Biomedical Engineering Reference
In-Depth Information
Table 5.1:
Overview of aneurysm dimensions present in our database
assessed via manual measurements done by a neuroradiologist. The table
shows mean
±
SD of the mean of the manual measurements by two
observers. Aneurysms were stratified according to their location
n
Neck (mm)
Width (mm)
Depth (mm)
ACoA
16
2
67
0
98
5
73
4
18
7
27
4
22
.
±
.
.
±
.
.
±
.
ACoP
10
3
.
03
±
0
.
88
4
.
45
±
1
.
79
5
.
50
±
1
.
99
MC
10
2
.
71
±
1
.
11
5
.
40
±
2
.
01
6
.
07
±
1
.
91
Pericallosal
1
3
.
50
±
NA
8
.
27
±
NA
9
.
52
±
NA
ICA
2
3
.
13
±
1
.
46
6
.
01
±
2
.
20
6
.
63
±
4
.
47
Berling, Germany) at a rate of 3 ml/s, starting the scanning 20 seconds after the
onset of contrast administration.
The acquired images were transferred to a SGI Indigo2 workstation (Sili-
con Graphics, Mountain View, CA) for viewing and postprocessing. The man-
ual quantification of the aneurysms was performed using 2D MIP images and
measuring tools provided by the console software Omnipro (Marconi; Haifa,
Israel). The clinical parameters needed for the planning of the endovascular
intervention were the maximum neck diameter, the maximum dome width,
and maximum dome depth of the aneurysm. As it is customary in clinical
routine, the measurements were carried out along several projection angles
and, from those, the neuroradiologist chose the view-angle producing maximal
measurements.
5.3.2
Computerized Protocol
Using the marching cubes algorithm, a 3D model of the aneurysm was recon-
structed from the zero-level set of
φ
. To make the computerized measurements
comparable to the manual gold standard, the models were rendered with a view-
point selected according to the criterion of maximality used to generate the MIP
images. Two points were manually pinpointed in the 3D scene, corresponding
to the end-point of the measured magnitude from that angle. Measurements
are then performed by projecting this points into the camera plane to emulate
MIP-based measurements.
