2. 1 .3.09
Artefacts caused by patient movement and breathing in
functional McG-Imaging
Joachim Eichhornt, U*" Lederl, Annegret Hemmannl, Jens Haueisen2,
Markus Huckl, Susanne Bluml, Peter Pohll, Maik Hoffmannl,
Helmut Ktihnertt, Huno.s Nowak2, Siegfried Mtillerr
ICardiomagnetic
Laboratory, Clinic of Internal Medicine (Department of Cardiology),
tBio-ugnetic Centre, University Hospital (Department of Neurology),
Friedrich Schiller University Jena, Erlanger Allee l0l, D-07740 Jena, Germany
Abstract: We investigated the influence of patient
movement and breathing on the reliability of
biomagnetic marker localization. Marker positions
before and after long term recordings were compared.
Our investigation of 88 measurements demonstrated a
mean deviation of 2 mm. Recording duration, artefacts
caused by body movement and breathing can be
by oil capsules in MRI. Both were put together by
a
special coordinate transformation algorithm [3].
neglected in compliant test subjects.
INTRODUCTION
Functional biomagnetic cardiac imaging needs exact
sensor positioning for source localization, especially in
successive or single long term recordings. Detection of
infrequent arrhythmias often needs recording durations of
some hundred seconds. The precision of functional source
localization (e.g. current density reconstruction on
myocardial surfaces) with composed sensor systems
depends on reconstruction algorithm, coordinate
transformation (MCG/MRI) and on the measurement
procedure. We commonly used 4 different sensor
positions for Myocardial Cunent Imaging [].
In this study we examined patient related influences
(body movement, breathing). We compared functional
biomagnetic localization before and after data aquisition.
The following questions had to be answered: How
much is the extend of patient movement during
measurement? Is it necessary to perform localization
before and after measurement or to stabilize the position
of the patient? Is there a difference between medianly and
laterally positioned body surface transformation markers?
What influence has the breathing.
METHOD
The study was performed at the Biomagnetic Centre of
Jena (50 channel twin dewar Philips MCG/MEG device
l2l, electromagnetically shielded room, sampling rate
1000 Hz). The position of the patient was determined
relative to the multichannel gradiometer system using l0
body surface markers (coilsets) fixed to certain points at
the anterior chest wall (see fig. l). Markers were replaced
O collsef
Figure
l:
coilsets and sensor positions
Before and after each MCG recording (100, 200, 600
sec.) the coilsets were localized with regard to the dewar
position. Every coilset has three orthogonal coils. A sinus
wave was recorded for localization (1.6 sec.) for each coil.
The localization result of one coilset was the average
position of 4.8 seconds (at least one breathing cycle) and
three coils.
We investigated 88 measurements (3-7 different
positions) of altogether 12 patients (2 female, l0 male,)
and 5 normal subjects (2 female, 3 male) with a mean age
of 43 + 16 years. We repeated the measuring procedure
with markers fixed on a board (same positioning pattern).
The difference (pre/post) was calculated for every
coordinate to identi$ a possible deviation in one typical
direction:
,lrrl
dev* : xpre-xport,l, dev"
dev
r:
I
I zws-zpostl
:
I
ypr.-yport
l,
.
The spatial distance was calculated for every 3D
localization (pre/post) to quantiff the artefacts caused by
patients movement or breathing:
Medical& BiologicalEngineering & Computing Vol. 34, Supplement 1, Part 2, 1996
The 1st International Conference on Bioelectromagnetism, June 9-13, 1996, Tampere, Finland
69
CONCLUSION
devtor:
Localization deviation of different anatomical postions
were compared (sternum, breast etc.). Far away situated
coilsets were excluded for the calculation of mean dev1o".
RESULTS
The mean 3D deviation of the localization pre/post
was 2
t
1.7 mm (maximum 5.6 mm).
We found an excellent mean deviation of surface
markers (about 2 mm) comparing localization before and
after measurement in test subjects. The influence of the
localization algorithm and noise can be neglected
(<<lmm). Breathing and patient movement had only
small influence on reproducibility of localization results.
The method choosen for localization in dewar coordinate
system of the patient's body is sufficient for long term
MCG recording.
REFERENCES
histogram of dev.(loc)
[1]
U. Leder, P. Pohl, F. Butkewitz, R. Huonker, R.
Rczanny, A. Michaelsen, H. Nowak, H. Ki.ihnert, S.
MUller: "Stromdichterekonstruktion auf realistischen
myokardialen Oberflächen als neue Methode des
nichtinvasiven elektrophysiologischen Imaging",
Biomed. Technik, vol. 40, pp 325-326, 1995
2,5
[2] O. Dössel,
3,5
mean of dev.(loc)
system
Figure 2: 3D deviation of localization
Fuchs,
This parameter includes all nonsystematic
errors
midpoint of sensor area). These postions were not used for
localization. Every recording position had
for
biomagnetic imaging'i Appl. Super-
l, pp. I 8 l3- l 825, 1993
M.
Fuchs,
H.-A.
Wischmantr,
M. Wagner, J. Krtiger,
[3]
coilset positions far away from the recommended
localization zone (x x 70, y N 70, z = -150 mm, origin:
,,Coordinate System Matching for Neuromagnetic and
Morphological Reconstruction Overlay", IEEE Trans.
Biomed. Eng. 1994
ACKNOWLEDGEMENT
This work was supported by the German BMBF grant
13N6736.
at least 3
dev1o".
Normal subjects had only slightly lower deviation (1.7
mm) compared to possibly less compliant patients (2.2
mm). Measurement duration had also little influence on
reproducibility of coilset position (100s: 1.2 ffiffi,
200/600s: 2 mm).
5
o
-94
t
o
Da
o
c(u.
oz
E
1
0
IDW
%%
1,00
2,00
6.00
number of trials
Figure 3: Measurement duration and 3D deviation
There was no higher spatial difference for female
subjects. Markers fixed on a board had a deviation of less
than 0.04 mm (2 %o of patient's mean devloJ.
70
J. Krtiger, K.M.
conductivity. vol.
(noise, patient movement, breathing). The subtraction of
coordinates (dev", devr, devr) indicated no significant
deviation in one direction.
Runaway values (deviation > 5 mm) occured in
coilset with a small
B. David, M.
Liidecke and H.A. Wischmann: "A 3l-channel SQUID
Medical & Biological Engineering & Computing Vol. 34, Supplement 1, Part 2, 1996
The 1st International Conference on Bioelectromagnetism, June 9-13, 1996, Tampere, Finland