|
ARRS,
American Roentgen Ray Society:
Virtual colonoscopy (VC) uses CT
images and a computer visualization system to virtually
navigate within a reconstructed 3-D model of the colon,
searching for polyps. There are currently two approaches
for VC. The first utilizes the actual 2-D images with
direct 3-D correlation for problem areas. The problem
with this approach is that it is frequently difficult
to differentiate oblique and longitudinally oriented
folds from small polyps. The second approach uses 3-D
endoscopic navigation of the colon with the 2-D images
used for problem solving. This latter approach mimics
actual colonoscopy. However, as with actual colonoscopy,
it is difficult for VC to ensure that all areas of the
mucosa have been evaluated. This is particularly true
in and around the haustral folds and around sharp turns
or bends.
This exhibit displays a novel approach
to solving this latter problem utilizing automatic field
of view monitoring. During fly-through guided navigation
the visualized colon surface can be marked and measured.
Counting the visible surface voxels and comparing this
number to the total of all surface voxels yields the
percentage of lumen surface visualized. After antegrade
and retrograde fly-throughs, all "patches"
of connected surface area not yet seen are identified,
measured and sorted by size. Clinically significant
patches, the smallest diameter defined being > 5mm,
are presented to the viewer, who quickly reviews each
by direct visualization by sequentially stepping through
the sorted list.
This exhibit will demonstrate how
this approach to CT colonoscopy ensures a near 100%
direct visualization of the mucosal surface. In addition,
it will be shown that this system's ease of use adds
little to the procedure time. This method of performing
VC may provide better direct visualization of the entire
colon than actual colonoscopy.
|
