The basic imaging paradigm of medical US has remained unchanged, namely image formation results from transmission‐reception of 180 degree backscatter assuming straight‐line propagation. Our group has developed Quantitative Ultrasound CT (USCT) that creates images of the breast from the full 3D scatter field including both transmitted and reflected signals. While modern US provides a two‐dimensional map of relative tissue “echogenicity,” USCT computes quantitative 3D maps of tissue acoustic properties including sound speed, attenuation, scatter density, etc., as approximations of components of the wave equation. The method corrects for refraction, absorption, multiple scattering and more. High resolution reflection tomography is performed in which the sound speed and attenuation maps are utilized for aberration correction to significantly improve image quality. Clinical measurements suggest that in breast tissue, benign lesions and cancerous lesions may be identified by these inherent acoustic. The scanner uses a multi‐frequency non‐linear 3D inverse‐ scattering algorithm and despite the historical computational complexity of the problem, our method is fast and practical. Conventional breast sonography is a notoriously difficult exam to perform; the quality is dependent on the skill of the operator as well as technical features of the scanner. In order to obtain the needed high resolution, the field of view in sonography is very small, which greatly complicates interpretation and localization of masses. USCT promises an automated whole‐breast scan providing a global view of the entire breast in 3D, facilitating comparison to prior exams in a reproducible geometry. Results of our trial with over 150 subjects with confirmed breast masses will be presented with detailed comparison to conventional sonography and MRI.Learning Objectives:1. Understand the principles of inverse‐scatter tomography.2. Understand the unique data acquisition and processing applied in ultrasound computed tomography.3. Understand the potential clinical advantages of quantitative whole breast ultrasound imaging.

Purpose: To establish the ability of our 3D registered inverse scattering based breast scanner to provide highly reproducible quantitative monitoring of breast health over short and long interval follow up, and to distinguish between malignant and benign lesions using quantitative estimates of speed of sound. Methods: Our group has developed a unique method of 3D ultrasound imaging based on the mathematical technique known as inverse scattering, which yields a volumetric map of the speed and attenuation of the breast. Resolution and quantitative accuracy is assessed using known phantoms. We scanned a group of 18 volunteers twice at the UCSD facility, first with a 960 element array and secondly with a 1536 array upgrade. We also scanned 2 volunteers ∼25 times over a period of 4 years. We used an earlier published result of Mast to correlate the speed values to bulk modulus in breast tissue. We correlated slices based on ‘landmarks’ observed in the ductal tissue pattern and noted that this ductal distribution was unique to each patient. We used the speed and attenuation maps for refraction and gain correction in ray based reflection imaging, allowing a 360 degree compounding of the reflection. Results: We are able to validate short and long term stability of the image as well as validate our ability to follow the change in tissue and mass characteristics. We observed a significant improvement of the speed image by using the 1536 element 3D array. Conclusions: We determined that the quantitative values of the speeds (bulk moduli) remained substantially constant over approximately 6 month periods for the UCSD cases, and several years for the TechniScan, Inc. study. These results are tabulated in the study. We determined that there is a significant difference between the speed of sound (bulk modulus) of malignant and benign lesions. TechniScan, Inc., Salt Lake City

We discuss the results obtained using the inversion and refraction corrected reflection (RFCR) algorithms described in the companion paper in these proceedings. We show images for three patients created with these algorithms, from data collected with our clinical device. We discuss their potential clinical relevance, and their relationship to other, more conventional imaging modalities.

A new ultrasound CT breast WBU (Warm Bath Ultrasound) Scanner developed by Techniscan Medical Systems, Salt Lake City, Utah, USA, and supported by Esaote Spa, Genoa, Italy, was installed for pre-clinical testing at the Department of Radiology, Clinical Radiology, University Medical Center Freiburg, Freiburg, Germany. The system acquires a volume dataset of both transmission and reflection of the whole breast. A 3D inverse scattering algorithm generates 3D maps of both speed of sound and attenuation of tissues, along with a 3D reflection map geometrically registered with the previous ones. The pilot study examination protocol included breast WBU, handheld ultrasound and mammography for every patient. MRI was used when recommended for diagnostic purposes. In cases of suspected malignancy, vacuum core biopsy was conducted. A total of 60 patients were included in the study. As result, 10 out of 12 carcinomas were detected in the WBU examinations. In 2 cases, carcinoma close to the chest wall laid outside the field of view. Fibroadenomas and cysts were depicted reliably. Lesion characteristics differ markedly from handheld ultrasound. Speed of sound and attenuation images may increase the lesions characterization reliability.

This paper discuss a fully 3D nonlinear algorithm that results in a 3D quantitative estimate of breast tissue characteristics and a refraction corrected reflection algorithm (RFCR) that utilizes these estimates. The data are obtained from a specially designed clinical ultrasound breast scanner and processed on the device. We discuss the data collection process, a fast solution to the forward problem and a concomitant fast inverse scattering solution for the imaging problem. We show how the resulting 3D tissue map is used in a refraction corrected reflection algorithm.

Reflection ultrasound (US) has been utilized as an adjunct imaging modality for over 30 years. TechniScan, Inc. has developed unique, transmission and concomitant reflection algorithms which are used to reconstruct images from data gathered during a tomographic breast scanning process called Warm Bath Ultrasound (WBU™). The transmission algorithm yields high resolution, 3D, attenuation and speed of sound (SOS) images. The reflection algorithm is based on canonical ray tracing utilizing refraction correction via the SOS and attenuation reconstructions. The refraction correction reflection algorithm allows 360 degree compounding resulting in the reflection image. The requisite data are collected when scanning the entire breast in a 33° C water bath, on average in 8 minutes. This presentation explains how the data are collected and processed by the 3D transmission and reflection imaging mode algorithms. The processing is carried out using two NVIDIA® Tesla™ GPU processors, accessing data on a 4-TeraByte RAID. The WBU™ images are displayed in a DICOM viewer that allows registration of all three modalities. Several representative cases are presented to demonstrate potential diagnostic capability including: a cyst, fibroadenoma, and a carcinoma. WBU™ images (SOS, attenuation, and reflection modalities) are shown along with their respective mammograms and standard ultrasound images. In addition, anatomical studies are shown comparing WBU™ images and MRI images of a cadaver breast. This innovative technology is designed to provide additional tools in the armamentarium for diagnosis of breast disease.

A new transmission ultrasound CT breast scanner (Techniscan Medical Systems, Inc.) was installed for pre-clinical testing at UCSD Medical Center. The scanner utilizes a 3D inverse scattering method to produce whole-breast tomographic images with resolution approximately 1.5 mm in plane, 3.5 mm slice profile and slice spacing of 1 mm. Sound speed accuracy and sensitivity were found to be highly linear (R2=0.99) over the wide range of 1370–1620 m/sec. Attenuation provided a wide image contrast and is able to localize and identify breast lesions. We present representative cases of human subjects enrolled in the pre-clinical study and describe future plans for the system.

Purpose: To present the design and performance of a new automated whole‐breast ultrasound computed tomography (USCT) system that employs 3D inverse‐scatter reconstruction. Method and Materials: The scanner (Techniscan Medical Systems, Inc., Salt Lake City, Utah) is installed at the University of California, San Diego to evaluate clinical performance including ability to detect and analyze breast masses. Patients lie prone on a table while opposing transmitter and receiver transducer arrays rotate 360° around the breast. Ultrasound plane waves (300 kHz–2MHz) are emitted every 2 degrees while the scattered beam is detected by a 960‐element six‐row transducer. In the same plane three B‐mode linear arrays acquire backscatter data from wideband signals (3.6–8.4 MHz). The arrays move up the breast in 2mm increments to scan the entire breast. Discrete frequency domain data is used by a proprietary 3D inverse‐scattering algorithm that incorporates multiple scattering within and between the planes. 2‐D coronal images of the entire breast are reconstructed as accurate quantitative maps of sound speed and attenuation along with aberration‐corrected high‐resolution reflection tomograms. Results: To date, more than 50 subjects were scanned with wide range in age (20–78), breast density and diagnostic outcome. Representative cases will be presented comparing mammography, sonography and multi‐planar USCT images. Included are benigns (cysts, fibroadenomas, fibrocystic disease) and biopsy‐proven malignancies (invasive ductal carcinomas, invasive tubular carcinoma and mixed lobular/ductal carcinoma). System performance and tissue characterization are excellent. Accuracy and linearity of sound speed measurements by USCT is very high (R2=0.988) over the range of 1400 to 1600 m/sec. Conclusion: The USCT system provides rapid, automated scanning with quantitative 3D images of the breast and substantially new information for characterizing breast masses. Conflict of Interest: Research sponsored in part by Techniscan Medical Systems, Inc.

TechniScan Medical Systems, Inc. is using novel inverse scattering methods to provide a unique method for calculating ultrasound characteristics of speed and attenuation of sound traveling through human tissue. In this paper we describe basic system parameters and results of the first in vivo patient studies. It is concluded that this novel inverse scattering method provides a unique method for noninvasive breast tissue characterization that could assist physicians in assigning probability of cancer to breast abnormalities identified but not resolved with currently available imaging techniques. This application could have a major impact on patient management decisions and has the potential to reduce the number of currently unavoidable breast biopsies that result in a benign outcome.

We present a method to incorporate the relaxation dominated attenuation into the finite-difference time-domain (FDTD) simulation of acoustic wave propagation in complex media. A dispersive perfectly matched layer (DPML) boundary condition, which is suitable for boundary matching to such a dispersive media whole space, is also proposed to truncate the FDTD simulation domain. The numerical simulation of a Ricker wavelet propagating in a dispersive medium, described by second-order Debye model, shows that the Ricker wavelet is attenuated in amplitude and expanded in time in its course of propagation, as required by Kramers-Kronig relations. The numerical results also are compared to exact solution showing that the dispersive FDTD method is accurate and that the DPML boundary condition effectively dampens reflective waves. The method presented here is applicable to the simulation of ultrasonic instrumentation for medical imaging and other nondestructive testing problems with frequency dependent, attenuating media.

Inverse scattering algorithms for reconstructing the physical properties of sea ice from scattered electromagnetic field data are presented. The development of these algorithms has advanced the theory of remote sensing, particularly in the microwave region, and has the potential to form the basis for a new generation of techniques for recovering sea ice properties, such as ice thickness, a parameter of geophysical and climatological importance. Moreover, the analysis underlying the algorithms has led to significant advances in the mathematical theory of inverse problems. In particular, the principal results include the following. (1) Inverse algorithms for reconstructing the complex permittivity in the Helmholtz equation in one and higher dimensions, based on layer stripping and nonlinear optimization, have been obtained and successfully applied to a (lossless) laboratory system. In one dimension, causality has been imposed to obtain stability of the solution and layer thicknesses can be obtained from the recovered dielectric profile, or directly from the reflection data through a nonlinear generalization of the Paley-Wiener theorem in Fourier analysis. (2) When the wavelength is much larger than the microstructural scale, the above algorithms reconstruct a profile of the effective complex permittivity of the sea ice, a composite of pure ice with random brine and air inclusions. A theory of inverse homogenization has been developed, which in this quasistatic regime, further inverts the reconstructed permittivities for microstructural information beyond the resolution of the wave. Rigorous bounds on brine volume and inclusion separation for a given value of the effective complex permittivity have been obtained as well as an accurate algorithm for reconstructing the brine volume from a set of values. (3) Inverse algorithms designed to recover sea ice thickness have been developed. A coupled radiative transfer-thermodynamic sea ice inverse model has accurately reconstructed the growth of a thin, artificial sea ice sheet from time-series electromagnetic scattering data.

Berenger's perfectly matched layer (PML) absorbing boundary condition for electromagnetic (EM) waves is derived to absorb 2-D and 3-D acoustic waves in finite difference time domain (FDTD) simulation of acoustic wave propagation and scattering. A PML medium suitable for acoustic waves is constructed. Plane wave propagation in the PML medium is solved for both 2-D and 3-D cases and explicit FDTD boundary conditions are derived. The equations show that a matched PML medium is a perfect simulation of free space in that a plane wave does not change its direction of propagation or its speed when it propagates from free space into a matched PML medium. FDTD simulation of a pulsed point source propagating in two dimensions is carried out to test the performance of the PML boundary for acoustic waves. Results show that an eight layer PML boundary condition reduces the reflected error 40 dB over Mur's second order boundary condition.