Sei sulla pagina 1di 7

Medical impedance tomography and process impedance tomography: a brief review

This article has been downloaded from IOPscience. Please scroll down to see the full text article.
2001 Meas. Sci. Technol. 12 991
(http://iopscience.iop.org/0957-0233/12/8/301)
Download details:
IP Address: 128.148.252.35
The article was downloaded on 09/11/2012 at 13:34
Please note that terms and conditions apply.
View the table of contents for this issue, or go to the journal homepage for more
Home Search Collections Journals About Contact us My IOPscience
INSTITUTE OF PHYSICS PUBLISHING MEASUREMENT SCIENCE AND TECHNOLOGY
Meas. Sci. Technol. 12 (2001) 991996 www.iop.org/Journals/mt PII: S0957-0233(01)22416-3
Medical impedance tomography and
process impedance tomography: a brief
review
Brian H Brown
Medical Physics and Clinical Engineering, University of Shefeld,
Royal Hallamshire Hospital, Shefeld S10 2JF, UK
E-mail: b.h.brown@shefeld.ac.uk
Received 2 March 2001, in nal form 31 May 2001, accepted for
publication 6 June 2001
Abstract
The resurgence of research into medical electrical impedance tomography
about 20 years ago was soon accompanied by a parallel development in
process impedance tomography. The interaction between these two research
communities was benecial to both groups. In recent years this interaction
has been very much reduced. This paper briey reviews the history and then
the current developments in medical impedance tomography.
Keywords: impedance, imaging, medical tomography, process tomography
1. Introduction
Several titles have been adopted to describe impedance
imaging systems. These include electrical tomography,
impedance imaging, resistance imaging, electrical resistance
tomography, impedance tomography and applied potential
tomography. The one in widest currency in terms of medical
imaging is electrical impedance tomography (EIT). The title
EIT has the advantage that you can consider it to include
resistive, capacitive and inductive tomographies.
All the proposed impedance imaging systems have a basic
methodology in common. They are all methods of imaging the
distribution of conductivity or permittivity within a volume.
This volume might be a part of the human body or it might
be the contents of a pipeline or vessel in a process control
situation. There are two basic stages to the production of an
impedance image:
(i) the collection of a set of N independent transfer
impedances and
(ii) the solution of an inverse problem in order to produce an
image from the set of transfer impedances.
This process can be illustrated in a medical context. We can
apply a set of electrodes around the chest as shown in gure 1
and then make the set of transfer impedance measurements
by applying currents and measuring voltages as illustrated in
gure 2. It is easy to show that, if 16 electrodes are used,
then only 104 (N(N 3)/2) independent transfer impedance
measurements can be made. These measurements can then be
used to produce an image such as that illustrated in gure 3.
Figure 1. Electrodes applied for thoracic imaging.
Body tissues have resistivities similar to those of
semiconductors (Duck 1990). Blood has a resistivity of
approximately 1.6 m. Lung has a higher resistivity of
about 10 m and this resistivity changes as air is inspired and
expired. It is relatively easy to forman image of the changes in
resistivity distribution during respiration. Figure 3 shows such
an image. However, this image illustrates clearly the major
disadvantage of EIT. The spatial resolution is poor because
there are only 104 independent measurements from which the
image can be reconstructed. This image resolves the two lungs
and the cardiac region but the spatial resolution is only of the
order of 30 mmfor a thorax which is 300 mmfromside to side.
However, there are balancing advantages which explain why
EIT is still under investigation and is used in several situations.
These advantages include
(i) its low cost,
0957-0233/01/080991+06$30.00 2001 IOP Publishing Ltd Printed in the UK 991
B H Brown
Figure 2. The sequence of transfer impedance measurements used to collect imaging data from 16 electrodes.
Figure 3. An EIT image showing the change in resistivity of the
lung during breathing. The maximum change in resistivity is by
approximately 50%.
(ii) that there is no hazard to the patient,
(iii) its simplicity of application,
(iv) the high speed of data collection (typically 25 frames s
1
)
and
(v) that tissue/material characterization is possible
This illustration of impedance imaging has used resistive
imaging and a medical image as an example. However,
capacitive imaging and inductive imaging are subject to very
similar constraints. In all cases current is made to pass through
the volume of interest and the resulting eld patterns are used
to produce an image of conductivity or permittivity. A set of
independent measurements is made and the number of these
constrains the resolution of the image which can be produced.
The objectives of the medical and process tomography
areas are very different and some of the advantages and
disadvantages are relevant to one area only. However, workers
in both areas were quick to see advantages in the development
of impedance tomography and have put considerable effort into
research on the subject.
2. A brief history of impedance imaging
EIThas quite a long history and should no longer be considered
a new technique. The rst publication of an impedance image
was that of Henderson and Webster (1978). They used a
two-dimensional matrix of 100 electrodes on one side of the
thorax and a single large electrode on the other to produce a
transmission image of the tissues. However, the rst published
tomographic images were those of Brown (1983) and Barber
et al (1983). They used 16 electrodes, injection of current
between adjacent electrodes and a back-projection method of
image reconstruction along isopotentials.
Impedance imaging was introduced into process tomogra-
phy during the 1980s and Professor Maurice Beck of Manch-
ester was instrumental in getting the UK Medical and Process
Tomography groups together on several occasions. This inter-
action gave rise to EIT systems being developed along similar
lines in both research areas and led to developments such as
capacitive and inductive tomography (Wang et al 1992, Yang
and York 1999, Grifths et al 1999, Korjenevsky et al 2000,
McCann et al 1999). The two groups drifted apart again during
the 1990s but perhaps this publication is evidence that they are
now coming together again in the new millennium.
During the last twenty years there have been many
research programmes and a number of general publications
in both medical and process tomography areas. For example,
in the medical EIT area there have been the following.
(i) Two European concerted action programmes on EIT
(called Applied Potential Tomography (APT) in the
early years). These gave rise to special issues of the
journal Physiological Measurement between 1987 and
1996 following meetings in Shefeld, Lyon, Copenhagen,
York, Barcelona, Ankara and Heidelberg.
(ii) A book introducing the technology published in 1990,
Electrical Impedance Tomography, ed J G Webster
(Bristol: Hilger)
(iii) Abook considering the possible clinical applications pub-
lished in 1993, Clinical and Physiological Applications of
EIT, ed D S Holder (London: UCL Press).
(iv) Acomprehensive reviewarticle which appeared following
the completion of the European Union concerted action
programmes in 1997, Imaging with electricity: report of
the European concerted action on impedance tomography,
J. Med. Engng Technol. 21 20132.
(v) A relatively recent review which followed the 10th
International Conference on Electrical Bio-impedance
held in Barcelona, Spain in 1998. Many of the
papers are publishedas Electrical Bioimpedance Methods,
992
Medical impedance tomography
Applications toMedicine andBiotechnology, Annals of the
New York Academy of Sciences, volume 873 (1999).
(vi) A very recent review of some medical applications by
Frerichs, Physiol. Meas. 21 R121 (2000).
(vii) A book called Bioimpedance and Bioelectricity, Basics
by S Grimnes and G Martinsen published by Academic
Press in 2000.
Several reviews and books on the process control applications
of impedance tomography have also appeared. For example,
see
(i) Process TomographyState of the Art (Beck et al 1998),
(ii) special issue on process tomography, Meas. Sci. Technol.
(1996),
(iii) a book called Process Tomography: Principles
Techniques and Applications (Williams and Beck 1995)
and
(iv) a review, Status of electrical tomography in industrial
applications (York 2001).
There is also a wide range of Web based sources
on medical and process EIT (see www.eit.org.uk and
www.tomography.umist.ac.uk). The volume of medical
research into EIT fell during the late 1980s because no
major medical applications had emerged and it appeared that
the limits of performance of EIT systems had almost been
reached. However, at that stage the use of multi-frequency
EIT began to be developed and this opened up ways around
some of the problems of medical EIT. One major problem
in medical EIT is in taking the body shape into account.
Body segments have complex shapes and these are the major
determinants of any set of transfer impedance measurements.
However, by using measurements made at one frequency as a
reference and then measurements made at another frequency
as a data set, it is possible to reduce the effects of body
shape to a second-order effect. By making multi-frequency
measurements, it was shown that tissue characterization using
impedance spectroscopy was possible. Biological tissues
have a complex impedance because of their cellular structure.
At low frequencies electrical current ows around cells,
whereas at high frequencies the current can penetrate the cell
membranes and hence ow through the intracellular spaces.
The signicance of this is further expanded in section 4.1.
3. Medical EIT
Opinions regarding what the main research problems in
medical EIT are vary, but the following three appear on most
lists:
(i) taking into account the curved paths of electrical eld
when reconstructing images,
(ii) taking into account the body shape and
(iii) making sufciently accurate measurements of the transfer
impedance between electrode or coil combinations.
The rst and third of these are common both to process
tomography and to medical tomograpy. The second is specic
to medical applications. The third problem arises because of
the nature of EIT. The result of making a transfer impedance
measurement from a volume conductor is a three-dimensional
integral. Every part of the volume conductor contributes to
every measurement and hence the contribution from any one
voxel will be small. This means that a large dynamic range is
required from the measurements. This can be very difcult to
achieve, particularly when the effect of electrode impedance
is taken into account. Many researchers have wished that
the electrodes could be made to disappear. By using
current sources with innite output impedance and recording
ampliers with innite input impedance, the electrodes can
be made to disappear, in that they should not affect the
measurements. However, current sources and ampliers are
never perfect, so electrode problems remain. One way to
solve the problem is to use inductive tomography, for which
no electrodes are required. Current can be induced into the
volume conductor magnetically and the resultant current in
the tissues produces opposing magnetic elds which can also
be detected. The instrumentation problems involved are very
considerable but this area is a very active one for current
research (Tozer et al 1999, Grifths et al 1999, Korjenevsky
et al 2000). There are other improvements to the hardware
which can be made to reduce the effects of electrodes and
these will be considered in section 4.
The following list gives some of the medical applications
of EIT which have been considered:
(i) gastric emptying measurement, particularly for neonates
(Vaisman et al 1999) ,
(ii) gastric pH measurement (Watson et al 1996),
(iii) lung ventilation monitoring (Khambete et al 2000) ,
(iv) optimization of ventilation during anaesthesia (Frerichs
et al 1999) ,
(v) assessment of lung water in neonates and in adults with
heart failure (Noble et al 1999) ,
(vi) detection of pulmonary embolus detection using both
ventilation and perfusion measurement (Leathard et al
1994) ,
(vii) head imaging with a viewto the detection of epileptic foci
(Tidswell et al 2001) ,
(viii) breast imaging to detect cancer (Kerner et al 2001, Ultchin
et al 2001) ,
(ix) measurement of tissue temperature during hyperthermia
(Conway et al 1992),
(x) measurement of cardiac output (Patterson and Zhang
2001, McArdle et al 1993) and
(xi) tissue characterization (Brown et al 2000) .
The subjects marked with a have been investigated as
research applications. Nearly all these applications involve
the measurement of changes in impedance. These changes
may be either temporal or frequency changes. One exception
is the measurement of lung water, but here the changes are very
large and so can be separated from the effects of body shape.
4. Future research in process and medical EIT
In addition to research into inductive tomography, there appear
to be three major areas of change in terms of the technology of
medical EIT. The rst concerns the physical construction of the
hardware, the second the use of multi-frequency data capture
and the third three-dimensional EIT. The rst and third have
seen parallel developments in process and medical EIT. The
second may be less relevant to process EITbecause most of the
993
B H Brown

10 20 30 40 50 60 70
5
10
15
20
25
30
D
e
p
t
h


>
>
>
>
current injection current injection
^ ^
Low frequency
10 20 30 40 50 60 70
5
10
15
20
25
30
D
e
p
t
h


>
>
>
>
High frequency
current injection current injection
^ ^
(a)
10 20 30 40 50 60 70
5
10
15
20
25
30
D
e
p
t
h


>
>
>
>
Low frequency
current injection current injection
^ ^
10 20 30 40 50 60 70
5
10
15
20
25
30
D
e
p
t
h


>
>
>
>
High frequency
current injection current injection
^ ^
(b)
Figure 4. The isopotentials when current is passed between two electrodes placed at nodes 15 and 55. In (a) the cells are distributed
randomly. It can be seen that the low and high frequency isopotential patterns are similar, although current does penetrate the cells at the
high frequencies. In (b) the cells have been placed in a layer, in a manner found in some epithelial tissues. In this case both high and low
frequency ispotential distributions are very different from those seen in (a).
volume conductors involved do not have complex impedances
in the frequency range where measurements are relatively easy
to make. However, if biological materials are to be handled in
processes such as those in biotechnology, then multi-frequency
EIT may well be relevant to process EIT.
4.1. Hardware
Nearly all medical EIT systems have used standard Ag/AgCl
electrodes attached to the body and leads to connect to the
data collection electronics. People realized at an early stage
that the leads were very important and that they limited the
achievable performance of the data collection system. Boot-
strapping has been used widely and even triaxial cables have
been used, but the limits of performance were probably reached
a few years ago. Shorter leads, or indeed getting rid of
the leads completely, was the only way to go but people
have been quite slow to take this direction. Process EIT
has probably led the way here. In part this was because the
investigation of capacitance tomography made placement of
the electronics very close to the electrodes absolutely essential.
My opinion is that the integration of the current injection
and potential measurement electronics with the electrode
assemblies is vital if further signicant improvements in
accuracy of measurement are to be obtained.
4.2. Multi-frequency-impedance spectroscopy
Biological tissues contain cells and this results in the tissue
impedance being complex. Typically the magnitude of the
tissue impedance will fall by at least 50% between 10 kHz
and 1 MHz. Put simply this is because current starts to
penetrate cell membranes as the frequency increases. This
is illustrated in gure 4(a), in which a mesh model has been
used to plot isopotentials at low and high frequencies. It
is possible to use the idea of differential EIT imaging to
observe changes of tissue impedance with frequency. By
using a measurement made at one frequency as a reference
and then using measurements made at other frequencies as
data, images showing how impedance changes with frequency
can be obtained. The impedance spectrum for each pixel
in the image can then be tted to a Cole model and hence
images showing the spatial distribution of the Cole parameters
determined. This has beenshowntobe animportant innovation
and allows quantitative measurements of tissue properties to
be imaged.
In medical EIT it is now becoming appreciated that an
impedance spectrum can perhaps be related quantitatively to
the structure of tissue. This point is illustrated in gure 4(b), in
which the cells which were randomly distributed in gure 4(a)
have now been placed in a layer. The impedance spectrum
994
Medical impedance tomography
that results is very much different. An impedance spectrum is
determined by the arrangement of cells within the tissue. This
is a very important observation and is the basis for recent work
showing that pre-malignant changes in cervical tissue can be
detected using impedance spectroscopy (Brown et al 2000).
Similar changes are being investigated in oesophageal and
bladder tissues, for which large nite element tissue models
are being developed.
4.3. Three-dimensional imaging
Most objects of interest are three-dimensional. However, the
images produced by most EIT research groups have treated the
object as a series of two-dimensional slices, by collecting data
from electrodes placed around the borders of a plane through
the object. Unfortunately, unlike other three-dimensional
imaging methods such as x-ray computed tomography, a slice
cannot be simply dened in this way in EIT. Electrical current
injected into a three-dimensional object cannot be conned
to a plane but ows above and below the plane dened by
the electrodes. This means that conductivity distributions
in these regions contribute to the measured signals. In
order to eliminate this problem, data collection and image
reconstruction must be considered as a fully three-dimensional
problem. This can be achieved by placing electrodes over the
whole surface of the object. Typically this is done by placing
electrodes in a series of planes around the object. Any pair of
electrodes, both within and between planes, can be used either
for applying current or for making potential measurements.
Reconstruction, although involving more measurements and
larger sensitivity matrices, is identical in principle to the two-
dimensional case (Metherall et al 1996).
5. Conclusions
EIT is a fast and inexpensive method of imaging material
properties with many possible applications in both process and
medical elds. However, the two areas of application are very
different and have many different constraints. For example, the
geometry is usually well controlled in process instrumentation,
in which electrodes may be placed around a regular and rigid
structure. However, electrodes placed around body segments
will have irregular geometry and the shape may change with
time. It is for this reason that capacitive tomography using
quite large electrodes is a reasonable way forwards for process
measurements but not for medical measurements. Another
difference is that, in many process control situations, the
electrical properties of the uid are known and the purpose of
the tomographic measurement is to assess the relative volumes
of the ow compartments. In the medical situation the body
tissues have unknown electrical properties, although the range
of likely values of conductivity and permittivity for particular
tissues is known. A further difference between the medical
and process situations is that the dynamic range of values
for conductivity and permittivity is often much greater in the
process area.
Multi-frequency measurements were made in the medical
EITeldas a means of reducingthe dependence onbodyshape.
However, the real value of spectral measurements has been
shown to be in tissue characterization. Because body tissues
are cellular, they have a characteristic impedance spectrum
which can be used to assess the function of particular organs. It
maybe that the use of process instrumentationinbiotechnology
will nd similar value in the use of tomographic spectroscopy.
Future EIT data collection systems are likely to be much
smaller and have wider bandwidths than present systems. They
may be completely wireless and use some of the developments
in wireless local area network (LAN) systems to return data
to a microprocessor. Such systems could well be utilized to
collect data without any ground reference and hence obtain
better immunity to noise. Impedance tomography is still
a fascinating subject for research and one in which greater
interaction between the medical and process areas can only be
of benet.
References
Barber D C and Brown B H (ed) 1994 Electrical impedance
tomography, special issue of Physiol. Meas. 15 (2A)
1996 Electrical impedance tomography, special issue of
Physiol. Meas. A 16 266
Barber D C, Brown B H and Freeston I L 1983 Imaging spatial
distributions of resistivity using applied potential tomography
Electron. Lett. 19 935
Beck M S, Dyakowski T and Williams R A 1998 Process
tomographystate of the art IEEE Trans. Instrum. Meas.
Control 20 16377
Boone K, Barber D and Brown B 1997 Review: imaging with
electricity: report of the European concerted action on
impedance tomography J. Med. Engng Technol. 21 20132
Brown B H 1983 Tissue impedance methods Imaging with
Non-ionising Radiations ed D F Jackson (Guildford: Surrey
University Press)
Brown B H and Barber D C (eds) 1992 Special issue on electrical
impedance tomography of Clin. Phys. Physiol. Meas. 13
(suppl A)
Brown B H, Barber D C and Tarrasenko L (eds) 1987 Electrical
impedance tomographyapplied potential tomography Clin.
Phys. Physiol. Meas. 8 (suppl A)
Brown B H, Tidy J, Boston K, Blackett A D, Smallwood R H and
Sharp F 2000 The relationship between tissue structure and
imposed electrical current ow in cervical neoplasia Lancet
355 8925
Conway J, Hawley M, Mangnall Y, Amasha H and van Rhoon G C
1992 Experimental assessment of electrical impedance imaging
for hyperthermia monitoring Clin. Phys. Physiol. Meas. 13
A1859
Duck F A 1990 Physical Properties of Tissue (New York:
Academic)
Frerichs I 2000 Electrical impedance tomography (EIT) in
applications related to lung and ventilation: a review of
experimental and clinical activities Physiol. Meas. 21 R121
Frerichs I, Hahn G, Schiffmann H, Berger C and Hellige G 1999
Monitoring regional lung ventilation by functional electrical
impedance tomography during assisted ventilation Ann. NY
Acad. Sci. 873 493505
Grifths H, Stewart W R and Gough W 1999 Magnetic induction
tomography: a measurement system for biological tissue Ann.
NY Acad. Sci. 873 33545
Grimnes S and Martinsen G 2000 Bioimpedance and
Bioelectricity, Basics (London: Academic)
Hames T J (ed) 1990 Overview of clinical applications Proc.
Copenhagen Meeting on Electrical Impedance Tomography
(Copenhagen 1990) (Shefeld: University of Shefeld)
Henderson R P and Webster J G 1978 An impedance camera for
spatially specic measurements of the thorax IEEE Trans.
Biomed. Eng. 25 2504
Holder D S (ed) 1993 Clinical and Physiological Applications of
EIT (London: UCL Press)
995
B H Brown
Kerner T, Hartov A, Soho S, Poplack S and Paulsen K 2001 Using
electrical impedance spectroscopy to image human breast:
practical considerations which inuence exam consistency
3rd EPSRC Network Meeting, UCL, London
Khambete N D, Brown B H and Smallwood R H 2000 Movement
artefact rejection in impedance pneumography using six
strategically placed electrodes Physiol. Meas. 21 7988
Korjenevsky A, Cheripenin V and Sapetsky A 2000 Magnetic
induction tomography: experimental realisation Physiol. Meas.
21 8994
Leathard A D, Brown B H, Campbell J H, Zhang F, Morice A H and
Tayler D 1994 A comparison of ventilatory and cardiac related
changes in EIT images of normal human lungs and of lungs
with pulmonary emboli Physiol. Meas. 15 A13746
McArdle F J, Turley A, Hussain A, Hawley K and Brown B H 1993
An in vivo examination of cardiac impedance changes imaged
by cardiosynchronous averaging Clinical and Physiological
Applications of EIT ed D S Holder (London: UCL Press)
pp 25768
McCann H, Yang W Q and Polydorides N P 1999 Information
retrieval by electrical capacitance tomography: evaluation of
an alternative algorithm and the importance of boundary
conditions 1st World Congress on Industrial Process
Tomography (Buxton, April) pp 20610
Meas. Sci. Technol. 1996 Special issue on process tomography 7
30815
Metherall P, Barber D C, Smallwood R H and Brown B H 1996
Three-dimensional electrical impedance tomography Nature
380 50912
Noble T J, Morice A H, Channer K S, Milnes P, Harris N D and
Brown B H 1999 Monitoring patients with left ventricular
failure by electrical impedance tomography Eur. J. Heart
Failure 1 37984
Patterson R and Zhang J 2001 Using EIT to observe cardiovascular
changes induced by the valsalva manoeuvre 3rd EPSRC
Network Meeting, UCL, London
Riu P (ed) 1999 Electrical bioimpedance methods, applications to
medicine and biotechnology Ann. NY Acad. Sci. 873
Tidswell A T et al 2001 Electrical impedance tomography of human
brain activity with a two-dimensional ring of scalp electrodes
Physiol. Meas. 22 17785
Tozer J C, Ireland R H, Barber D C and Barker A T 1999 Magnetic
impedance tomography Ann. NY Acad. Sci. 873 3539
Ultchin Y, Nachaliel U and Ori A 2001 Indirect calculation of breast
tissue impedance values 3rd EPSRC Network Meeting, UCL,
London
Vaisman N, Weintrop N, Blumental A, Yosefberg Z and Vardi P
1999 Gastric emptying in patients with type 1 diabetes mellitus
Ann. NY Acad. Sci. 873 50611
Wang M, Dicken F J and Beck M S 1992 Improved electrical
impedance tomography system and data collection protocols
Proc. 1st European Concerted Action on Process Tomography
pp 4053
Watson S J, Smallwood R H, Brown B H, Cherian P and
Bardhan K D 1996 Determination of the relationship between
the pH and conductivity of gastric juice Physiol. Meas. 17 217
Webster J G (ed) 1990 Electrical Impedance Tomography (Bristol:
Hilger)
Williams R A and Beck M S 1995 Process Tomography: Principles
Techniques and Applications (London: Butterworth
Heinemann)
Yang W Q and York T A 1999 New ac-based capacitance
tomography system Proc. IEE 146 4753
York T 2001 Status of electrical tomography in industrial
applications J. Electron. Imaging at press
996

Potrebbero piacerti anche