Published online before print February 9, 2012, doi: 10.3174/ajnr.A3039
AJNR 2012 33: E61

B. Pulli^a and A.J. Yoo^a
aDivision of Diagnostic and Interventional Neuroradiology
Massachusetts General Hospital
Harvard Medical School
Boston, Massachusetts

We have read with great interest the letter by Dr Kloska¹ and a subsequent reply²concerning a study by Sharma et al,³ investigating the perfusion imaging correlate of CTA source image (CTA-SI) lesions and would like to share our own data on this subject. On the basis of our findings, we agree with Dr Kloska that evaluation of acute ischemia with CTA-SI on state-of-the-art multisection CT scanners depends primarily on the imaging protocol.

In a recent investigation at our institution,⁴ we analyzed 100 patients with acute ischemic stroke who underwent CTA-SI by using 2 different acquisition protocols with close follow-up DWI. We correlated CTA-SI hypoattenuation volumes with concurrent DWI hyperintense volumes and found that with the older protocol designed for a 4- or 16-section CT scanner (Table), CTA-SI volume correlated well with DWI volume. When a 64-section scanner was installed at our medical center, we modified our CTA acquisition protocol to improve arterial phase opacification (Table), and this was associated with significant overestimation of concurrent DWI volumes. Analysis of the CTA protocol parameters revealed that a shorter time from contrast injection to imaging of the ischemic territory was the single best predictor of CTA-SI overestimation of the infarct core on DWI. Imaging too soon likely prevented the contrast from traversing the collaterals and reaching the infarct bed. Most interesting, atrial fibrillation was associated with CTA-SI overestimation with the older protocol, which used a fixed delay, presumably related to delayed contrast-arrival time from reduced cardiac output.

Given these data, we conclude that CTA-SI, just like CT perfusion, requires protocol validation before being used to assess acute infarct size in clinical practice. Moreover, protocol parameters should be standardized. The Table illustrates the variable parameters that have been used in studies evaluating CTA-SI for acute ischemic stroke. Until these issues are fully addressed, CTA-SI should be used cautiously to inform treatment decisions in patients with acute ischemic stroke.

References

1. Kloska SP. CT angiographic source images with modern multisection CT scanners: appropriate injection protocol is crucial. AJNR Am J Neuroradiol 2011;32: E93 » FREE Full Text
2. Aviv RI, Sharma M, Lee TJ. Reply. AJNR Am J Neuroradiol 2011;32: E94 » FREE Full Text
3. Sharma M, Fox AJ, Symons S, et al. CT angiographic source images: flow- or volume-weighted? AJNR Am J Neuroradiol 2011;32: 359–64 » Abstract/FREE Full Text
4. Pulli B, Schaefer PW, Hakimelahi R, et al. Acute ischemic stroke: infarct core estimation on CT angiography source images depends on CT angiography protocol. Radiology 2012;262: 593–604 » Abstract/FREE Full Text
5. Schramm P, Schellinger PD, Fiebach JB, et al. Comparison of CT and CT angiography source images with diffusion-weighted imaging in patients with acute stroke within 6 hours after onset. Stroke 2002;33: 2426–32 » Abstract/FREE Full Text
6. Schramm P, Schellinger PD, Klotz E, et al. Comparison of perfusion computed tomography and computed tomography angiography source images with perfusion-weighted imaging and diffusion-weighted imaging in patients with acute stroke of less than 6 hours’ duration. Stroke 2004;35: 1652–58 » Abstract/FREE Full Text
7. Wittkamp G, Buerke B, Dziewas R, et al. Whole brain perfused blood volume CT: visualization of infarcted tissue compared to quantitative perfusion CT.Acad Radiol 2010;17: 427–32 » CrossRef » Medline

Reply

Published online before print February 9, 2012, doi: 10.3174/ajnr.A3051
AJNR 2012 33: E62

R. Aviv^a
aAssociate Professor Medical Imaging
University of Toronto and Sunnybrook Health Science Centre
Toronto, Ontario, Canada

T.Y. Lee^b
bRobarts Research Institute and Lawson Health Research Center
University of Western Ontario
London, Ontario, Canada

M. Sharma^c
cUniversity Hospital London
London, Ontario, Canada

We thank Drs Pulli and Yoo for their interest in our results and their subsequent confirmation of our study findings.¹ Our work highlighted the danger of extrapolating findings from previous-generation scanners to newer, faster scanners. In particular, before our publication, there was an established opinion that CTA-signal intensities (SIs) were CBV-weighted and could, in addition to providing vascular data, provide an estimate of infarct core. This observation was true for older-generation scanners and was likely due to the increased scanning time required to cover the region of interest. Slower scanning times allowed contrast to reach the blood vessels in the ischemic/infarct region associated with prolonged transit time.

With the advent of modern scanners, the observation is no longer true. The role of CTA, in our opinion, remains the delineation of vascular anatomy, including the site of occlusion and presence of collaterals. CTA protocols should be optimized for vascular, not parenchymal, imaging. Estimation of infarct volume on CTA-SI was useful when CTP was not widely available but should be relegated to the history books. CTP studies provide outstanding estimates of core and tissue at risk with the area under the curve associated with thresholded parameters approaching 0.90 for core determination.² For those centers where CTP is unavailable, we have shown that a postcontrast CT (PCT) following CTA demonstrates hypoattenuation that closely estimates core. PCT can be obtained at a lower dose than CTP. We strongly favor CTP over a PCT approach and have previously shown that infarct can be more confidently and correctly diagnosed by using CTP than noncontrast CT or CTA source images.³

References

1. Sharma M, Fox AJ, Symons S, et al. CT angiographic source images: flow- or volume-weighted? AJNR Am J Neuroradiol 2011;32: 359–64 » Abstract/FREE Full Text
2. Bivard A, Spratt N, Levi C, et al. Perfusion computer tomography: imaging and clinical validation in acute ischaemic stroke. Brain 2011;134: 3408–16 » Abstract/FREE Full Text
3. Hopyan J, Ciarallo A, Dowlatshahi D, et al. Certainty of stroke diagnosis: incremental benefit with CT perfusion over noncontrast CT and CT angiography. Radiology 2010;255: 142–53 » Abstract/FREE Full Text