Note: Descriptions are shown in the official language in which they were submitted.
CA 02492903 1998-05-05
s ELECTRICAL IMPEDANCE METHOD AND APPARATUS FOR DETECTING AND
DIAGNOSING DISEAES
FIELD OF THE INVENTION
The present invention relates to a method and apparatus for detecting or
io diagnosing disease states in a living organism by using a plurality of
electrical
impedance measurements.
BACKGROUND OF THE INVENTION
is Methods for screening and diagnosing diseased states within the body are
based
on sensing a physical characteristic or physiological attribute of body
tissue, then
distinguishing normal from abnormal states from changes in the characteristic
or
attribute. For example, X-ray techniques measure tissue physical density,
ultrasound
measures acoustic density, and thermal sensing techniques measures differences
in
2o tissue heat. Another measurable property of tissue is its electrical
impedance; i.e., the
resistance tissue offers to the flow of electrical current through it. Values
of electrical
impedance of various body tissues are well known through studies on intact
humans or
from excised tissue made available following therapeutic surgical procedures.
In
addition, it is well documented that a decrease in electrical impedance occurs
in tissue
2s as it undergoes cancerous changes. This finding is consistent over many
animal
species and tissue types, as summarized by Pethig and Kell'. Human breast
cancers,
in particular, have shown similar changes in studies such as those of
Chaudhary et a12
and Surowiec et a13. Both groups examined surgically excised normal and
malignant
human breast tissue and obtained similar results; i.e., on average, the
electrical
3o impedance of breast cancer tissue was about one-third that of the normal
surrounding
breast tissue.
Electrical impedance imaging has been proposed to create a picture of
electrical
impedance differences within a body region4.5.6 much as an X-ray provides a
picture of
3s differences in physical density. One of the incentives to do so is the
potential
application of electrical impedance imaging as a screening technique for
breast cancer,
either as a replacement of or supplement to X-ray mammography. Mammography has
reasonable sensitivity for detecting abnormalities when present, but the
technique fails
CA 02492903 1998-05-05
s to detect about 5 to 15% of breast cancers. This is due to several factors,
including
concealment of the cancer by overlying normal, but dense, breast tissue,
failure of
mammography to image certain portions of the breast, as well as errors in
perception.
Mammography has relatively low success for distinguishing malignancies from
other
abnormalities, and of the approximately 500,000 breast biopsies performed in
the
to United States each year because of an abnormality detected on mammography,
only
15 to 30% of the biopsies reveal cancer. This lack of specificity not only
results in
needless anxiety and an unnecessary procedure, but adds a significant cost to
the
breast cancer screening program.
is There have been a number of reports of attempts to detect breast tumors
using
electrical impedance imaging. 45,6,7,8,9,10 However, there are basic problems
when trying
to construct an image from impedance data. The paths through tissue of X-rays
are
straight lines. In contrast, electrical current does not proceed in straight
lines or in a
single plane; it follows the path of least resistance, which is inevitably
irregular and 3
2o dimensional. As a result, the mathematics for constructing the impedance
image is
very complex and requires simplifying assumptions that greatly decrease image
fidelity
and resolution. Not surprisingly, in view of the image reconstruction
difficulties, either
no clinical data were published in any of these reports, or if they were, the
images were
of low resolution and difficult to interpret.
2s
SUMMARY OF THE INVENTION
The present invention scans for the presence or absence of breast
abnormalities,
particularly benign and malignant tumors. While not intending to be bound by
any
3o particular theory, the method of the invention may arise from the following
assumptions
and hypotheses:
1. The tumor will occur in only one breast.
2. Both breasts are structurally similar, and therefore can be expected to be
3s approximate mirror images with respect to their impedance characteristics.
3. If impedance measurements are taken in a multiplicity of directions or
paths across
the breast (I call this an impedance scan), the presence of tumors, which are
known
2
CA 02492903 1998-05-05
s to have a significantly lower impedance than the normal tissue they replace,
will
distort or change the impedance in at least some of the paths of current flow.
4. The magnitude of decreased impedance is greater for malignant tumors than
for
benign ones, providing a method for differentiating between these tumor types.
5. There will always be some differences in impedance between breasts in the
normal
to individual; but these differences will be less than the differences when a
cancer is
present. In additional, cancer differences may have unique characteristics.
This
concept of "difference between differences" (impedance difference between
breasts
without a cancer versus impedance difference between breasts with a cancer) is
one of the novel methods used for data analysis.
is
Impedance data from scanning both breasts are compared and differences are
displayed and analyzed. The use of impedance differences "subtracts out" an
otherwise complex and voluminous amount of impedance data produced by most
current paths that, while irregular and 3 dimensional, are nevertheless
substantially
2o impedance-identical because the paths were virtually identical in both
breasts. The
differences that remain are much more manageable analytically, and can be used
to
identify abnormalities.
An important aspect of the invention is the electrode array, designed and
fabricated
2s so that electrode position and spacing are, as closely as possible,
identical in the two
breasts in order not to introduce artifactual differences related to the array
itself.
Another aspect of the invention that contributes to the simplicity and
reliability of the
impedance scanning method, is the design and implementation of the apparatus
for
acquiring, displaying, and analyzing the data.
Whereas a primary objective of the present invention is to provide a novel and
improved method and apparatus for detecting and locating breast cancers, the
invention can also be applied to other diseases or conditions in which there
is a
distinguishable difference in electrical impedance in the tissue as a result
of the disease
3s or condition. For example, the occurrence of a deep venous thrombosis in
the thigh or
leg would cause a change in the circulatory dynamics which would be reflected
by a
change in the electrical impedance of the affected region.
3
CA 02492903 1998-05-05
s A further objective of the present invention is to provide a novel and
improved
method and apparatus for detecting and locating diseases or conditions in any
region of
the body in which the electrical impedance of the region containing the
disease or
condition can be compared to an essentially identical, normal body region; for
example,
right and left forearms, right and left thighs, or right and left calves.
io
A still further objective of the present invention is to provide a novel and
improved
method and apparatus for detecting and locating diseases or conditions in any
region of
the body in which the electrical impedance of the region containing the
disease or
condition can be compared to another normal body region that, while not
entirely
is identical, is consistently and constantly different; for example, right and
left sides of the
abdomen.
DESCRIPTION OF THE DRAWINGS
zo FIG. 1 is an illustration of the four electrode impedance measurement
technique;
FIG. 2 is an illustration of the breast electrode array of the invention;
FIG. 3A is an illustration of a positioning template for the breast electrode
array;
FIG. 3B is an illustration of a positioning ring for the breast electrode
array;
FIG. 4 shows an implementation of lead wiring for the breast electrode array;
2s FIG. 5 is a block diagram of the apparatus of the invention;
FIG. 6 is a plot of an impedance matrix, a method of the present invention,
obtained from a subject with cancer in the left breast;
FIG. 7 is a bar plot of the 28 Zgame subset of impedances, obtained from a
subject
with cancer in the right breast;
3o FIG. 8 illustrates the four possible chord lengths of a circle when there
are eight
equally spaced points on its circumference;
FIG. 9 is a chord plot of the 28 Zsame subset of impedances for the same
subject
used for FIG. 7. The breasts are symbolized as circles and the 28 impedances
have
been normalized, their anatomic direction represented by chords of a circle,
and their
3s magnitude represented by the density of shading;
FIG. 10 is a further refinement of FIG. 9 in which mirror image matching
impedance
chords have been removed; and
4
CA 02492903 1998-05-05
s FIG. 11 is a diagnostic algorithm used for breast cancer screening based on
impedance scanning.
DETAILED DESCRIPTION OF THE INVENTION
io Electrical Impedance and the Four Electrode Measurement Techni4ue.
Electrical impedance is most accurately measured by using four electrodes as
shown in FIG. 1. The outer pair of electrodes 1 is used for the application of
current 1,
and the inner pair of electrodes 2 is used to measure the voltage V that is
produced
is across the tissue (or generally, material) 3 by the current. The current I
flowing
between electrodes 1 is indicated by the arrows 4. The impedance Z is the
ratio of V to
I; i.e., Z = VlI. It is well known that using separate electrode pairs for
current injection
and voltage measurement produces a more accurate measurement of impedance
because polarization effects at the voltage measurement electrodes are
minimized.
ao
Impedance consists of two components, resistance and capacitive reactance (or
equivalently, the magnitude of impedance and its phase angle). Both components
are
measured, displayed, and analyzed in the present invention. However, for the
purpose
of explanation of the invention, only resistance will be used and will
interchangeably be
2s referred to as either resistance or the more general term impedance.
The Breast Electrode Arrax
FIG. 2 discloses a breast electrode array 5 of the invention that has eight
electrode
3o pairs, each pair consisting of an outer electrode 8 for current injection
and an inner
electrode 9 for voltage measurement. The illustrated implementation of the
array has a
main section 6 and a tail section 7. Eight pairs of rectangular electrodes in
circular
orientation are shown, but there are many alternatives that could be
advantageously
used with the present invention: more electrode pairs; different electrode
shapes; other
3s shapes for the main body and tail sections of the array; and other
geometrical
arrangements of the electrodes, e.g., radial sectors with three or more
electrodes.
Regardless of the electrode arrangement, four electrodes must be used for each
s
CA 02492903 1998-05-05
s impedance measurement, two outer electrodes between which current is
injected, and
two inner electrodes at which voltage is measured. The electrodes are attached
to the
skin side of the main section 6 of the array 5 and are made of an electrically
conductive, self adhesive material so that when the array is positioned on the
skin and
pressed against it, the adhesive quality of the electrodes assures good skin
fixation.
io Alternatively, additional adhesive material can be used at various
positions on the main
section 6 and/or the tail section 7 of the array. In order to assure impedance
is
measured in all regions of the breast, electrode arrays 5 are made in
different sizes for
use in women with different breast cup sizes.
is For clarity of description, lead connections for the electrodes are not
shown in FIG.
2. The material used for the main section 6 of the array 5, and to a lesser
degree for
tail section 7, may be flexible to allow the array to conform to the shape of
the breast.
Shape conformity is further aided by cutouts or darts 11 that allow the
material to
overlap and thereby prevent it from crumpling and possibly lifting part or all
of an
2o electrode off the skin. More accurate and consistently identical
positioning of electrode
arrays on both breasts is aided by the index marks 10 shown at four locations
on the
inner edge of the main section 6. Before applying an array to the skin surface
of the
breast, a transparent, flexible positioning template 12, shown in FIG. 3A., is
positioned
on the breast with central cutout 15 centered about the nipple and the
template rotated
2s so that crosshair line 16 is aligned with the body's vertical axis and
crosshair 17 is
aligned with the body's transverse (horizontal) axis. An ink or other mark is
made on
the skin surface through the cutouts 14 of the positioning template 12. The
template is
then removed and the electrode array is applied with its index marks 10 shown
in FIG.
2 aligned with the ink marks on the skin. An alternative embodiment of a
device for
3o positioning the breast electrode array, an array positioning ring 18, is
shown in FIG. 3B.
It consists of a ring 19 that has an outer diameter equal to the distance
between the tips
of diametrically opposed index marks 10 of FIG. 2. The ring 19 of FIG. 3B has
four
notches 20 corresponding to each of the index marks 10 of FIG. 2. Fine
crosshairs 21
extend from the inner side of ring 19 between diametrically opposed notches,
with the
3s central junction of the crosshairs serving to center the array positioning
ring 18 over the
nipple.
6
CA 02492903 1998-05-05
s FIG. 4 shows an implementation of electrode lead wiring for the breast
electrode
array. Electrically conductive material is deposited or otherwise applied to
both surfaces
of the main section 6 (the darts 11 are not shown in this drawing) and the
tail section T
of the array as follows: On the skin side of main section 6 the conductive
material has
the form of "U" shaped electrode connection areas 22 and 23 for outer and
inner ring
to electrodes respectively which are attached in position over these
connection areas.
Many other shapes could be used for the connection areas, the primary
consideration
being a large enough area to ensure low resistance electrical continuity with
the skin
elctrodes. Fine, conductive pathways (or leads) 24 and 25 are applied to the
non-skin
side of main section 6 and tail section 7. The flexible material from which
the main and
is tail sections are fabricated is non-conductive, and so insulates conductive
pathways 24
and 25 from electrode connection areas 22 and 23. At one end each lead 24 and
25
penetrates through main section 6 and is soldered to (or otherwise electrially
attached
to) the electrode connection areas 22 and 23 respectively. At their other end
the leads
bunch in the tail section 7 to make individual electrical contact with
conductive fingers
20 26 to form a ribbon type connector 27.
Alternative embodiments of electrode arrays are possible that would not
necessitate the attachment of adhesive electrodes to the subject's skin. For
example,
the subject could lay prone on a table with an opening for the breasts to fall
freely
2s downward. A flat plate, or cone, or other shaped holder with an array of
electrodes on
its upper surface, could then be moved upward, guided by landmarks on the
breast
and/or chest wall, to compress the breast to the extent required for good
electrode
contact. Further compression may serve beneficially to bring a tumor closer to
the
electrodes, or create a breast shape more conducive to analysis. A variant of
this
3o method would position the subject between 0 and 90°, say at
45° to the horizontal,
again allowing the breasts to fall through an opening with, in this
embodiment, a shelf at
a suitable angle, say 45°, to guide the breasts. Another variant of
this method would
have the subject erect, as in conventional X-ray mammography, and use, for
example,
mediolateral oblique and craniocaudal compression, as in conventional X-ray
3s mammography procedure, but with electrode arrays in the compression plates.
Acquiring Impedance Data
CA 02492903 1998-05-05
FIG. 5 discloses a basic block diagram of the data acquisition and analysis
apparatus 29 for automatically obtaining, processing and analyzing impedance
measurements. For the purposes of illustration, the apparatus 29 will be
described as
employed for screening, locating and diagnosing breast cancer. However, it
should be
io recognized that the method and apparatus of the invention can be employed
in a
similar manner for screening or diagnosis at other body sites and for other
conditions
and diseases. A breast electrode array 5 of 16 electrodes arranged as two
concentric
circles of eight electrodes each is show in FIG. 5. Conventional ECG
monitoring
electrodes, cut down to appropriate size with the metal tab connector intact,
were used.
is A plurality of leads 28 is intended, in this illustration, to number 16,
one for each
electrode. More generally, more electrodes, other electrode arrangements, and
other
electrode types, could be used in this application to produce more detailed
and useful
results. Also, the device and method of this invention contemplate the use of
a variety
of electrode arrays and leads, depending on other applications for which the
apparatus
20 29 is used.
In preparation for the acquisition of impedance data, the subject lies supine
on an
examining table and the skin in the area the array will be placed on is gently
cleansed
with alcohol to debride the surface. The breast electrode array 5 is carefully
oriented
2s with respect to body axes and centered about the nipple as previously
described, and
then lightly pressed against the skin over each electrode to fix the array to
the skin.
The breast electrode array 5 as described for the present invention may not be
reusable, in which case another array would be used for the second breast. It
may,
however, be possible to modify the design of the array to allow more than one
use.
As previously described, the four electrode technique is used to measure
electrical
impedance. Continuous 50 kilohertz sine wave current is injected between two
of the
eight electrodes in the outer ring of the breast electrode array 5. Use of
this frequency
and waveform is standard practice for many bioimpedance applications, but
there is an
3s extended range of useable frequencies and, to a lesser degree, other
waveforms. For
the right breast, the electrode pairs are numbered clockwise 1 to 8, and for
the left
breast, electrode pair numbering is counterclockwise so that mirror-imaged
electrode
s
CA 02492903 1998-05-05
s pairs will always be compared. The injected current, whose amplitude is low
enough
that it is imperceptible, creates electric field potentials (voltages)
throughout the entire
breast and adjacent chest wall. In particular, it creates voltages at the
eight electrodes
in the inner ring of the electrode array. Measuring the voltage difference
between any
two of the inner electrodes, and dividing it by the value of the current
injected between
io the two outer electrodes gives, by Ohm's law, the value of the impedance.
For
example, if current 1~,3 is defined as being applied between the outer
electrodes of
electrode pairs 1 and 3, and voltage V~,~ is measured between the inner
electrodes of
electrode pairs 1 and 7, the resultant impedance Z is
Z = V,,~
~t,l
The apparatus 29 is multichannel device that is connected to electrode leads
28
from the breast electrode array 5. A central control unit, consisting of
central
2o processing unit (CPU) 32 and RAM and ROM memories 33 and 34, selects in
rapid
succession, one set of four electrodes at a time, a multiplicity of sets of
two outer
electrodes for current injection and two inner electrodes for voltage
measurement to
perform what I call an impedance scan. Four electrode impedance measurement is
made by the conventionally designed impedance section 30. The analog-to-
digital
2s (A/D) converter 31 is of known type, and converts the analog impedance
measurement
to a digitized form. Depending on the number of array electrodes, more than
one A/D
converter may be needed. Digital input data from the A/D converter 31 are
processed
by the CPU, where they undergoing real time analyses for error checking,
routing to the
monitor 35 for display of raw or processed data, as well as storage in memory
for
so further analysis and output to the monitor 35 and printer 36.
The total number of possible combinations of two current electrodes and two
voltage electrodes is very large. However, mathematical and electrical circuit
theory
can show that there are only 49 such combinations that are independent and
that all
3s other combinations can be calculated from the set of 49. This set is
obtained as
follows: Current is applied between the outer electrodes of electrode pairs 1
and 2 and
then, in turn, the voltage between the inner electrode of electrode pair 1 and
all other
inner electrodes are measured, i.e., V~,2, V~,3 ... V~,s. Dividing each of
these voltages by
1~,2, the current between outer the electrodes of electrode pairs 1 and 2,
gives the first
9
CA 02492903 1998-05-05
s seven impedance values. Current is next applied between the outer electrodes
of
electrode pairs 1 and 3, 1~,3, which will create a new pattern of electric
field potentials.
Then, the voltage is again measured between the inner electrode of electrode
pair 1
and all other inner electrodes (V~,2, V~,3 ... V~,6). Dividing each of the
voltages by 1~,3
gives the next seven impedance values. This process is repeated for current
applied
Io between the outer electrodes of electrode pairs 1 and 4, 1 and 5, 1 and 6,
9 and 7, and
1 and 8, to produce, finally, seven sets of seven impedance values. Placing
these
impedance values (elements) in a 7-row by 7-column grid results in what I call
the
impedance matrix.
is There is a special subset of 7 impedance values in the 49 element set -
those that
use the same pair of electrodes for current injection and voltage measurement;
for
example, current I ~,3 applied between the outer electrodes of electrode pairs
1 and 3,
and voltage V~,3 measured between inner electrodes of the same electrode pairs
gives
impedance
z = v,,3
i1,3
2s I call impedances in this Subset Zsame type impedances. They are:
Z1,2 Z1,3 z1,4 Z1,5 Z1,6 Z1,7
There is additional value, as will be disclosed under Data Analysis, in
measuring all
3o possible ZSame impedances, another 21 measurements, as listed below:
Z~.3 Z2,4 Z2,5 ZZ,6 Z2,7Z~,B
Z3.4 Z7.5 Z3,6 Z3,7Z3,B
Z4,5 Z4,6 Z4,7Z4,8
Z5,6 Z5.7Z5,8
35 Z6,d
Z7,8
Therefore, a complete set of impedance measurements for one breast, when the
4o illustrated eight pair electrode array is used, consists of 49 measurements
for the
impedance matrix, and another 21 measurements to obtain all values for
ZSa~,,e,
resulting in a total of 70 impedance measurements for each breast. I call the
organized process of selecting lead sets and obtaining these measurements an
io
CA 02492903 1998-05-05
s impedance scan. As the values are being acquired, their accuracy and
reliability are
checked in real time by a novel error detection program in the central control
unit that
uses algorithms based on 1 ) expected values of impedances related to their
position in
the matrix, 2) expected ratios of the resistive and reactive components of
impedance,
and 3) a comparison of the 21 measured Z~me values listed above to their
calculated
to values derived from the impedance matrix.
Data Analxsis
In the method of this invention, the breast is considered as a non
homogeneous,
is electrically conducting object with M + 1 electrode pairs (to be referred
to in this
discussion simply as an "electrode," one that can be used for both current
injection and
voltage measurement without electrode polarization). I assign one electrode as
the
reference electrode with zero potential. The current at the reference
electrode is the
sum of the currents that are applied to the other M electrodes. The impedance
matrix Z
2o relates the currents li, the current through the ith electrode, and the
voltages Vi, the
potential difference between the ith electrode and the reference electrode,
where i =
1,2,3,...,M, as follows:
Vi h
Vz IZ
V3 Ia
- Z
x
VN IN
30
which can be condensed as V = Z x I.
For an object with M + 1 electrodes as described above, the impedance matrix Z
is
defined as an M x M matrix:
11
CA 02492903 1998-05-05
Z11Z122'13 . Z1M
"'21x'22Z'23 ' 'Z2M
'
x'31x'32Z'33 Z'3M
ZMlT'M2Z'M3
Each matrix element Zij (i,j = 1,2,3,...,M) is equal to Vi/lj when all
currents except
the current at the jth electrode are equal to zero. In a given subject, the
impedance
matrix Z is unique for a given pattern of breast electrodes and therefore
represents the
to "signature" of the breast. Once available, the Z matrix (or R matrix, or Xc
matrix, if Z is
resolved into its resistive and capacitive reactive components) can be used in
the
following ways in the present invention to screen for or diagnose disease:
1. Perform pattern recognition analyses on the matrix;
is 2. Examine the determinant, eigenvalues and eigenvectors of the matrix; and
3. Compute the joule losses for each breast by evaluating P = IT x Z x I*
where T
denotes matrix transposition and * denotes complex conjugation. ff P is
significantly
lower for one breast, it may indicate that a cancer is present in the breast
with the
lower P.
The method of the present invention recognizes that an impedance matrix is not
an
image of the structure of the underlying breast, and indeed the complexity and
impracticality of attempting to construct impedance images is purposely
avoided.
Instead, a relatively simple new test procedure called an impedance scan is
performed
2s in an organized fashion with inventive devices to ensure accuracy and
reproducibility of
results in general and precise mirroring of the procedure between two sides or
two
regions in particular. Impedance values are obtained in a manner that allows
them to
be organized into an impedance matrix. These values, the elements of the
matrix,
depend on the electrical characteristics of the underlying structures. Any
electrical
3o change or abnormality in a structure will produce a change in the impedance
matrix.
And if there is a (reasonably) identical normal structure or region, detection
and
diagnosis of the difference produced by the abnormality is greatly simplified
by
comparison with the normal structure and by clinical experience that has
previously
related the difference with the disease state that produced it.
12
CA 02492903 1998-05-05
A pilot study to examine the performance of the described invention was
performed
on two groups of subjects. The first group (biopsy group) consisted of100
subjects
scheduled for breast biopsy because of a suspicious area in one breast
revealed by X-
ray mammogram. Immediately before the biopsy, impedance scanning was performed
to on both breasts. The pathology report on the biopsy material was obtained
later. The
second group of 20 subjects (control group) were randomly selected from those
undergoing routine yearly screening mammograms. All were confirmed to have no
abnormalities. An example of a graphical plot of an impedance matrix
(specifically, the
resistive component) in a subject found to have an infiltrating ductal
adenocarcinoma in
is her left breast is shown in FIG. 6. The x-axis indicates the selection of
outer electrodes
for current injection; e.g., between electrodes 1 and 2, between 1 and 3, ...
between 1
and 8. As described previously, for each such selection, seven values of
impedance
are obtained by measuring voltages between electrodes 9 and 2, 1 and 3, ... 1
and 8.
In the graph, these seven impedance values are shown in order proceeding from
left to
2o right for each current selection, with impedance (resistance) magnitude
indicated by the
y-axis height. As predicted by the known decrease in impedance of malignant
tissue,
the matrix plot for the breast with a cancer has lower impedances. This effect
is seen
throughout in this example; in other subjects, it occurred in most, but not
all regions.
2s Associated with certain types of matrices, including the impedance matrices
as
structured in the present invention, are values called eigenvalues and vectors
called
eigenvectors. These words are Anglo-German hybrids which use the German
"eigen"
for characteristic or particular. Characteristic or particular in the sense
that by
mathematical analysis each 7 x 7 impedance matrix can be represented by a set
of
3o seven numbers, i.e., seven eigenvalues, that are unique to that matrix.
Furthermore,
associated with each of these numbers (eigenvalues) is a unique, 7D vector,
its
eigenvector. Since the eigenvalues and eigenvectors characterize the matrix,
and the
matrix in turn is sensitive to tissue changes resulting from disease, an
object of the
present invention is the use of eigenvalues and eigenvectors as a means of
detecting
3s and diagnosing disease states. The number of eigenvalues and eigenvectors
available
for this purpose will vary with the size of the impedance matrix, increasing
as the
number of electrodes used in the array becomes larger.
13
CA 02492903 1998-05-05
Another object of the method of the present invention is to use the special
set of
impedance values, referred to previously as Zsa,r,e, as a means of detecting
and
diagnosing disease states. This set, for a 7 x 7 matrix, has 28 values and is
shown
below. Larger matrices will have correspondingly larger sets. The 28 Zsame
values
to from both breasts can be plotted and compared; for example, in the bar plot
Z~,= Z~,~ Z,,q Z,,S Z,,6Z~,7Z,,e
Z2,3 Z2,q Z2,5 Z2,6Z2,7Z2,8
Z3,q Z3.5 Z1~6Z3.7Z3.8
zq,5 Zq,bZq,7Zq,e _
Z5.6,7 Z5,8
5
Z
7 76.e
~
~.7
Z7.g
of FIG.7, obtained in the pilot study from another subject found to have early
cancer of
the right breast (intraductal carcinoma). Most of the bars in the breast with
the cancer
are lower in amplitude (smaller impedance) than corresponding bars on the
normal
side.
2s Yet another method of the present invention for the display and analysis of
Zee".,e
data is as follows: First, all impedance values are normalized to cancel out
the effect
on irnpedance of differences in electrode separation. For example (viewing
breast
array 5 in FIG. 5), even if the tissue impedance is constant throughout, the
impedance
reading between electrodes 1 and 5 will be larger than the value between
electrodes 1
3o and 2 because of the greater spatial separation between the former. As
shown in FIG.
8, for an eight pair, circular electrode array there are four separation
distances or
chords: a 22.5° chord 37, a 45° chord 38, a 67.5° chord
39, and a 90° chord 40. The
effect of electrode separation on impedance can be cancelled out and an
estimate of
normalized or specific impedance obtained (impedance per unit length) by
setting the
3s distance between adjacent electrodes at 1 and retaining the measured value
of
impedance between adjacent electrode pairs (22.5° chords), but dividing
the
impedance measured between 45° chords, 67.5° chords, and a
90° chords by 1.85,
2.41 and 2.61 respectively. The divisors were obtained from basic
trigonometric
calculations. Next, the minimum and maximum values of the 56 normalized
impedance
4o values (28 per side) for a given subject are used to define the impedance
range for that
subject, and the range is then subdivided into eight equal smaller size
ranges, or bins.
14
CA 02492903 1998-05-05
s Using as an example the same subject from FIG. 7 (right breast cancer), it
can be seen
in FIG. 9 that the range of normalized impedances varied from 85 to 292 ohms,
giving
eight bins of 85 to 110, 111 to 136, 137 to 162, ... and 267 to 292 ohms. The
right and
left breasts are symbolized in FIG. 9 by circles, and the impedances between
electrode
pairs are represented as wide lines or chords. The magnitude of each impedance
is
to indicated by the density of grey shading (or alternatively by color) with
the the lowest
impedances (those within the 85 to 110 ohm bin) assigned black, impedances at
the
next level (the 111 to 136 ohm bin) are assigned dark grey, and so on until
the highest
impedances (the 267 to 292 ohm bin) which are assigned white. I call FIG. 9 a
chord
plot of normalized impedances.
is
By calculating normalized impedance to more closely represent impedance per
length, displaying the results as shaded chords across a circle, as in FIG.9,
then
removing all mirror image matching impedance chords, leaving only differences
between the two sides, as shown in FIG. 10, further extends the usefulness of
the
2o method. It is now apparent that not only are the lower impedances in the
right breast,
but they are clustered at the inner aspect of the breast. This finding was
confirmed by
the X-ray mammogram for this subject - the cancer was in the mid inner right
breast. I
call FIG. 10 a match-removed chord plot of normalized impedances.
2s Another, yet further extension of the disclosed method becomes necessary on
observation of difference Zsar"e impedance scans as displayed in FIG. 10 but
for normal
subjects. They, too, have some differences because there are always small
anatomic
or physiologic side-to-side differences in homologous body structures. This
required
the establishment of a database in which side-to-side impedance differences
were
3o examined when both breasts were normal (control group) and when one of the
breasts
had a cancer (cancer group). This was done in order to provide a set of
statistics to
distguish between impedance scans from normal subjects and those in which
there
was an underlying disease condition. The statistics are:
3s 1. Mean of Algebraic Differences: The algebraic side-to-side difference
beween each
of the 28 Zsame impedance values is taken, and the mean algebraic difference
for the
28 values is calculated.
is
CA 02492903 1998-05-05
s 2. Mean of Absolute Differences: The absolute side-to-side difference beween
each of
the 28 Zsame impedance values is taken, and the mean absolute difference for
the 28
values is calculated.
3. Number of Matches: The number of mirror image matching impedance chords,
referred to in the discussion of FIG.10. A related statistic, the Number of
to Mismatches = 28 - Number of Matches.
4. Mean of Algebraic Differences/ Number of Mismatches.
5. Mean of Absolute Differences/ Number of Mismatches
Confidence intervals about these means are calculated based on measures such
is as one or two standard deviations. Some or all of these statistics, and/or
other ones
derived from the impedance matrix, eigenvalues or eigenvectors, are used as
part of a
diagnostic algorithm for breast cancer screening, as shown in FIG. 11. The
mean side-
to-side difference statistics generated by the impedance scan are examined
(preferentially by computer) to determine whether a significant number of the
means or
2o certain subsets of means, weighted or otherwise, fall within the confidence
limits of
normal or cancer groups. If the determination is for the normal group, the
impedance
scan is normal; if the determination is for the cancer group, a breast cancer
is
suspected, and the match-removed chord plot of normalized Zsame impedances is
obtained to indicate tumor location, followed by other procedures such as X-
ray
2s mammography.
3s
16
CA 02492903 1998-05-05
' Pethig, R., and D.B. Kell. The passive electrical properties of biological
systems: their significance in physiology, biophysics
and biotechnology. Phys. Med. Biol. 32:933-970, 1987.
2 Chaudhary, S.S., R.K. Mishra, A. Swarup, and J.M. Thomas. Dielectric
properties of normal 8c malignant human breast
tissues at radiowave and microwave frequencies. Indian J. Biochem. Biophys.
21:76-79,1984.
3 Surowiec, A.J., S.S. Stuchly, J.R. Barr, and A. Swarup. Dielectric
properties of breast carcinoma and the surrounding tissues.
IEEE Trans. Biomed Engng. 35:257-263, 1988.
° Jossinet, J., C. Fourcade, and M. Schmitt. A study for breast imaging
with a circular array of impedance electrodes. Proc. Yth
Int.Conf. Bioelectrical Impedance, 1981, Tokyo, Japan, 83-86.
5 Jossinet, J.C., and E. Mbock-Mbock. Technical implementation and evaluation
of a bioelectrical breast scanner. Proc. 10th Int
I 5 Conf. IEEE Engng. Med Biol., 1988, New Orleans, USA, (Imped. Imaging II).
6 Skidmore, R., J.M. Evans, D. Jerkins, and P.N.T. Wells. A data collection
system for gathering electrical impedance
measurements from the human breast. Clin. Phys. Physiol. Meas. 8:99-102, 1987.
' Morimoto, T., S. Kimura, Y. Konishi, K. Komaki, T. Uyama, Y. Monden, Y.
Kinouchi, and T. Iritani. A study of the electrical
bio-impedance of tumors. J. Invest. Surg. 6:25-32, 1993.
a Pipemo, G., E.H. Frei, and M. Moshitzky. Breast cancer screening by
impedance measurements. Front. Med Biol.
Engng.2:111- 117, 1990.
9 Man, B., B.D. Sollish, M. Moshitzky, Y. Choukron, and E.H. Frei. Results of
preclinical tests for breast cancer detection by
dielectric measurements. Xlllnt. Conf. Med Biol. Engng. 1979, Jerusalem,
Israel. Springer Int., Berlin, 1980, 30.4
'°Sollish, B.D., E.H. Frei, E. Hammerman, S.B. Lang and M. Moshitzky.
Microprocessor-assisted screening techniques. Isr. J.
Med Sci. 17:859-864, 1981.
17
