By Jaime E Zelaya Jr MDPhD Student PhD Year I OHSU School of Medicine amp Department of Biomedical Engineering Dr Sandra Rugonyis Lab zelayaohsuedu Image Parameters To protect patient privacy I excluded the patient demographic information from the abdominal CT scan images U ID: 914401
Download Presentation The PPT/PDF document "Information Needed for Segmenting Patien..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Information Needed for Segmenting Patient-Specific Abdominal Aortic Aneurysms Using Serial CT Scan Images with Contrast
By: Jaime E. Zelaya, Jr.
MD/PhD Student, PhD Year I
OHSU School of Medicine & Department of Biomedical Engineering
Dr. Sandra
Rugonyi’s
Lab
zelaya@ohsu.edu
Slide2Image Parameters
To protect patient privacy, I excluded the patient demographic information from the abdominal CT scan images. Unfortunately, in doing so, other important image information, i.e. image length, is excluded. Thus, I will provide the image parameters specific to the patient of interest on this slide. Should you need more information about the images, please feel free to e-mail me.
You may start segmenting at image 41 or a few images before and stop segmentation at the iliac bifurcation which starts at about image 201.
Spacing between slices
: 1mm
DFOV (Dual Field of View):
376mm x 376mm
Tilt
: 0
Magnification
: 1.0x
Image matrix
: 512 pixels x 512 pixels
A few parameter derivations I made include:
Pixel size
is determined by dividing the DFOV length by the number of pixels. In this case, the pixel size is: 376mm/512pixels = 0.734375mm/pixel
The
number of pixels in the z-direction
can be found by using the slice spacing length and dividing it by the pixel size: 1mm/(0.734375mm/pixel) = about 1.36 pixels.
Voxel
size ratios in the x, y, and z-direction
: 1:1:1.36 (obtained from relationship of the distances of each pixel in the respective directions, 0.73mm:0.73mm:1mm).
Slide3CT Scan Images Showing Contours used to Obtain Patient-Specific Surface and Fluid Geometrical Models
This is a transverse CT scan of a patient’s abdomen. The aortic aneurysm is the ellipsoid structure located at the center of the image. Other nearby structures are identified
.
Anterior
Anterior
Anterior
Right Kidney
Right
Right
Left Kidney
Left
Left
Posterior
Posterior
Posterior
Abdominal Aortic Aneurysm (AAA)
AAA Lumen
Calcifications in AAA wall
Vertebral Body & Spine
AAA Thrombus
The surface contour of the abdominal aortic aneurysm is traced in yellow. To construct surface geometry, I traced the surface contours on every CT scan cross-sectional image and used the coordinates defined in the contour.
The lumen contour of the abdominal aortic aneurysm is traced in red. To construct the lumen geometry, I traced the lumen contours on every CT scan cross-sectional image and used the coordinates defined in the contour.
Slide4Abdominal Aortic Aneurismal Surface Shell Geometry Obtained for the Given Patient’s Data in Different Perspectives
Top Anterior View
Right View
Slide5Abdominal Aortic Aneurismal Surface Shell Geometry Obtained for the Given Patient’s Data in Different Perspectives Continued… . . .
Posterior View
Left View
Slide6Abdominal Aortic Aneurismal Surface Shell Geometry Obtained for the Given Patient’s Data in Different Perspectives Continued… . . .
Bottom Posterior Right View
Note that the structure has open ends.
Slide7Abdominal Aortic Aneurismal Lumen
Shell Geometry Obtained for the Given Patient’s Data in Different
Perspectives
For this patient, the lumen geometry obtained does not look significantly different in shape when compared to the AAA surface geometry.
Top Anterior View
Right View
Slide8Abdominal Aortic Aneurismal Lumen
Shell Geometry Obtained for the Given Patient’s Data in Different
Perspectives Continued . . .
Posterior View
Left View
Slide9Abdominal Aortic Aneurismal Lumen
Shell Geometry Obtained for the Given Patient’s Data in Different
Perspectives Continued . . .
Bottom Posterior Right View
Note that the
lumen geometry
also has
open ends.
Slide10Some Difficulties Encountered during the Contour Trace Procedure
Sometimes the outer wall is difficult to distinguish from the CT scan because of limited resolution.
In such a case, one can draw a contour with a slight distance away from the lumen boundary defined by the contrast.
Difficult to Discern AAA Outer Wall Boundary
Slide11Some Difficulties Encountered during the Contour Trace Procedure for Surface and Lumen Geometries
This is the 49
th
image in the patient CT scan data that is provided. Notice that
part of the
bulge of the aneurysm
is quite large that it appears
next to the celiac trunk/common hepatic
artery in addition to its usual location.
As one progresses caudally, the cross-section images show how the “divided” aneurysm begins to “merge”. The images shown from left to right are the 52
nd
, 55
th
, and 61st images in the provided data series, respectively.
Image: 49
Image: 55
Image: 52
Image: 61
Aneurysm Appears Divided in Cross-Section
Part of AAA Bulging
AAA
Slide12Some Difficulties Encountered during the Contour Trace Procedure for Surface and
Lumen Geometries
Continued
. . .
Image: 49
Image: 52
Image: 55
Image: 61
In a situation like this, one can either trace surface
and lumen contours
that gradually grow from image to image until ultimately reaching the desired
boundary
.
If this approach is used, there will be an error in the geometrical
representations since important parts of the aneurysm are not considered in some images. I used this approach to generate the aneurismal surface and lumen geometries shown in the previous slides since the program I used for segmentation did not allow me to trace contours in two separate areas.
If possible, a more preferable option is to trace the surface
and lumen contours for each of the apparently divided parts of the aneurysm to obtain a more accurate representation of the geometry.
These serial images show how to trace contours for the generation of a surface geometry in the situation where a cross-section image displays parts of the same large aneurysm separately. The lumen contours can be traced in a similar fashion.
Slide13After completing the contour tracing process in a program written by one of the post-docs in my laboratory, the output consisted of .
obj
and .
mtl
files that included the x, y, and z coordinates for specific points defined within the contours. Then, I opened the .
obj
file using Mesh Lab, a free downloadable software, to display the geometries shown in this power point presentation.
Please feel free to contact me if you need further information or assistance. I look forward to working with you. Thanks for your collaboration.