Transcript Coronary Phantom Design for K-edge Angiography John

Coronary Phantom Design for K-edge Angiography
John Jorgensen1, Sarah Pachtman1, Punam Patel1, Marcus Spallek1
Advisors: Paul King, Ph.D.1; Frank E. Carroll, M.D. 2
1
Department of Biomedical Engineering, Vanderbilt University, Nashville, TN
2Vanderbilt University Medical Center Department of Radiology and Radiological Sciences
Coronary Phantom Design
Overview of MX200
Figure 1. Diagram of MX200’s laser paths and
major elements.
MXISystems, Inc. in conjunction with Vanderbilt
University and the Office of Naval Research has
developed a laser synchrotron (Inverse Compton
scattering) X-ray source for the energy region from
10 keV to 50 keV. It is capable of producing X-ray
fluences sufficiently high that a single, 9-ps
pulse produces a complete medical X-ray image. The
device is a tunable, monochromatic x-ray (MX200) to
be used in medical imaging. Current research focuses
on the application for intravenous coronary
angiography.
Imaging
• The phantom was not constructed with the material that was specified. Due to the
specifications of the printing machine used to manufacture the prototype, Z-Corp’s ZP-102
plaster-based powder was used to create the prototype. Images were shot using a semimonochromatic method:
Device Description
The coronary phantom consists of a hollow chamber, with a 5 mm diameter tube for saline pump
attachment. The outer surface is comprised of interconnected tubes. The three vertical tubes have
internal diameter of 1 mm. The horizontal tubes range in diameter from 2 to 5 mm. Two openings
are directly connected to the ring grid and allow for fluid filing and emptying. The tubes do not open
into the chamber. The device is one solid unit designed to be filled with fluids for imaging with the
MX200.
http://www.mxisystems.com/mx200.html
•Images were taken with two polychromatic tubes running at 50KVP, having different effective
X-ray energies:
• Molly tube (Molybdenum) with an aluminum filter.
•gives an X-ray spectrum with a peak around 18KeV provided our “below the K-edge
of Iodine” shots.
•Tungsten tube with an aluminum filter running at 50KVP.
•has a peak X-ray output in the range of 50 KeV (values rounded off in data analysis as
50KeV since it is still simply polychromatic)
• Results were imaged on MAR345 digital detector running in 100µ resolution mode. They
were then downloaded to a PC and processed using ImageJ.
Subtraction Method
Social Impact
• Images were taken at 18 and 50KeV
• A portion of incident X-rays was selected and the exposure was summed in the same area of
both images.
• The images were normalized to each other so that the area of selection of each image
contained the same exposure by multiplying the lesser exposed by the ratio of the two
selections.
• Using ImageJ Image Calculator the “difference” between the high energy and the low energy
was created, making the areas that might have increased absorption due to k-edge effects more
pronounced.
• Hypersensitivity reaction to iodine (3-12% patients)
• Increase patient safety:
• reduce the amount of potent contrast agents used during angiography
• decrease the dose of ionizing radiation to the patient by employing K-edge Imaging
techniques
Project Deliverables
Phantom Specifications
• Design a coronary phantom
3”
• Test contrast agent concentrations
30”
Results
θ
•Current clinical dilution with saline solution is 1:16
•Using monochromatic plain films it is possible to see 1:256 dilutions
Tan θ = 3/30 = 0.10
θ ~ 6°
• Determine optimal angle for stereo imaging
Design Considerations
Figure 2. Initial images taken with an empty phantom (left) and
with saline (right).
• Radio translucent at 35keV
Figure 3. Images taken with 1:256 dilution with the subtraction
method (left) and differential method (right).
•K-edge of Iodine is 33.2KeV
• Flexible material for a pulsatile model
Conclusion
• Fluid perfusion capabilities
• Phantom Size
• 7.5cm diameter Approximate size of canine heart
Material Testing
•Vasculature
• Size Range 5 - 1 mm
• Coronary arteries represented on surface
• Curved channels
• Hollow vessels
Finger
imaged at
Finger
imaged at
19 keV
26 keV
Market Analysis
•Product is a reusable coronary phantom specific for imaging research for the MX200.
•The determination of the optimal angle for stereoscopic imaging as a function of distance will
benefit the radiological community.
•Optimal cost to performance ratio with an associated manufacturing cost of $20 for materials
testing, while the current industry manufacturing cost is ~$5000.
•Several samples of 2component polyurethane resins
tested for radio opaque
appearance at 35 keV.
•Smooth-On Reoflex™ 20
liquid rubber chosen as
ideal material because of
translucent appearance on
detector.
•High grayscale value
indicates degree of x-ray
absorbance and
translucency.
Smooth-On Reoflex™
20
Shore A
20 A
Durometer
Tear
60 pli
Strength
(pli)
Tensile
200 psi
Strength
(psi)
Elongation 1000%
at Break
3-D images of the empty phantom indicated that one of the walls separating the arteries from the main
chamber was damaged. It was noted that when air was blown into the arterial portion of the phantom
the air leaked out of all ports, indicating the presence of a hole in the wall between the inside chamber
and the surface arteries. The phantom was filled with a small amount of saline and a leak was
detected.
The images were taken with an iodine dilution of 1:256. A balloon filled with water was inserted in
the center chamber while the contrast solution was filled in the arteries (image on left). Note the
balloon and water appear very opaque and would require high dose exposures to image through the
phantom. The 1:256 dilution we were able to image is still significantly lower than the dose of about
1:16 given to a CT patient. The image on the right was edited using a differential method. This
method relies on logic function and provided a clear image of the contrast agent over the subtraction
method.
The 3-D movie shows the phantom with rotations of 3, 6, 9, 12 degrees. The 6° is found to be the
optimal stereo angle for imaging.
Acknowledgements
This project was made possible through the help of the following contributors:
Scott Degenhardt; Robert Traeger; Dr. Frank Carroll; Philip Davis; Z-Corporation