and therefore, breast cancer risk), as well as detect and help differentiate breast lesions. Finally, our sound-speed tomograms may also be a useful tool to monitor the clinical response of breast cancer patients to neo-adjuvant chemotherapy.

Abstract

We report on a continuing assessment of the in-vivo performance of an operator independent breast imaging device based on the principles of acoustic tomography. This study highlights the feasibility of mass characterization using criteria derived from reflection, sound speed and attenuation imaging. The data were collected with a clinical prototype at the Karmanos Cancer Institute in Detroit MI from patients recruited at our breast center. Tomographic sets of images were constructed from the data and used to form 3-D image stacks corresponding to the volume of the breast. Masses were identified independently by either ultrasound or biopsy and their locations determined from conventional mammography and ultrasound exams. The nature of the mass and its location were used to assess the feasibility of our prototype to detect and characterize masses in a case-following scenario. Our techniques generated whole breast reflection images as well as images of the acoustic parameters of sound speed and attenuation. The combination of these images reveals major breast anatomy, including fat, parenchyma, fibrous stroma and masses. The three types of images are intrinsically co-registered because the reconstructions are performed using a common data set acquired by the prototype. Fusion imaging, utilizing thresholding, is shown to visualize mass characterization and facilitates separation of cancer from benign masses. These initial results indicate that operator independent whole-breast imaging and the detection and a characterization of cancerous breast masses are feasible using acoustic tomography techniques.

Abstract

Breast ultrasound tomography is a rapidly developing imaging modality that has the potential to impact breast cancer screening and diagnosis. A new ultrasound breast imaging device (CURE) with a ring array of transducers has been designed and built at Karmanos Cancer Institute, which acquires both reflection and transmission ultrasound signals. To extract the sound-speed information from the breast data acquired by CURE, we have developed an iterative sound-speed image reconstruction algorithm for breast ultrasound transmission tomography based on total-variation (TV) minimization. We investigate applicability of the TV tomography algorithm using in vivo ultrasound breast data from 61 patients, and compare the results with those obtained using the Tikhonov regularization method. We demonstrate that, compared to the Tikhonov regularization scheme, the TV regularization method significantly improves image quality, resulting in sound-speed tomography images with sharp (preserved) edges of abnormalities and few artifacts.

Abstract

Multiple tissue parameters are useful for early breast cancer detection and diagnosis. Ultrasound tomography is a new, important imaging modality for our computerized ultrasound risk evaluation (CURE) prototype, which employs a ring transducer array to scan the whole breast in a water tank. We use our bent-ray time-of-light ultrasound tomography to reconstruct sound speeds of breasts with different densities. In addition, we develop an ultrasound attenuation reconstruction method based on complex- signal energy ratios. We apply our reconstruction techniques to in vitro and in vivo breast data acquired using the CURE device. Our reconstruction results demonstrate that CURE has great potential to reliably measure multiple mass characteristics by combining sound-speed, attenuation and reflection images.

Abstract

We report and discuss clinical breast imaging results obtained with operator independent ultrasound tomography. A series of in-vivo experiments were carried out using a recently upgraded clinical prototype based on the principles of ultrasound tomography. The in-vivo performance of the prototype was assessed by imaging patients at the Karmanos Cancer Institute. Our techniques successfully demonstrated in-vivo tomographic imaging of breast architecture in both reflection and transmission imaging modes. Masses as small as 6 mm in size were detected. These initial results indicate that operator- independent whole-breast imaging and the detection of cancerous breast masses are feasible using ultrasound tomography techniques. This approach has the potential to provide a low cost, non-invasive, and non-ionizing means of evaluating breast masses. Future work will concentrate on extending these results to larger trials.

Abstract

Commercial medical ultrasound scanners produce images of human tissue by interpreting the information carried by echoes reflected from structures contained inside it. In this paper, a new generation of toroidal ultrasound arrays is used to measure both the signals reflected and transmitted through the tissue. It is shown that transmission measurements encode complete information about the gross structure of the tissue and lead to images that are superior to those obtained from reflection measurements alone. Experimental results are provided for a gel phantom and a human breast in vivo.

Abstract

Properly accounting for ultrasound scattering from heterogeneities within the breast is essential for high-
resolution ultrasound breast imaging. This requires a reflectivity image reconstruction method capable of accurately handling ultrasound scattering. We develop an optimized ultrasound-wave propagator for reflectivity image reconstruction using pulse-echo ultrasound signals. The method is based on a solution of one-way wave equation and recursive inward continuation of ultrasound wavefields in the frequency-space and frequency-wavenumber domains using a heterogeneous sound-speed model of the breast obtained from tomography. It minimizes ultrasound phase errors during wavefield inward continuation while maintaining the advantage of high computational efficiency. Pulse-echo ultrasound imaging tests for a numerical breast phantom demonstrate that our optimized method has the potential to improve the reliability and accuracy of ultrasound breast imaging.

Abstract

Objective and motivation: Time-of-flight (TOF) tomography used by a clinical ultrasound tomography device can efficiently and reliably produce sound–speed images of the breast for cancer diagnosis. Accurate picking of TOFs of transmitted ultrasound signals is extremely important to ensure high-resolution and high-quality ultrasound sound–speed tomograms. Since manually picking is time-consuming for large datasets, we developed an improved automatic TOF picker based on the Akaike information criterion (AIC), as described in this paper.
Methods: We make use of an approach termed multi-model inference (model averaging), based on the calculated AIC values, to improve the accuracy of TOF picks. By using multi-model inference, our picking method incorporates all the information near the TOF of ultrasound signals. Median filtering and reciprocal pair comparison are also incorporated in our AIC picker to effectively remove outliers.
Results: We validate our AIC picker using synthetic ultrasound waveforms, and demonstrate that our automatic TOF picker can accurately pick TOFs in the presence of random noise with absolute amplitudes up to 80% of the maximum absolute signal amplitude. We apply the new method to 1160 in vivo breast ultrasound waveforms, and compare the picked TOFs with manual picks and amplitude threshold picks. The mean value and standard deviation between our TOF picker and manual picking are 0.4 ls and 0.29 ls, while for amplitude threshold picker the values are 1.02 ls and 0.9 ls, respectively. Tomograms for in vivo breast data with high signal-to-noise ratio (SNR) (25 dB) and low SNR (18 dB) clearly demonstrate that our AIC picker is much less sensitive to the SNRs of the data, compared to the amplitude threshold picker.
Discussion and conclusions: The picking routine developed here is aimed at determining reliable quantitative values, necessary for adding diagnostic information to our clinical ultrasound tomography device – CURE. It has been successfully adopted into CURE, and allows us to generate such values reliably. We demonstrate that in vivo sound–speed tomograms with our TOF picks significantly improve the reconstruction accuracy and reduce image artifacts.

Abstract

We report and discuss clinical breast imaging results obtained with operator independent ultrasound tomography. A series of breast exams are carried out using a recently upgraded clinical prototype designed and built on the principles of ultrasound tomography. The in-vivo performance of the prototype is assessed by imaging patients at the Karmanos Cancer Institute. Our techniques successfully demonstrate in-vivo tomographic imaging of breast architecture in both reflection and transmission imaging modes. These initial results indicate that operator-independent whole-breast imaging and the detection of cancerous breast masses are feasible using ultrasound tomography techniques. This approach has the potential to provide a low cost, non-invasive, and non-ionizing means of evaluating breast masses. Future work will concentrate on extending these results to larger trials.

Abstract

The development of ultrasound tomography for the detection of breast cancer could have a major impact on the effectiveness of current diagnostic tools. Here, the potential of ultrasound tomography is investigated by means of a new generation of toroidal ultrasound arrays that can measure both the signals reflected and transmitted through human breast, simultaneously. Experiments performed on phantoms and human breast in vivo are used to compare continuous wave (CW) insonification versus wideband (WB) excitation. It is shown that while transmission diffraction tomography has little benefit from WB excitation, reflection tomography is greatly improved due to the low signal-to-noise ratio of reflection measurements.

Abstract

A novel clinical prototype, CURE (Computed Ultrasound Risk Evaluation), is used to collect breast tissue image data of patients with either benign or malignant masses. Three types of images, reflection, sound speed and attenuation, are generated from the raw data using tomographic reconstruction algorithms. Each type of image, usually presented as a gray scale image, maps different characteristics of the breast tissue. This study is focused on fusing all three types of images to create true color (RGB) images by assigning a different primary color to each type of image. The resulting fused images display multiple tissue characteristics that can be viewed simultaneously. Preliminary results indicate that it may be possible to characterize breast masses on the basis of viewing the superimposed information. Such a methodology has the potential to dramatically reduce the time required to view all the acquired data and to make a clinical assessment. Since the color scale can be quantified, it may also be possible to segment such images in order to isolate the regions of interest and to ultimately allow automated methods for mass detection and characterization.

Abstract

Ultrasound attenuation parameters of breast masses are closely related to their types and pathological states, therefore, it is essential to reliably estimate attenuation parameters for quantitative breast tissue characterization. We study the applicability of three different attenuation tomography methods for ultrasound breast imaging using a ring transducer array. The first method uses the amplitude decays of signals transmitted through the breast to reconstruct attenuation coefficients. The second method employs the spectral ratios between the pulse propagating through the breast and that through water to obtain attenuation parameters. The third method makes use of the complex energy ratios estimated using the amplitude envelopes of transmitted signals. We use in vitro and in vivo breast data acquired with a clinical ultrasound breast imaging system (CURE) to compare these tomography methods. Our results show that the amplitude decay method yields attenuation coefficients with more artifacts than the other two methods. There is bias and variability in the estimated attenuation using the spectral ratio due to its sensitivity to different temporal band-widths and signal-to-noise-ratios of the data. The method based on the complex signal energy ratio is more robust than the other two methods and yields images with fewer artifacts.

Abstract

To improve clinical breast imaging, a new ultrasound tomography imaging device (CURE) has been built at the Karmanos Cancer Institute. The ring array of the CURE device records ultrasound transmitted and reflected ultrasound signals simultaneously. We develop a bent-ray tomography algorithm for reconstructing the sound-speed distribution of the breast using time-of-flights of transmitted signals. We study the capability of the algorithm using a breast phantom dataset and over 190 patients' data. Examples are presented to demonstrate the sound-speed reconstructions for different breast types from fatty to dense on the BI-RADS categories 1-4. Our reconstructions show that the mean sound-speed value increases from fatty to dense breasts: 1440.8 m/ s (fatty), 1451.9 m/ s (scattered), 1473.2 m/ s(heterogeneous), and 1505.25 m/ s (dense). This is an important clinical implication of our reconstruction. The mean sound speed can be used for breast density analysis. In addition, the sound-speed reconstruction, in combination with attenuation and reflectivity images, has the potential to improve breast-cancer diagnostic imaging. The breast is not compressed and does not move during the ultrasound scan using the CURE device, stacking 2D slices of ultrasound sound-speed tomography images forms a 3D volumetric view of the whole breast. The 3D image can also be projected into a 2-D "ultrasound mammogram" to visually mimic X-ray mammogram without breast compression and ionizing radiation.

Abstract

Ultrasound reflection imaging is a promising imaging modality for detecting small, early-stage breast cancers. Properly accounting for ultrasound scattering from heterogeneities within the breast is essential for high-resolution and high-quality ultrasound breast imaging. We develop a globally optimized Fourier finite-difference method for ultrasound reflectivity image reconstruction. It utilizes an optimized solution of acoustic-wave equation and a heterogeneous sound-speed distribution of the breast obtained from tomography to reconstruct ultrasound reflectivity images. The method contains a finite-difference term in addition to the split-step Fourier implementation, and minimizes ultrasound phase errors during wavefield inward continuation while maintaining the advantage of high computational efficiency. The accuracy analysis indicates that the optimized method is much more accurate than the split-step Fourier method. The computational efficiency of the optimized method is one to two orders of magnitude faster than time-reversal imaging using a finite-difference time-domain wave-equation scheme. Our new optimized method can accurately handle ultrasound scattering from breast heterogeneities during reflectivity image reconstruction. Our numerical imaging examples demonstrate that the optimized method has the potential to produce high-quality and high-resolution ultrasound reflectivity images in combination with a reliable ultrasound sound-speed tomography method.

Abstract

The classical diffraction limit excludes the possibility of resolving features of an object which are spaced less than half a wavelength apart when scattering experiments are performed from the far field. However, recently it has been shown that this limit could be a consequence of the Born approximation that neglects the distortion of the probing wave as it travels through the object to be imaged. Such a distortion, which is due to the multiple scattering phenomenon, can encode unlimited resolution in the radiating component of the scattered field thus leading to super resolution. In this context, a resolution better than λ/3 has been reported in the case of elastic wave probing [F. Simonetti, Phys. Rev. E 73, 036619 (2006)], λ being the wavelength of the wave illuminating the object. This paper demonstrates a resolution better than λ∕4 and approaching λ∕6 for objects immersed in a water bath probed by means of a ring transducer array that excites and detects ultrasonic pressure waves in a full view configuration. This is achieved despite the presence of a high level of noise in the measurements (the signal to noise ratio was below 0dB). Moreover, while previous papers have provided experimental evidence of super resolution for objects small compared to the wavelength, here the case of extended objects is also investigated.

Abstract

This paper investigates the sampling criterion needed to image objects within a circular ring array. The array consists of transducer elements deployed along a circular aperture at regular angular intervals. Each transducer excites waves which propagate towards the center of the array and detects outgoing fields traveling towards it. It is shown that while with conventional linear apertures the sampling criterion is dictated by the wavelength of the probing wave only, in the case of a circular aperture the sampling depends on the size of the object relative to the wavelength and its position with respect to the aperture.

Abstract

Waveform tomography results are presented from 800 kHz ultrasound transmission scans of a breast phantom, and from an in vivo ultrasound breast scan: significant improvements are demonstrated in resolution over time-of-flight reconstructions. Quantitative reconstructions of both sound-speed and inelastic attenuation are recovered. The data were acquired in the Computed Ultrasound Risk Evaluation (CURE) system, comprising a 20 cm diameter solid-state ultrasound ring array with 256 active, non-beamforming transducers. Waveform tomography is capable of resolving variations in acoustic properties at sub-wavelength scales. This was verified through comparison of the breast phantom reconstructions with x-ray CT results: the final images resolve variations in sound speed with a spatial resolution close to 2 mm. Waveform tomography overcomes the resolution limit of time-of-flight methods caused by finite frequency (diffraction) effects. The method is a combination of time-of-flight tomography, and 2-D acoustic waveform inversion of the transmission arrivals in ultrasonic data. For selected frequency components of the waveforms, a finite-difference simulation of the visco-acoustic wave equation is used to compute synthetic data in the current model, and the data residuals are formed by subtraction. The residuals are used in an iterative, gradient-based scheme to update the sound-speed and attenuation model to produce a reduced misfit to the data. Computational efficiency is achieved through the use of time-reversal of the data residuals to construct the model updates. Lower frequencies are used first, to establish the long wavelength components of the image, and higher frequencies are introduced later to provide increased resolution.

Abstract

Sound-speed tomography images can be used for cancer detection and diagnosis. Tumors have generally higher sound speeds than the surrounding tissue. Quality and resolution of tomography images are primarily determined by the insonification/illumination aperture of ultrasound and the capability of the tomography method for accurately handling heterogeneous nature of the breast. We investigate the capability of an efficient time-of-flight tomography method using transmission ultrasound from a ring array for reconstructing sound-speed images of the breast. The method uses first-arrival times of transmitted ultrasonic signals emerging from non-beamforming ultrasound transducers located around a ring. It properly accounts for ray bending within the breast by solving the eikonal equation using a finite-difference scheme. We test and validate the time-of-flight transmission tomography method using synthetic data for numerical breast phantoms containing various objects. In our simulation, the objects are immersed in water within a ring array. Two-dimensional synthetic data are generated using a finite-difference scheme to solve acoustic-wave equation in heterogeneous media. We study the reconstruction accuracy of the tomography method for objects with different sizes and shapes as well as different perturbations from the surrounding medium. In addition, we also address some specific data processing issues related to the tomography. Our tomography results demonstrate that the first-arrival transmission tomography method can accurately reconstruct objects larger than approximately five wavelengths of the incident ultrasound using a ring array.

Abstract

Although mammography is the gold standard for breast imaging, its limitations result in a high rate of biopsies of benign lesions and a significant false negative rate for women with dense breasts. In response to this imaging performance gap we have been developing a clinical breast imaging methodology based on the principles of ultrasound tomography. The Computed Ultrasound Risk Evaluation (CURE) system has been designed with the clinical goals of whole breast, operator-independent imaging, and differentiation of breast masses. This paper describes the first clinical prototype, summarizes our initial image reconstruction techniques, and presents phantom and preliminary in vivo results. In an initial assessment of its in vivo performance, we have examined 50 women with the CURE prototype and obtained the following results. (1) Tomographic imaging of breast architecture is demonstrated in both CURE modes of reflection and transmission imaging. (2) In-plane spatial resolution of 0.5 mm in reflection and 4 mm in transmission is achieved. (3) Masses > 15 mm in size are routinely detected. (4) Reflection, sound speed, and attenuation imaging of breast masses are demonstrated. These initial results indicate that operator-independent, whole-breast imaging and the detection of breast masses are feasible. Future studies will focus on improved detection and differentiation of masses in support of our long-term goal of increasing the specificity of breast exams, thereby reducing the number of biopsies of benign masses.

Abstract

Ultrasonic imaging has the potential to enhance our capability to detect and diagnose breast cancers, but its imaging quality and resolution need to be significantly improved. We make use of the principle of the time-reversal mirror to develop an image-reconstruction method for ultrasonic breast imaging. It reconstructs images of scatterers (e.g., tumors) that generate/scatter ultrasonic waves by backpropagating measured ultrasonic signals into a heterogeneous breast model on computers using the principle of time-reversal mirror. We use solutions of the (two-way) full wave equation and one-way wave equation in heterogeneous media for backpropagation. We found that the one-way wave-equation-based imaging method can produce higher-resolution images than the two-way propagation-based imaging method when the data acquisition aperture is limited (for a linear transducer array). With a full aperture, our imaging results demonstrate that imaging with time-reversed ultrasound can produce high-quality images of the breast.

Abstract

Ultrasound imaging is widely used in medicine because of its benign characteristics and real-time capabilities. Physics theory suggests that the application of tomographic techniques may allow ultrasound imaging to reach its full potential as a diagnostic tool allowing it to compete with other tomographic modalities such as x-ray computer tomography, and MRI. This paper describes the construction and use of a prototype tomographic scanner and reports on the feasibility of implementing tomographic theory in practice and the potential of ultrasound (US) tomography in diagnostic imaging. Data were collected with the prototype by scanning two types of phantoms and a cadaveric breast. A specialized suite of algorithms was developed and utilized to construct images of reflectivity and sound speed from the phantom data. The basic results can be summarized as follows. (i) A fast, clinically relevant US tomography scanner can be built using existing technology. (ii) The spatial resolution, deduced from images of reflectivity, is 0.4 mm. The demonstrated 10 cm depth-of-field is superior to that of conventional ultrasound and the image contrast is improved through the reduction of speckle noise and overall lowering of the noise floor. (iii) Images of acoustic properties such as sound speed suggest that it is possible to measure variations in the sound speed of 5 m/s. An apparent correlation with x-ray attenuation suggests that the sound speed can be used to discriminate between various types of soft tissue. (iv) Ultrasound tomography has the potential to improve diagnostic imaging in relation to breast cancer detection.

Abstract

Reflection imaging has the potential to produce higher-resolution breast images than transmission tomography; however, the current clinical reflection imaging technique yields poor-quality breast images due to speckle. We present a new ultrasonic breast imaging method for obtaining high-resolution and clear breast images using ultrasonic reflection data acquired by a new ultrasonic scanning device that provides a better illumination of targets of interest than the clinical B-scan. The new imaging method is based on the solution of the wave equation in Cartesian coordinates and is implemented using Fast Fourier Transform algorithms. We apply the new ultrasonic breast imaging method to two ultrasonic data sets obtained using an experimental ultrasound scanner recently developed by the Karmanos Cancer Institute. One data set was acquired for a "cyst" phantom using 360 transmitter positions and 321 receiver positions along a 20-cm diameter ring. Another data set was collected with 180 transmitter positions and 1601 receiver positions along a 30-cm diameter ring with the breast specimen located at the center of the ring. We report on the breast imaging results for these two data sets using the new breast imaging method. The results demonstrate that the wave-equation-based ultrasonic breast imaging has the potential to produce high-resolution breast images.

Abstract

The Karmanos Cancer Institute is developing an ultrasound device for measuring and imaging acoustic parameters of human tissue. This paper discusses the experimental results relating to tomographic reconstructions of phantoms and tissue. The specimens were scanned by the prototype scanner at a frequency of 1.5 MHz using 2 microsecond pulses. The receivers and transmitters were positioned along a ring trajectory having a diameter of 20 cm. The ring plane is translated in the vertical direction allowing for 3-D reconstructions from stacked 2-D planes of data. All ultrasound scans were performed at 10 millimeter slice thickness to generate multiple tomographic images. In a previous SPIE paper we presented preliminary results of ultrasound tomographic reconstruction of formalin-fixed breast tissue. We now present new results from data acquired with the scanner. Images were constructed using both reflection-based and transmission based algorithms. The resulting images demonstrate the ability to detect sub-mm features and to measure acoustic properties such as sound speed. Comparison with conventional ultrasound indicates the potential for better margin definition and acoustic characterization of tissue.

Abstract

Extremely high quality data was acquired using an experimental ultrasound scanner developed at Lawrence Livermore National Laboratory using a 2D ring geometry with up to 720 transmitter/receiver transducer positions. This unique geometry allows reflection and transmission modes and transmission imaging and quantification of a 3D volume using 2D slice data. Standard image reconstruction methods were applied to the data including straight-ray filtered back projection, reflection tomography, and diffraction tomography. Newer approaches were also tested such as full wave, full wave adjoint method, bent-ray filtered backprojection, and full-aperture tomography. A variety of data sets were collected including a formalin-fixed human breast tissue sample, a commercial ultrasound complex breast phantom, and cylindrical objects with and without inclusions. The resulting reconstruction quality of the images ranges from poor to excellent. The method and results of this study are described including like-data reconstructions produced by different algorithms with side-by-side image comparisons. Comparisons to medical B-scan and x-ray CT scan images are also shown. Reconstruction methods with respect to image quality using resolution, noise, and quantitative accuracy, and computational efficiency metrics will also be discussed.

Abstract

In contrast to standard reflection ultrasound (US), transmission US holds the promise of more thorough tissue characterization by generating quantitative acoustic parameters. We compare results from a conventional US scanner with data acquired using an experimental circular scanner operating at frequencies of 0.3 - 1.5 MHz. Data were obtained on phantoms and a normal, formalin-fixed, excised breast. Both reflection and transmission-based algorithms were used to generate images of reflectivity, sound speed and attenuation. Images of the phantoms demonstrate the ability to detect sub-mm features and quantify acoustic properties such as sound speed and attenuation. The human breast specimen showed full field evaluation, improved penetration and tissue definition. Comparison with conventional US indicates the potential for better margin definition and acoustic characterization of masses, particularly in the complex scattering environments of human breast tissue. The use of morphology, in the context of reflectivity, sound speed and attenuation, for characterizing tissue, is discussed.

Abstract

Tomographic images of tissue phantoms and a sample of breast tissue have been produced from an acoustic synthetic array system for frequencies near 500 kHz. The images for sound speed and attenuation show millimeter resolution and demonstrate the feasibility of obtaining high-resolution tomographic images with frequencies that can deeply penetrate tissue. The image reconstruction method is based on the Born approximation to acoustic scattering and is a simplified version of a method previously used by Andre (Andre, et. al., Int. J. Imaging Systems and Technology, Vol 8, No. 1, 1997) for a circular acoustic array system. The images have comparable resolution to conventional ultrasound images at much higher frequencies (3-5 MHz) but with lower speckle noise. This shows the potential of low frequency, deeply penetrating, ultrasound for high-resolution quantitative imaging.

Abstract

New ultrasound data, obtained with a circular experimental scanner, are compared with data obtained with standard X-ray CT. Ultrasound data obtained by scanning fixed breast tissue were used to generate images of sound speed and reflectivity. The ultrasound images exhibit approximately 1 mm resolution and about 20 dB of dynamic range. All data were obtained in a circular geometry. X-ray CT scans were used to generate X-ray images corresponding to the same 'slices' obtained with the ultrasound scanner. The good match of sensitivity, resolution and angular coverage between the ultrasound and X-ray data makes possible a direct comparison of the three types of images. We present the results of such a comparison for an excised breast fixed in formalin. The results are presented visually using various types of data fusion. A general correspondence between the sound speed, reflectivity and X-ray morphologies is found. The degree to which data fusion can help characterize tissue is assessed by examining the quantitative correlations between the ultrasound and X-ray images.

Abstract

Mammography is not sufficiently effective for women with dense breast tissue. In North America and Europe, women with dense breasts are at much higher risk for developing breast cancer. Consequently, many breast cancers go undetected at their treatable stage. Improved cancer detection and characterization for women with dense breast tissue is urgently needed. Our clinical study has shown that ultrasound tomography (UST) is an emerging technique that moves beyond B-mode imaging by its transmission capabilities. Transmission ultrasound provides additional tissue parameters such as sound speed, attenuation, and tissue stiffness information. For women with dense breasts, these parameters can be used to assist in detecting malignant masses within glandular or fatty tissue and differentiating malignant and benign masses. This paper focuses on the use of waveform ultrasound sound speed imaging and tissue stiffness information generated using transmission data to characterize different breast tissues and breast masses. In-vivo examples will be given to assess its effectiveness.

Abstract

Ultrasound tomography generates several different imaging stacks. This includes reflection, sound speed, and attenuation images. The images visualize different acoustic parameters which are useful for assessing different types of breast diseases or tissues. Typically, a radiologist views the images to determine a diagnosis for a patient. However, a learning algorithm can be trained to predict diagnoses based on the features contained within the image. Thus, we present a method to extract features from an ultrasound tomography image and label them. The extracted features with the associated label of benign or malignant are fed to a machine learning algorithm which trains a classifier model (the agent). Extracted features from an unlabeled image are then labeled according to the agent. In particular, the differences in tissue acoustic parameters and lesion heterogeneity within the tumor and its surrounding peritumoral region have great diagnostic potential. Ultimately, a radiologist has to work quickly, thus we will also demonstrate that machine learning tools can be used quickly on clinically relevant time scales.

Abstract

Objectives: Ultrasound tomography (UST) is an emerging whole-breast 3-dimensional imaging technique that obtains quantitative tomograms of speed of sound of the entire breast. The imaged parameter is the speed of sound which is used as a surrogate measure of density at each voxel and holds promise as a method to evaluate breast density without ionizing radiation. This study evaluated the technique of UST and compared whole-breast volume averaged speed of sound (VASS) with MR percent water content from noncontrast magnetic resonance imaging (MRI).

Materials and Methods: Forty-three healthy female volunteers (median age, 40 years; range, 29–59 years) underwent bilateral breast UST and MRI using a 2-point Dixon technique. Reproducibility of VASS was evaluated using Bland-Altman analysis. Volume averaged speed of sound and MR percent water were evaluated and compared using Pearson correlation coefficient.

Results: The mean ± standard deviation VASS measurement was 1463 ± 29ms−1 (range, 1434–1542ms−1). There was high similarity between right (1464 ± 30ms−1) and left (1462 ± 28ms−1) breasts (P = 0.113) (intraclass correlation coefficient, 0.98). Mean MR percent water content was 35.7% ± 14.7% (range, 13.2%–75.3%), with small but significant differences between right and left breasts (36.3% ± 14.9% and 35.1% ± 14.7%, respectively; P = 0.004). There was a very strong correlation between VASS and MR percent water density (r2 = 0.96, P < 0.0001).

Conclusions: Ultrasound tomography holds promise as a reliable and reproducible 3-dimensional technique to provide a surrogatemeasure of breast density and correlates strongly with MR percent water content.

Abstract

There has been a major need to better understand the biological characteristics of triple-negative breast cancers. Compared with estrogen receptor (ER)-positive cancers, several magnetic resonance (MR) imaging findings have been reported as characteristic findings. However, information regarding their location has not been described. Our study was to compare the location of triple-negative breast cancers with that of ER-positive breast cancers using magnetic resonance (MR) imaging. The locations of 1102 primary breast cancers (256 triple-negative and 846 ER-positive) in 1090 women (mean, 52.1 years) were reviewed using three-dimensional (3D) coordinates. The x-axis measurement was recorded as the transverse distance from the posterior nipple line; y-axis measurement as the anteroposterior distance from the chest wall; z-axis measurement as the superoinferior distance from the posterior nipple line. The association between breast cancer subtype and tumor location was evaluated using multiple linear regression analysis. Triple-negative breast cancers were significantly closer to the chest wall than ER-positive breast cancers in absolute (1.8 cm vs. 2.3 cm, P < .0001) and normalized (0.21 vs. 0.25, P < .0001) y-axis distances. The x- and z-axes distances were not significantly different between triple-negative and ER-positive breast cancers. Multiple linear regression analysis revealed that age, mammographic density, axillary nodal status, and triple-negative subtype were significantly associated with absolute and normalized distances from the chest wall (all P < .05). Our results show that triple-negative breast cancers have a tendency toward a posterior or prepectoral location compared with ER-positive breast cancers.

Abstract

Objective: To determine the clinical display thresholds of an ultrasound tomography (UST) prototype relative to magnetic resonance (MR) for comparable visualization of breast anatomy and tumor rendering.
Materials and Methods: The study was compliant with HIPAA, approved by the IRB, and performed after obtaining informed consent. Thirty-six women were imaged with MR and our UST prototype. The UST scan generated reflection, sound speed and attenuation images. The reflection images were fused with the components of sound speed and attenuation images that achieved thresholds to represent parenchyma and/or solid masses using an image arithmetic process. Qualitative and quantitative comparisons of MR and UST clinical images were used to identify anatomical similarities, and optimized thresholds for tumor shapes and volumes.
Results: Thresholding techniques generated UST images comparable to MR for visualizing fibrous stroma, parenchyma, fatty tissues, and tumors, of which 25 were cancer and 11 benign. Optimized sound speed thresholds of 1.46±0.1 km/s and 1.52±0.03 km/s were identified to best represent the extent of fibroglandular tissue and solid masses, respectively. An arithmetic combination of attenuation images using the threshold of 0.16±0.04 dB/cm further characterized benign from malignant masses. No significant difference in tumor volume was noted between benign or malignant masses by UST or MR (p>0.1) using these universal thresholds.
Conclusion: UST demonstrated the ability to image and render breast tissues in a manner comparable to MR. Universal UST threshold values appear feasible for rendering of the size and distribution of benign and malignant tissues without intravenous contrast.

Abstract

OBJECTIVE: The objective of our study was to determine the clinical display thresholds of an ultrasound tomography prototype relative to MRI for comparable visualization of breast anatomy and tumor rendering.

SUBJECTS AND METHODS: Thirty-six women were imaged with MRI and our ultrasound tomography prototype. The ultrasound tomography scan generated reflection, sound-speed, and attenuation images. The reflection images were fused with the components of the sound-speed and attenuation images that achieved thresholds to represent parenchyma or solid masses using an image arithmetic process. Qualitative and quantitative comparisons of MRI and ultrasound tomography clinical images were used to identify anatomic similarities and optimized thresholds for tumor shapes and volumes.

RESULTS: Thresholding techniques generated ultrasound tomography images comparable to MR images for visualizing fibrous stroma, parenchyma, fatty tissues, and tumors. In 25 patients, tumors were cancerous and in 11, benign. Optimized sound-speed thresholds of 1.46 ± 0.1 and 1.52 ± 0.03 km/s were identified to best represent the extent of fibroglandular tissue and solid masses, respectively. An arithmetic combination of attenuation images using a threshold of 0.16 ± 0.04 dB/cm (mean ± SD) further characterized benign from malignant masses. No significant difference in tumor volume was noted between benign or malignant masses by ultrasound tomography or MRI (p > 0.1) using these universal thresholds.

CONCLUSION: Ultrasound tomography is able to image and render breast tissues in a manner comparable to MRI. Using universal ultrasound tomography threshold values for rendering the size and distribution of benign and malignant tissues appears feasible without IV contrast material.

Abstract

The objective of this study is to present imaging parameters and display thresholds of an ultrasound tomography (UST) prototype in order to demonstrate analogous visualization of overall breast anatomy and lesions relative to magnetic resonance (MR). Thirty-six women were imaged with MR and our UST prototype. The UST scan generated sound speed, attenuation, and reflection images and were subjected to variable thresholds then fused together into a single UST image. Qualitative and quantitative comparisons of MR and UST images were utilized to identify anatomical similarities and mass characteristics. Overall, UST demonstrated the ability to visualize and characterize breast tissues in a manner comparable to MR without the use of IV contrast. For optimal visualization, fused images utilized thresholds of 1.46±0.1 km/s for sound speed to represent architectural features of the breast including parenchyma. An arithmetic combination of images using the logical .AND. and .OR. operators, along with thresholds of 1.52±0.03 km/s for sound speed and 0.16±0.04 dB/cm for attenuation, allowed for mass detection and characterization similar to MR.

Abstract

The purpose of this study was to investigate the performance of an ultrasound tomography (UST) prototype relative to magnetic resonance (MR) for imaging overall breast anatomy and accentuating tumors relative to background tissue. The study was HIPAA compliant, approved by the Institutional Review Board, and performed after obtaining the requisite informed consent. Twenty-three patients were imaged with MR and the UST prototype. T1 weighted images with fat saturation, with and without gadolinium enhancement, were used to examine anatomical structures and tumors, while T2 weighted images were used to identify cysts. The UST scans generated sound speed, attenuation, and reflection images. A qualitative visual comparison of the MRI and UST images was then used to identify anatomical similarities. A more focused approach that involved a comparison of reported masses, lesion volumes, and breast density was used to quantify the findings from the visual assessment. Our acoustic tomography prototype imaged distributions of fibrous stroma, parenchyma, fatty tissues, and lesions in patterns similar to those seen in the MR images. The range of thresholds required to establish tumor volume equivalency between MRI and UST suggested that a universal threshold for isolating masses relative to background tissue is feasible with UST. UST has demonstrated the ability to visualize and characterize breast tissues in a manner comparable to MRI. Thresholding techniques accentuate masses relative to background anatomy, which may prove clinically useful for early cancer detection.

Abstract

The objective of this study is to investigate a potential low-cost alternative to MRI, based on acoustic tomography. Using MRI as the gold standard, our goals are to assess the performance of acoustic tomography in (i) depicting normal breast anatomy, (ii) imaging cancerous lesions and (iii) accentuating lesions relative to background tissue using thresholding techniques. Fifteen patients were imaged with MRI and with an acoustic tomography prototype. A qualitative visual comparison of the MRI and prototype images was used to verify anatomical similarities. These similarities suggest that the prototype can image fibrous stroma, parenchyma and fatty tissues, with similar sensitivity to MRI. The prototype was also shown to be able to image masses but equivalency in mass sensitivity with MRI could not be established because of the small numbers of patients and the prototype's limited scanning range. The range of thresholds required to establish tumor volume equivalency suggests that a universal threshold for isolating masses relative to background tissue is possible with acoustic tomography. Thresholding techniques promise to accentuate masses relative to background anatomy which may prove clinically useful in potential screening applications. Future work will utilize larger trials to verify these preliminary conclusions.

Abstract

Ultrasound tomography (UST) is an emerging modality that can offer quantitative measurements of breast density. Recent breakthroughs in UST image reconstruction involve the use of a waveform reconstruction as opposed to a raybased reconstruction. The sound speed (SS) images that are created using the waveform reconstruction have a much higher image quality. These waveform images offer improved resolution and contrasts between regions of dense and fatty tissues. As part of a study that was designed to assess breast density changes using UST sound speed imaging among women undergoing tamoxifen therapy, UST waveform sound speed images were then reconstructed for a subset of participants. These initial results show that changes to the parenchymal tissue can more clearly be visualized when using the waveform sound speed images. Additional quantitative testing of the waveform images was also started to test the hypothesis that waveform sound speed images are a more robust measure of breast density than ray-based reconstructions. Further analysis is still needed to better understand how tamoxifen affects breast tissue.

Abstract

Women with high breast density have an increased risk of developing breast cancer. Women treated with the selective estrogen receptor modulator tamoxifen for estrogen receptor positive breast cancer experience a 50% reduction in risk of contralateral breast cancer and overall reduction of similar magnitude has been identified among high-risk women receiving the drug for prevention. Tamoxifen has been shown to reduce mammographic density, and in the IBIS-1 chemoprevention trial, risk reduction and decline in density were significantly associated. Ultrasound tomography (UST) is an imaging modality that can create tomographic sound speed images of the breast. These sound speed images are useful because breast density is proportional to sound speed. The aim of this work is to examine the relationship between UST-measured breast density and the use of tamoxifen. So far, preliminary results for a small number of patients have been observed and are promising. Correlations between the UST-measured density and mammographic density are strong and positive, while relationships between UST density with some patient specific risk factors behave as expected. Initial results of UST examinations of tamoxifen treated patients show that approximately 45% of the patients have a decrease in density in the contralateral breast after only several months of treatment. The true effect of tamoxifen on UST-measured density cannot yet be fully determined until more data are collected. However, these promising results suggest that UST can be used to reliably assess quantitative changes in breast density over short intervals and therefore suggest that UST may enable rapid assessment of density changes associated with therapeutic and preventative interventions.

Abstract

Objective: To evaluate the combination of tumor volume and sound speed as a potential imaging marker for assessing neoadjuvant chemotherapy (NAC) response.

Methods: This study was carried out under an IRB-approved protocol (written consent required). Fourteen patients undergoing NAC for invasive breast cancer were examined with ultrasound tomography (UST) throughout their treatment. The volume (V) and the volume-averaged sound speed (VASS) of the tumors and their changes were measured for each patient. Time-dependent response curves of V and VASS were constructed individually for each patient and then as averages for the complete versus partial response groups in order to characterize differences between the two groups. Differences in group means were assessed for statistical significance using t-tests. Differences in shapes of group curves were evaluated with Kolmogorov–Smirnoff tests.

Results: On average, tumor volume and sound speed in the partial response group showed a gradual decline in the first 60 days of treatment, while the complete response group showed a much steeper decline (P < 0.05). The shapes of the response curves of the two groups, corresponding to the entire treatment period, were also found to be significantly different (P < 0.05). Furthermore, large simultaneous drops in volume and sound speed in the first 3 weeks of treatment were characteristic only of the complete responders (P < 0.05).

Conclusion: This study demonstrates the feasibility of using UST to monitor NAC response, warranting future studies to better define the potential of UST for noninvasive, rapid identification of partial versus complete responders in women undergoing NAC.

Abstract

As part of an ongoing assessment of the in-vivo performance of a operator independent breast imaging device, based on acoustic tomography, we report on new results obtained with patients undergoing neoadjuvant chemotherapy. Five patients were examined with the prototype on multiple occasions corresponding in time to their chemotherapy sessions. Images of reflection, sound speed and attenuation, representing the entire volume of the breast, were reconstructed from the exam data and analyzed for time-dependent changes during the treatment period. It was found that changes in acoustic properties of the tumors could be measured directly from the images. The measured properties include reflectivity, sound speed and attenuation, leading to measurable changes in the volume, shape and internal attributes of the tumors. These measurements were used to monitor the response of the tumors to the therapy with the long term goal of correlating results with pathological and clinical outcomes. Comparisons with tumor size changes based on traditional US and MRI indicates potential for accurate, quantifiable tracking of tumor volume. Furthermore, our tentative results also show declines in internal properties of the tumors, possibly relating to a reduction in tissue stiffness and/or density. Future work will include an expansion of the study to a larger cohort of patients for determining the statistical significance of our findings.