Abstract:
HTML
Purpose: To improve the sensitivity and specificity of spiral CT for
pulmonary embolism using subtraction images to identify perfusion defects.
Methods and Materials: Three to 8 methacrylate beads measuring 4-5 mm diameter
were embolized into the pulmonary vasculature of 8 anesthetized, ventilated
juvenile pigs to simulate occlusive segmental pulmonary emboli. Two animals did
not receive emboli for a total of 10 animal studies. CT scans were obtained on
an 8 row multidetector CT scanner (GE Lightspeed Ultra) using 1.25-mm
collimation and reconstruction. Intravenous contrast medium was administered
via an 18 gauge cathether in the brachial vein at two rates (1.5 and 3 ml/s). A
non contrast CT scan was acquired in a caudal-cranial direction. Following this
scan the contrast injection was started and a second contrast enhanced CT (CTA)
was acquired after a 20 second prep delay. Both scans were obtained within a
single 45-second breath-hold. Following imaging the animals were euthanized and
methacrylate casts of the pulmonary arteries made which formed the gold
standard for emboli location. Perfusion images were created by subtracting the
pre-contrast scan from the CTA scan on a slice by slice basis. The studies were
randomized and interpreted by 3 blinded observers on a medical workstation in
three groups: CTA images alone and on a second reading in combination with
perfusion images. The location of the emboli and the diagnostic confidence (5
point scale, 5 confident, 1 unsure) was recorded.
Results: 46 emboli were injected creating 37 independent embolic sites and 35
separate perfusion defects. Compared to CTA alone, at 1.5 ml/s the addition of
perfusion images increased the mean sensitivity and specificity by 6 and 3%
respectively. At 3 ml/s the increase was 5 and 3% respectively. Using the 3
ml/s injection rate and perfusion images, two observers obtained 95%
sensitivity at 96% specificity. Compared to the interpretation of the CTA
alone, the combined interpretation of CTA and perfusion scan increased
confidence from 4.5 to 4.77 at 1.5 ml per sec and 4.77 to 4.93 at 3 ml per
second.
Conclusion: The addition of subtraction perfusion images to CTA for pulmonary
emboli increased sensitivity, specificity and confidence over CTA alone. (J.R.M.
received a grant from GE Medical Systems.)
Questions about this event email: jmayo@vanhosp.bc.ca
Mayo MD, J,
Enhanced Detection of Pulmonary Emboli through the Use of Subtraction CT Perfusion Images. Radiological Society of North America 2003 Scientific Assembly and Annual Meeting, November 30 - December 5, 2003 ,Chicago IL.
http://archive.rsna.org/2003/3107331.html
