International Workshop on Medical Ultrasound Tomography

Nov 1, 2017, 8:30 AM → Nov 3, 2017, 6:00 PM Europe/Berlin

Wed, November 1

Opening + Introduction talk

Conveners: Nicole Ruiter (Karlsruhe Institute of Technology), Torsten Hopp (Karlsruhe Institute of Technology)

1 Opening ceremony

Speakers: Jeffrey Bamber (The Institute of Cancer Research), Nicole Ruiter (Karlsruhe Institute of Technology), Torsten Hopp (Karlsruhe Institute of Technology)

2 Current challenges in breast screening and diagnosis

Speaker: Peter Littrup (Karmanos Cancer Institute)

Session 1:Imaging and inversion I

Conveners: Andreas Fichtner (ETH Zurich), Jürgen Hesser (Heidelberg University)

3 One-dimensional Marchenko inversion in stretched space

In acoustic inversion, we aim to retrieve the density and compressibility of an unknown medium from single-sided reflection data by solving a nonlinear integral equation. If the wavefield is known throughout the medium, this equation simplifies as a linear map between the medium properties and the reflection data. We aim to realize such a map by substituting the solution of a Marchenko equation into an integral representation in stretched space.

Speaker: Joost van der Neut (Delft University of Technology)

4 Ultrasound imaging from reflection data

When only reflection data are available, imaging with the wave equation is much more difficult then in the presence of transmission measurements. This is because reflection data are spectrally incomplete due to fact that frequencies close to zero can not be used. In the talk we describe the mathematical problem and suggest a numerical method of data completion that generates the missing low frequency data by a purely mathematical procedure.

Speaker: Frank Natterer (University of Münster)

5 Using 2-D approximation of the 3-D incident field for Born inversion

In a real ultrasound measurement system, acoustic waves show 3-D propagation characteristics. However, in a majority of recently developed wave based imaging methods, the inverse problem is solved in a 2-D framework due to computational restrictions. This reduction causes problems during inversion. In this work, we use a 2-D approximation for the measured 3-D incident field and show that using this approximated incident field for Born inversion (BI) improves the resulting reconstruction.

Speaker: Ulas Taskin (Delft University of Technology)

10:30 AM Coffee break

Session 2: Imaging and inversion II

Conveners: Frank Natterer (University of Münster), Joost van der Neut (Delft University of Technology)

6 USCT Image Reconstruction: Acceleration using Gauss-Newton Preconditioned Conjugate Gradient

Ultrasound transmission tomography offers quantitative characterization of the tissue or materials by their speed of sound and attenuation. Reconstruction of such images is an inverse problem which is solved iteratively based on a forward model of paraxial approximation of the Helmholtz equation and thus is time-consuming. Hence, developing optimizers that decrease this time, in particular reducing the number of forward propagations is important for clinical practice.

Speaker: Hongjian Wang (Heidelberg University)

7 3D time-domain spectral-element ultrasound waveform tomography

Waveform inversion for ultrasound computed tomography is a promising tool to image the acoustic properties of breast tissue. We present a technique for time-domain 3D acoustic waveform inversion that combines the spectral-element method with source encoding strategies and a quasi-Newton trust-region method. The key objective is to reduce the computational cost of waveform inversion by randomly encoding the emitters in space and by introducing time lags based on the acquisition geometry.

Speaker: Christian Boehm (ETH Zurich)

8 Joint reconstruction of ultrasound and photoacoustic tomography images

Photoacoustic computed tomography (PACT) is a bioimaging modality that seeks to reconstruct an estimate of the absorbed optical energy density within an object. We propose a joint reconstruction problem in which the speed-of-sound (SOS) distribution is concurrently estimated along with the sought-after absorbed optical energy density from the photoacoustic measurement data.

Speaker: Mark Anastasio (Washington University in St. Louis)

12:00 PM Lunch

Session 3: Imaging and inversion III

Conveners: Neb Duric (Delphinus Medical Technologies and Karmanos Cancer Institute), Peter Huthwaite (Imperial College London)

9 Medical ultrasound tomography: lessons from exploration geophysics

Speaker: Gerhard Pratt (University of Western Ontario)

10 Non-linear Ultrasonic Computed Tomography (USCT) for soft and hard tissue imaging

This paper presents the theoretical, numerical and experimental frameworks for breast and bone imaging. The difficulties raised are somewhat different as, in soft tissues, the very small fluctuations to be quantified suffer from their very low values. This poor echogenic index generally induces low detection probability, for instance in the case of large diffuse masses. In bone imaging, the difficulties arise from the very high contrast that alters the propagation of the ultrasonic waves.

Speaker: Philippe Lasaygues (Aix Marseille Univ, CNRS, Centrale Marseille, Laboratory of Mechanics and Acoustics, Marseille, France)

11 Scattering Computed Tomography visualizing density distributions by back-propagation

Ultrasound computed tomography (UCT) has various imaging modes; sound speed profile, attenuation profile and gradient of acoustic impedance. In these imaging modes, ultrasound scattering tomography gives high resolution images highly relating to anatomical structure. In order to visualize density distribution as new efficient information in diagnoses, we use echo scattering signals and information of sound speed profile obtained by UCT.

Speaker: Daisuke Kondo (The University of Tokyo)

2:40 PM Coffee break

Session 4: Imaging and inversion IV

Conveners: Koen van Dongen (Delft University of Technology), Mark Anastasio (Washington University in St. Louis)

12 Multiscale full-waveform inversion: From the whole earth to the human breast

Speaker: Andreas Fichtner (Swiss Federal Institute of Technology (ETH) Zürich)

13 Real-Time Ultrasound Transmission Tomography based on Bézier Curves

Refraction-corrected ray-tracing methods in Ultrasound Transmission Tomography (USTT) are able to reconstruct the acoustic properties of tissues with good accuracy when using medium and high frequencies (>3MHz). However, these techniques, such as the Fast Marching Method (FMM), are time-consuming. In this work, we propose a real-time reconstruction methodology that exploits the large computing capabilities of current GPUs.

Speaker: Joaquin L. Herraiz (University Complutense of Madrid)

14 High resolution breast ultrasound tomography with HARBUT

Breast ultrasound tomography can produce a map of acoustic properties through the breast, greatly increasing the potential for improved cancer diagnosis. The use of ultrasound is cheaper than MRI and safer than traditional mammography. One particular challenge seen in breast ultrasound tomography is the low resolution caused by ray behaviour commonly assumed in most reconstruction algorithms; by neglecting diffraction effects the algorithm is unable to reliably image small scale features.

Speaker: Peter Huthwaite (Imperial College London)

15 3D imaging of the breast using full-waveform inversion

Full-waveform inversion (FWI) is an iterative algorithm that produces high-quality images from experimental data. The petroleum industry has invested large sums to make this method more efficient, robust and reliable, especially for complex and large 3D datasets. However, its application to medical imaging, especially in 3D, is still in its infancy. Our objective is to efficiently implement FWI to invert 3D ultrasound datasets and study the implications of using low-frequency transducers.

Speaker: Oscar Calderon Agudo (Imperial College London)

Discussion session: Imaging and inversion

Convener: Koen van Dongen (Delft University of Technology)

Thu, November 2

Session 5: Ultrasound tomography systems

Conveners: Jeffrey Bamber (Institute of Cancer Research), Takashi Azuma (The University of Tokyo)

16 Ultrasound Computed Tomography: Historically Guided Musings

Speaker: James F. Greenleaf (Mayo Clinic)

17 A Multi-Modal Ultrasound Breast Imaging System

The Multimodal Ultrasound Breast Imaging System (MUBI) is a joint development of the Spanish National Research Council (CISC) and the Complutense University of Madrid (UCM). It is intended to be a flexible platform for multi-modal ultrasound imaging research, mainly oriented to breast diagnosis. Up to now, the following imaging techniques have been implemented: Phased-Array full angle spatial compound (FASC), Acoustic Radiation Force Imaging (ARFI) and Ultrasound Computed Tomography (USCT).

Speaker: Jorge Camacho (Spanish National Research Council (CSIC))

18 First steps towards the Delft Breast Ultrasound Scanning System (DBUS)

Ultrasound is gaining interest as a modality to scan for tumors in dense breasts of young women. Tissue characterisation by full wave form inversion (FWI) has the potential to improve the accuracy of breast ultrasound. However, many of the existing systems use frequencies above 1 MHz. With these high frequencies it is difficult to apply FWI due to the computational costs involved. Consequently, we investigate the applicability of low frequencies together with FWI for breast cancer detection.

Speaker: Koen van Dongen (Delft University of Technology)

19 Hybrid Optoacoustic and Ultrasound Imaging System with a Multi-Segment Detector Array

The high complementarity of the optoacoustic (OA) and pulse-echo ultrasound (US) modalities makes the combined usage of these imaging technologies highly advantageous. Yet, due to the different physical contrast mechanisms, the development of detection technology optimally suited for image acquisition in both modalities remains a major challenge. Here we demonstrate a multi-segment detector array whose novel design is meant to optimally support both ultrasound and optoacoustic image acquisition.

Speaker: Elena Mercep (iThera Medical, GmbH)

10:30 AM Coffee break

Session 6: System design

Conveners: Ivana Ivanovic Balic (SonoView Acoustic Sensing Technologies), Torsten Hopp (Karlsruhe Institute of Technology)

20 Source selection for ultrasound waveform tomography: real data case from USCT Data Challenge 2017

Waveform inversion for Ultrasound Computed Tomography (USCT) is an emerging high-resolution imaging technique for breast cancer screening. Despite its potential, the involved computational burden is challenging. This mainly depends on the total number of wave propagation simulations solved during the inversion procedure. Thus, the computational cost can be reduced by only selecting the emitting transducers that best resolve the information about the breast tissue and avoid redundant experiments.

Speaker: Naiara Korta Martiartu (ETH Zurich)

21 Upper Bound of Accuracy for Self-Calibration of a 3D Ultrasound Tomography System without Ground Truth

A Newton’s method based self-calibration was presented in previous work for a 3D Ultrasound Tomography System (USCT). The method sequentially calibrates a complex USCT system with 2041 transducers based on time-of-flight (TOF) measurements. A direct evaluation of the calibration result was not possible due to unknown ground truth. In this work we present a method to estimate the upper boundary for the calibration accuracy.

Speaker: Wei Yap Tan (Karlsruhe Institute of Technology)

22 System Design of a Flexible 512-Channel Platform for Ultrasound Computer Tomography

Breast cancer has become one of the primary diseases threatening women’s health. Ultrasound computer tomography (USCT) has attracted many researchers’ attention for its potential in early diagnosis of breast cancer. In USCT system, transducer with large number of elements are often be adopted[4-5]. Platform with large number of channels is needed, In this paper, we describes the system design of a 512-channel USCT platform.

Speaker: Junjie Song (Huazhong University of Science and Technology, China)

12:00 PM Lunch

Session 7: Ultrasound transducer

Conveners: Hartmut Gemmeke (KIT), Philippe Lasaygues (Aix Marseille Univ, CNRS, Centrale Marseille, Laboratory of Mechanics and Acoustics, Marseille, France)

23 Latest developments in ultrasound transducer technology and future directions

Speaker: Nicolaas de Jong (Technical University of Delft)

24 Technologies for Piezoceramic Composites used for Ultrasonic Transducers

Piezoelectric ultrasonic transducers are widely used in many fields, such as medical ultrasonic, nondestructive testing and structural health monitoring. The limits of piezoelectric bulk ceramics in high electromechanical coupling, broad bandwidth, good acoustic impedance matching and high resolution of acoustically separated elements lead to attention on piezoelectric 1-3 composites with improved electrical, electromechanical and mechanical properties.

Speaker: Holger Neubert (Fraunhofer IKTS Dresden)

25 Dice-and-fill single element octagon transducers for next generation 3D USCT

At KIT a 3D USCT system is under development. The system is optimized for SAFT imaging and has a multistatic setup of 2041 ultrasound transducers with approx. 1MHz 3dB bandwidth and 36° 3dB opening angle for 2.5MHz. The USCT groups transducers in a semi-elliposoidal aperture surrounding a ROI of 10x10x10cm$^3$. To increase the ROI for a next USCT generation, the opening angle of a future transducer should be increased to approx. 60° while other characteristics should be preserved or improved.

Speaker: Michael Zapf (Karlsruhe Institute of Technology)

2:40 PM Coffee break

Poster session

26 Comparison of registration strategies for USCT-MRI image fusion: preliminary results

Comparing USCT to well-known MRI is an essential step to evaluate the clinical value of USCT. Yet the different conditions of the breast either embedded in water (USCT) or in air (MRI) prevent direct comparison of tissue structures due to non-linear deformations of the breast in 3D. Previously we presented an image registration simulating the MRI subjected to buoyancy. In this work we compare it with a new strategy by applying gravity to USCT to match the breast shape in the MRI.

27 Comparison of two ray tracing methods for sound speed imaging

Sound speed imaging in USCT is a kind of functional imaging mode which can provide valuable information for differentiating normal tissue and tumor. Sound speed imaging in USCT is a kind of functional imaging mode which can provide valuable information for differentiating normal tissue and tumor. One of the key problems of sound speed reconstruction is ray tracing. This paper compares the sensitivity and the accuracy of the finite difference method and the linear interpolation method.

28 Contrast resolution enhancement of Ultrasonic Computed Tomography (USCT) using a wavelet-based method – Preliminary results in bone imaging

In the case of bone imaging, current medical echography does not allow imaging under the cortical zone; no information is available on cortical depth, underlying medullary tissue or marrow. Many authors have discussed ultrasonic imaging of bones. Their main objective was to evaluate the thickness of the diaphysis in axial and transverse mode, and to estimate the velocity of the sound of a wave. Our group has been focusing on the cross-sectional radial imaging process, using USCT

29 Fast reflectivity imaging in 3D using SAFT

The computational burden for 3D Synthetic Aperture Focusing Technique (SAFT) is large as for each voxel the delay for each acquired A-scan has to be calculated, e.g. O(N$^5$) for N$^3$ voxels and N$^2$ A-scans. If the 3D distribution of speed of sound is applied to correct the delays for objects with varying speed of sound the computation time increases further. This overview paper presents the implementations for 3D SAFT developed by the KIT group and discusses their performance.

30 Improved temperature measurement and modeling for 3D USCT II

Ultrasound computer tomography for breast cancer relies on accurate SOS information, mostly disturbed by the temperature dependency of H$^2$O as medium. The current KIT USCT system (3D USCT II) incorporates many temperature sensors: two accurate calibrated and many low accuracy temp. sensors for spatially densely sampled 3d information. The accurate knowledge of the temperature is important for accurate and sharp imaging as in the USCT setup approx. 50% of the US travel distance is through water. Over an USCT measurement duration the sensors are digitized in second’s interval. While many experiments are represented accurately some experiments needs an enhanced and robustified approach.

31 Measurement of the speed of sound, attenuation and mass density of fresh breast tissue

An improved distinction between benign and malignant lesions is necessary to enhance the diagnostic value of breast ultrasonography. This requires a better quantitative characterization of tissues and hence accurate knowledge about the acoustic medium parameters. Current data on these parameters is inconsistent due to variations in applied measurement protocols. We tested a standardized protocol for measuring the speed of sound, attenuation and mass density of freshly excised breast tissues.

32 Minimum-variance beamforming for ultrasound computer tomography imaging

The breast cancer has become the most common type of cancer among women throughout the world.The traditional delay-and-sum (DAS) beamforming has been widely used in USCT imaging. In this paper, minimum-variance (MV) beamforming method is applied to improve the image quality for USCT. The USCT image is expected to have less noises and artifacts, and higher resolution and contrast.

33 Object Classification and Localization with an Airborne Ultrasound Imaging System

An airborne ultrasound imaging system (ABUS) was developed at KIT for reflection tomography. The prototype system consists of sixteen ultrasonic transducers surrounding a region of interest (ROI) of defined shape with a diameter of 50 cm. The transducers have a center frequency of 200 kHz and a bandwidth of 20 kHz. The prototype aims to demonstrate possible industrial applications for object classification and localization with airborne ultrasound.

34 Piezofibre composite transducers for next generation 3D USCT

At KIT a 3D USCT is under development. The system is optimized for SAFT imaging and has a multistatic setup of 2041 transducers with approx. 1MHz 3dB bandwidth and 36° 3dB opening angle for 2.5MHz. The USCT groups transducers in a semi-elliposoidal aperture surrounding a ROI of 10x10x10cm$^3$. To increase the ROI for a next USCT generation, the opening angle of a future transducer design should be increased to approx. 60° while other characteristics should be preserved or even improved.

35 The USCT reference data base

Ultrasound Computer Tomography (USCT) is an emerging technology mostly aimed at breast cancer imaging. Following the idea of open science a USCT reference data base (http://ipeusctdb1.ipe.kit.edu/~usct/challenge/) is established with open and easy to use data and code interfaces. The aim is to promote and facilitate the exchange of available reconstruction algorithms and raw data sets from different USCT devices throughout the growing USCT community.

36 Visualisation of Ultrasound Computer Tomography Breast Data Set

Visualising volume images is a key factor in diagnosing and detecting early breast cancer. However, the standard visualisation approaches still revolve around 2D images slides. Although 3D visualisation is important in diagnosing and treating breast cancer, yet the potential is not entirely realized with slices based visualisation. In this paper, we present an interactive 3D web-based visualisation tool for the breast data set which overcomes this limitation and offers new possibilities.

Discussion session: Systems

Convener: Nicole Ruiter (Karlsruhe Institute of Technology)

Walking tour to conference dinner

6:30 PM Conference Dinner

Fri, November 3

Session 8: Clinical applications I

Conveners: James Greenleaf (Mayo Clinic), Jeffrey Bamber (Institute of Cancer Research)

37 Challenges in Breast Ultrasound

Speaker: Helmut Madjar (DKD HELIOS Klinik Wiesbaden)

38 Ultrasound Tomography for Breast Cancer Screening

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. We have been developing a new technique for breast imaging that is based on ultrasound tomography which quantifies tissue characteristics for development of more objective criteria.

Speaker: Neb Duric (Delphinus Medical Technologies and Karmanos Cancer Institute)

39 3D Ultrasound Tomography for Breast Cancer Diagnosis at KIT: an Overview

3D USCT emitting and receiving spherical wave fronts overcomes the limitations of 2D systems by offering a nearly isotropic 3D point spread function, a large depth of field, less loss of out-of-plane reflections and fast 3D data acquisition. Yet 3D devices for clinical practice require a more complex hard- and software due to the huge data rate, time-consuming image reconstruction, and large number of small transducers. The here reviewed KIT 3D USCT is a 3D prototype for clinical studies.

Speaker: Nicole Ruiter (Karlsruhe Institute of Technology)

40 Breast Tissue Characterization with Sound Speed and Tissue Stiffness Imaging

In this study, we are going to assess the ability of SoftVue’s waveform sound speed and through-transmission based stiffness image to render a variety of breast tissue and breast masses.

Speaker: Cuiping Li (Delphinus Medical Technologies, Inc., USA)

10:30 AM Coffee break

Session 9: Image processing

Conveners: Peter Littrup (Delphinus Medical Technologies), Richard Prager (University of Cambridge)

41 Postprocessing workflow of 3D-USCT: bridging the gap to the clinic

As first USCT systems are approaching clinical application, it is an essential task to prepare the reconstructed images for intuitive diagnosis and conform to clinical standards. We describe our post-processing workflow consisting of automated breast segmentation, image fusion, DICOM export, our dedicated USCT DICOM viewer and the methods to transfer images to the clinic.

Speaker: Torsten Hopp (Karlsruhe Institute of Technology)

42 Tissue Characterization With Ultrasound Tomography Machine Learning

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.

Speaker: Gursharan Sandhu (Delphinus Medical Technologies, Inc.)

43 Challenges and applications of registering 3D Ultrasound Computer Tomography with conventional breast imaging techniques

For evaluation of 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 developed a patient-specific image registration method, which simulates different breast positions through biomechanical modelling. We give a review of its recent applications and raise open challenges.

Speaker: Patricia Cotic Smole (Karlsruhe Institute of Technology)

12:00 PM Lunch

Session 10: Clinical application II

Conveners: Michael Zapf (Karlsruhe Institute of Technology), Nicole Ruiter (Karlsruhe Institute of Technology)

44 Optoacoustic tomography and the pathway to commercialization

Speaker: Christian Wiest (iThera Medical GmbH)

45 Feasibility study on USCT for brain imaging to estimate artifacts and image distortion caused by bone propagation

We are trying to image inside the skull with USCT with the goal of examination of acute stroke during the emergency conveyance. For transcranial USCT, attenuation and phase distortion after the pass through the skull would degrade its imaging performance. The purpose of this study is to establish the reconstruction method for the transcranial USCT through the process of a frequency optimization and an artifact cancellation to recover the energy attenuation and the focus quality, respectively.

Speaker: Yuki Hayashi (The Univ. of Tokyo)

46 Multi-perspective ultrasound imaging of abdominal aortas

Ultrasound functional imaging of large vessels is challenged by regional low contrast of the wall. Moreover, displacement tracking in the lateral direction suffers from poor lateral resolution. These challenges can be tackled in superficial vasculature with beam steering and compounding. However, in deep vessels like the abdominal aorta, beam steering is not possible because the aorta will not be in the field-of-view. In this study, multi-perspective imaging of the abdominal aorta is proposed.

Speaker: Niels Petterson (Eindhoven University of Technology)

2:40 PM Coffee break

Session 11: Clinical application III

Conveners: Gursharan Sandhu (Delphinus Medical Technologies, Inc.), Jorge Camacho (Spanish National Research Council (CSIC))

47 A simple method for acoustic properties determination of cancerous tissue and its implementation into the clinical workflow

During USCT examination, tissue morphology, as well as acoustic properties such as speed of sound are extracted. Data based upon malignant tissue is limited, so is its reliability as a prognostic factor indicating a malignant or benign nature. It’s hard to obtain tumour tissue because of their pathological use for staging. Therefore, this work will present a simple method of acoustic properties measurement and its implementation in a clinical workflow.

Speaker: Frank Wolfram (Lung Cancer Centre SRH Waldklinikum Gera)

48 Clinical ultrasound breast tomography using Softvue®: a preliminary in vitro and in vivo assessment

Softvue® (Delphinus Inc.) ultrasound tomography (UST) provides images of sound speed (SS), attenuation coefficient (AC), a combination of SS and AC called stiffness, and relative reflectivity. We have compared volume-averaged SS (VASS) with MRI volume-averaged water versus fat, to quantify breast density (O’Flynn et al., 2017). Here, clinical reproducibility is assessed, stiffness compared with MRI, and VASS compared with mammographic density. Performance is estimated using phantoms.

Speaker: Jeffrey Bamber (Institute of Cancer Research)

49 The New Generation of the KIT 3D USCT

The first clinical studies with our current prototype, 3D USCT II, enabled us to identify the necessary improvements for transition of the method to clinical practice. The main goals are to improve the contrast of reflection tomography, the resolution of transmission tomography, and to optimize imaged volume of the breast. Furthermore, the fabrication of the transducer arrays needs to be cost-effective and the data acquisition and readout time should be reduced.

Speaker: Hartmut Gemmeke (KIT-IPE)

Discussion session: Clinical applications

Conveners: Jeffrey Bamber (Institute of Cancer Research), Neb Duric (Delphinus Medical Technologies and Karmanos Cancer Institute)

Summary talk + closing

50 Potential impact of ultrasound tomography in the clinic

Speaker: Peter Littrup (Karmanos Cancer Institute)

51 Closing ceremony

Speakers: Jeffrey Bamber (Institute of Cancer Research), Koen van Dongen (Delft University of Technology), Neb Duric (Delphinus Medical Technologies and Karmanos Cancer Institute), Nicole Ruiter (Karlsruhe Institute of Technology), Torsten Hopp (Karlsruhe Institute of Technology)