Project 1 Tissue characterisation by endoscopic sonoelastography and Haptic Imaging

MedViz – a Research & Development cluster in Bergen. MedViz is a cluster of groups performing interdisciplinary research in advanced image analysis and visualisation bridging the gap between “bench and bedside”. MedViz also includes imaging physics and fundamental biomedical translational research. Our mission is to promote the development of biomedical and clinical imaging technologies with the ultimate goal of improving diagnosis and treatment of patients. MedViz wants to develop new clinical methods for potential commercialisation by means of denovo software for analysis, decision support and visual communication.

Project leader: PhD Candidate Roald Flesland Havre, Institute of Medicine

Endoscopic ultrasonography (EUS) combines endoscopy with high frequency ultrasonography for intraluminal use. Thus, detailed images of small lesions in and outside the gastrointestinal (GI) tract can be obtained. So far, neither ultrasound nor other imaging techniques can identify tissue in regard to separating benign (inflammation) from malignant tissue. Sonoelastography is an ultrasound based elasticity imaging technique which can be applied for analyzing tissue stiffness. Recently this technique has also become available as an ultrasound modality which is integrated in EUS-systems (S-EUS). The method has been regarded as a possible tool for identifying cancer tissue. In our experience, different types of benign and malignant tumours may show different pattern depending on the tissue characteristics. Therefore, it is important to perform in vitro and in vivo studies which can enhance the understanding of S-EUS images. If images can precisely identify malignant tissue, this would be of major clinical importance and reduce the need for obtaining biopsies, which always carries a certain risk. Most tumours have combined inflammatory and malignant areas and if a biopsy is necessary, S-EUS can be helpful in identifying the best area to obtain tissue material. In our experience, S-EUS seems to be a valuable new technology for improving diagnosis of malignant tissue.

Haptic imaging is a visualisation principle where data is displayed in a tactile manner [1]. Besides the surface structure sensing, haptic imaging has been also investigated to support interaction on volume data directly [2]. One important advantage of this method is that the elastic properties of tissue can be examined by palpating and feeling the acquired image data. Thus, tissues with different stiffness and elasticity may be examined during or after finishing the examination of the patient.

Objectives

The overall clinical goal is to improve endosonographic imaging of GI tumours with the aim to separate inflammatory tissue and fibrosis from malignant tissue.
To reduce and possibly avoid obtaining biopsies for histopathological analysis.
To develop a haptic imaging software/hardware platform that can support EUS elastographic diagnosis using alternative ways of visualisation of the elasticity data obtained from biological tissue and phantoms.
Compare EUS-sonoelastography and haptic imaging with histological findings (EUS- guided biopsies and resected tissue).

Methods

Our studies will initially be performed using linear curved-array ultrasound endoscopes. These instruments consist of an oblique forward-looking endoscope with a linear, convex curved-array electronic transducer, which produces a real-time sector scan of a plane parallel to the long axis of the endoscope (Fig.1). The accessory channel is running into the imaging plane, allowing inserted accessories, such as a biopsy needle, to be recorded in the real-time. As with other electronic US endoscopes the probe is linked with a conventional ultrasound machine enabling utilization of its advanced facilities, such as an integrated software program for sonoelastography and different Doppler functions including real-time duplex imaging. Haptic imaging using echoendoscopes is at the present time not possible. We are currently working on a concept for a software-hardware approach to haptic exploration of elastography data. Within this project we want to research new algorithmic solutions for a computer assisted interpretation of elastography data for the endoscopic examinations. The new data interpretation approach will be worked out in cooperation with Institute of Informatics at UiB (Prof. Helwig Hauser and 1. aman. Dr. Ivan Viola) and with Norsk Hydro (Jens Grimsgaard and Gunnar Halvorsen). The vendor-provided software that combines elastography measurements with B-mode images will be enriched by adding the haptic interface. This enables palpation-like interaction on the acquired data during or after the intervention. This novel interaction technique will be first to allow to remotely palpate the internal structures in the patients body that are not accessible to clinicians hands otherwise.

R&D Plan

The first part of the project is currently running and a paper which describes the influence of different sonoelastographic scanning parameters has recently been submitted for publication (R.F. Havre et al.). Furthermore, the technique has been applied for examination of nearly 40 patients with benign and malignant tumors and EUS fine needle and thru-cut biopsies have been obtained from some patients. This study will be extended and published in an international journal in 2008. The second part of the project, related to haptic imaging, has already started by a Master thesis project that is supervised jointly by the clinical and informatics partners. The first task in the new project is to design a software framework that combines different co-registered modalities in one visualization approach. Then haptic support will be added to the framework and the EUS elastography stiffness values will be mapped to the haptic feedback. A test environment will be provided by Norsk Hydro for early proof-of-concept tests using the Phantom haptic device. The implementation will be followed by a first evaluation from the clinical and methodological sides in order to establish design guidelines for EUS elastography haptic imaging software and hardware. The challenge making an EUS device which can be used also for haptic imaging will take time, probably running for 35 years, including testing the equipment in vitro and in patients.

Figure 1: Screenshot from a first version of a prototype showing 3D and 2D elastography and the corresponding B-mode image.

Name	E-mail	Institution	Acad. pos.	Role in project
Roald Havre	This e-mail address is being protected from spambots. You need JavaScript enabled to view it	Helse Bergen	PhD Stud.	clinical evaluation
Odd Helge Gilja	This e-mail address is being protected from spambots. You need JavaScript enabled to view it	Helse Bergen	Professor	clinical evaluation
Svein Ødegaard	This e-mail address is being protected from spambots. You need JavaScript enabled to view it	Helse Bergen	Professor	clinical evaluation
Ivan Viola	This e-mail address is being protected from spambots. You need JavaScript enabled to view it	II.UiB	PostDoc	methodology in visu-alization and haptics
Helwig Hauser	This e-mail address is being protected from spambots. You need JavaScript enabled to view it	II.UiB	Professor	methodology in visualization
Gunnar Halvordsen	This e-mail address is being protected from spambots. You need JavaScript enabled to view it	Norsk Hydro		methodology in haptic imaging
Jens Grimsgaard	This e-mail address is being protected from spambots. You need JavaScript enabled to view it	Norsk Hydro		methodology in haptic imaging

References

[1] Atlas of Endoscopic Ultrasonography. The upper gastrointestinal tract. Editors of book: Ødegaard S, Nesje LB, Gilja OH. 2007:1-208. Fagbokforlaget, Bergen, Norway. ISBN 978-82-450-0335-2.
[2] Gilja OH, Hausken T, Hatlebakk JG, Ødegaard S, Berstad A, Viola I, Giertsen C, Gregersen H. Advanced Imaging and Visualization in Gastrointestinal Disorders. World J Gastroenterol 2007;13(9):1408-21.
[3] R.L. Williams II, J.N. Howell, and R.R. Conatser Jr., 2007, Digital Human Modeling for Palpatory Medical Training with Haptic Feedback, in Handbook of Digital Human Modeling for Applied Ergonomics and Human Factors Engineering, Society of Automotive Engineers.
[4] K.L. Palmerius, Direct Volume Haptics for Visualization, PhD thesis at Linköping University in Sweden, 2007.