Biomedical Engineering Reference
In-Depth Information
Fig. 26.8
(
a
)Stiffness
η
vs connective tissue content
f
ct
circles
and
squares
indicate the reference
and the target measuring sites, respectively. (
b
) Box plots of the relative differences in the stiffness
parameter
η
p
/η
n
divided into the categories 'desmoplasia' ('
D
'), 'other' ('
O
'), and 'lesion' ('
L
');
'
n
' stands for the reference measuring site and '
p
' for the target measuring site. The relative in-
crease in the stiffness parameter is significant and 100 % sensitive for the category 'desmoplasia'
and the 'target' site. Figure
26.8
(b) shows the box plots for the relative increase in
η
with separated evaluation of lesion type 'desmoplasia', 'other', and all 'lesion'. On
the current database, it was possible to reach 100 % sensitivity with
η
for detection
of 'desmoplasia'; the confidence interval is large, but it is still highly significant, the
median indicating at 2
.
4-fold increase in the stiffness parameter. For the selected
threshold value the corresponding specificity is 63 %.
The acquired aspiration data enabled the calibration of constitutive models, based
on the FE-model and the inverse analysis procedure. The results obtained for aver-
age observations from measurements on normal tissue were reported in Sect.
26.3.2
,
assuming a given model for liver capsule.
26.4 Discussion
26.4.1 Protocol for Open Surgery Measurements
A significant improvement was achieved in the repeatability of intraoperative aspira-
tion experiments in the present study. In all previous applications the aspiration test
was performed as combination of indentation and aspiration without measurement
of the contact force, where the former led to unknown contact and far-field boundary
conditions. This is considered as the main reason for the pronounced scatter in the
measuring data of previous studies (Nava,
2007
). The new measuring protocol min-
imizes the contact force between probe head and organ, with the surgeon following
as much as possible the motion of the liver during the measurement. This provided
better-defined and hence more repeatable kinetic and kinematic initial and boundary
conditions. It was possible to reduce the systematic error by a factor of 5 to a co-
efficient of variation of about 5 %. One other reason for better reproducibility was
the fact that always the same surgeon trained to operate the aspiration device per-
formed the measurements. This is good accuracy for intraoperative measurements,
