CT Perfusion Imaging in Cerebral Ischemia
To the Editor:
We read with interest the Guidelines and Recommendations for Perfusion Imaging in Cerebral Ischemia by the Writing Group on Perfusion Imaging, from the Council on Cardiovascular Radiology of the American Heart Association.1 The authors have succeeded admirably in summarizing the important recent developments of perfusion imaging and their relevance to clinical investigations of cerebral ischemia. We have been involved in the development of the First-Pass Bolus Tracking Methodology of CT Perfusion for over a decade and would like to comment on a few issues raised by the authors concerning this particular methodology.
1. Radiation dose.
The Table is a comparison of the effective dose equivalent2 (HE) of each of the three perfusion imaging methods discussed in the Guidelines that involve the use of ionizing radiation. The effective dose equivalents for the 2 computed tomography (CT) techniques are estimated using the methodology published by Huda et al 3 based on the CT dose index values published for LightSpeed QXi scanners (General Electric Medical Systems). For CT scanners of other models or from other manufacturers, the values in the Table can be scaled proportionally according to the computed tomography dose index value of the scanner relative to the LightSpeed QXi scanner.
The Table shows that CT perfusion (CTP) imaging does not necessarily give a higher radiation dose to the subject than XeCT and SPECT perfusion imaging. In comparison, a screening head CT scan has an effective dose equivalent of 1.5 mSv.3 With the recent interest in using a saline chaser to shorten the duration of the first transit of the contrast bolus through the brain, the scan duration for CTP can be reduced to around 25 to 30 s resulting in an even lower dose of 1.8 to 2.2 mSv.
2. Single versus multiple functional parameters.
While xenon-enhanced computed tomography (XeCT) and single photon emission computed tomography (SPECT) can only measure cerebral blood flow (CBF), CTP can simultaneously measure CBF, cerebral blood volume (CBV) and mean transit time (MTT). Although the results need to be confirmed in a larger clinical trial, Wintermark et al5 have shown that CBV and CBF together can be used to differentiate reversible from non-reversible ischemia. A corollary of this result is that CBV could be viewed as the surrogate marker for time since ischemia. In this sense, CTP imaging could provide a more complete assessment of cerebral ischemia than either XeCT or SPECT in that both the severity and duration of ischemia would be assessed.
3. Arterial input function (AIF).
The authors are correct in their conclusion that measurement of the arterial input function from proximal branches of middle cerebral artery (MCA) or anterior cerebral artery (ACA) is affected by partial volume averaging (PVA). 6 It is also true that this PVA effect is more prominent when a distal MCA, or ACA branches are used for the determination of the AIF when the slice of interest is higher than the usual one through the basal ganglia. However, methods are available for the correction of the PVA effect. We described a method in a prior publication, which requires an independent calibration of the scanner,6 and thus may not be convenient to use in clinical departments. A more convenient method is to reference the area under the AIF to that of the time-density curve of a venous sinus in one of the acquired CT slices. The assumption is that the venous sinus used is large relative to the limiting spatial resolution of CT scanners (0.67 to 0.71 mm) for the PVA effect to be negligible. If that is the case, the ratio of the two areas should give the fraction by which the AIF has been underestimated due to the PVA effect. This very convenient method has already been implemented in one of the commercially available CTP imaging software.
4. Mean transit time.
It is stated that the MTT measured by CTP is not “the transit time through the same volume in which the CBV is determined” instead it is the transit time between MCA and a venous sinus. In CTP imaging, the MTT map is determined by the deconvolution of the AIF and the tissue time-density curve corresponding to each pixel of the map. The deconvolution determines the impulse residue function7 for the tissue volume in each pixel and the transit time through that tissue volume is calculated as the ratio of the area to the height of the impulse residue function as prescribed by Meier and Zieler.8,9 It is important to note that the calculation of MTT in CTP does not involve a deconvolution between the AIF and the venous sinus time-density curve to determine the transit time between MCA and the venous sinus. The venous sinus time-density curve is used just for PVA correction as discussed above in point #3.
5. Blood–brain barrier disruption.
There is concern that CTP imaging may return erroneous CBF and CBV values when the blood–brain barrier becomes permeable following ischemia. We have extended the basic CTP tracer kinetics model to account for blood-tissue permeability to contrast agent.10 Furthermore, 2 of our prior publications have validated the extended CTP imaging method in the measurement of CBF in a brain tumor model11 as well as the measurement of tumor blood flow in a soft tissue tumor model12 against a gold standard—microspheres (class I data).
6. Imaging speed.
The characterization by the authors that echoplanar magnetic resonance (MR) is faster than CT is a simplification of a complicated issue. A proper comparison of speed between CT and MR should include considerations of spatial resolution and signal-to-noise of the resultant images not just on the raw speed of image acquisition. As discussed earlier good spatial resolution is required for PVA correction of the AIF. If we compare the spatial resolution achieved with CTP with those of echoplanar MR, then the comparison may not be as favorable to MR. For instance, an echoplanar T2* weighted spin-echo imaging sequence can acquire a slab of 11 × 6 mm slices every 1.6 s at an in-plane resolution of 4.8 mm (phase encode [y-] direction) and 1.7 mm (read [x-] direction), whereas a CTP sequence can acquire a slab of 4 × 5 mm slices every 0.5 to 1.0 s at an in-plane resolution of 0.67 to 0.71 mm in both x and y directions. CTP can measure the time-density curve from a venous sinus unaffected by the PVA effect. This is critical for the correction of the PVA effect on the AIF. Currently used MR echoplanar imaging sequences, because of its poorer spatial resolution, cannot implement the same correction method for the PVA effect on the AIF as CTP imaging can.
Writing Group on Perfusion Imaging, From the Council on Cardiovasscular Radiology of the Am Heart Association. AHA Scientific Statement. Guidelines and recommendations for perfusion imaging in cerebral ischemia. Stroke. 2003; 34: 1084–1104.
International Commission on Radiological Protection. Recommendations of the International Commission on Radiological Protection. ICRP Publication 26. Oxford: Pergamon Press; 1977.
Huda W, Sandison GA, Lee TY. Patient doses from computed tomography in Manitoba from 1977 to 1987. Br J Radiol. 1989; 62: 138–144.
Wintermark M, Reichhart M, Thiran JP, Maeder P, Chalaron M, Schnyder P, Bogousslavsky J, Meuli R. Prognostic accuracy of cerebral blood flow measurement by perfusion computed tomography, at the time of emergency room admission, in acute stroke patients. Ann Neurol. 2002; 51: 417–432.
Cenic A, Nabavi DG, Craen RA, Gelb AW, Lee TY. Dynamic CT measurement of cerebral blood flow: a validation study. AJNR Am J Neuroradiol. 1999; 20: 63–73.
Bassingthwaigthe JB, Chinard FP, Crone C, Lassen NA, Perl W. Definitions and terminology for indicator dilution methods. In: Crone C, Lassen NA, eds. Capillary Permeability. Copenhagen: Munskgaard; 1970.
Meier P, Zierler KL. On the theory of the indicator-dilution method for measurement of blood flow and volume. J Appl Physiol. 1954; 6: 731–744.
Zierler KL. Equations for measuring blood flow by external monitoring of radioisotopes. Circ Res. 1965; 16: 309–321.
Cenic A Navbavi DG, Craen RA, Gelb AW, Lee TY. CT method to measure hemodynamics in brain tumors: validation and application of cerebral blood flow maps. AJNR Am J Neuroradiol. 2000; 21: 462–470.
Speaking for my colleagues in the Writing Group on Perfusion Imaging from the Council on Cardiovascular Radiology of the American Heart Association, I would like to make a few points in response to the excellent, informative letter of Dr. Lee and his colleagues.
1. In point 2, they state that the combination of computed tomography perfusion (CTP)-derived cerebral blood volume (CBV) and cerebral blood flow (CBF) can differentiate reversible from non-reversible ischemia. All neuroscientists would applaud if that statement were proven to be true. I would suggest the word, “might”. Animal studies with comparative techniques acting as gold standards, and much larger controlled clinical studies will be necessary to ascertain the truth. “Reversibility” can be either spontaneous or through interventions such as the administration, intravenously or intra-arterially, of a thrombolytic agent, among many possibilities. Such treatments have their own sets of variables that add to the complexity of the prediction of tissue viability.
The use of CBV plus CBF is similar to the magnetic resonance (MR) paradigm of using diffusion and perfusion, with CBV paralleling diffusion, and CBF the perfusion parameter. Similar problems regarding “reversibility” may confront CTP as they do with MR perfusion/diffusion imaging. Recent articles suggest that diffusion as measured with MR is not as simple as initially thought. Physiologically, a focus of ischemia is heterogeneous. Within this focus there is a mixture of diffusion values, dependent on a variety of factors, including local collateral flow. This collateral allows CBV to increase initially as an attempt to maintain tissue oxygenation, but then the average CBV value drops as this autoregulatory mechanism fails. While a low CBV value probably indicates that some infarction is present, how much? Is any of this process reversible? Is there a population of cells that can be rescued? Over what time period?
It is said that CBV might be a surrogate marker for “time since ischemia”. Given the physiological discussion above, it is difficult to understand that statement. CBV depends on, to a greater degree, the status of the collateral circulation, not the time since the insult. It is a reflection of severity—the failure of an autoregulatory mechanism.
It is also said that CTP can provide a “more complete assessment” of ischemia than either XeCT or SPECT. By more complete, the authors mean the multiple calculated parameters. That may or may not be helpful in making difficult clinical decisions. As discussed in our article,1 there has been an extensive experience using CBF as determined with XeCT, a single quantified value, to predict not only the potential reversibility of the ischemic process, but also the propensity for edema formation and hemorrhage. SPECT has also demonstrated an ability to predict hemorrhage with thrombolytic treatment. Thus, we should insist on “efficacy” of a technique for clinical utilization, whether that is with one or more parameters, rather than on the number of variables that can be calculated.
2. In point 5, the authors cite their excellent work on calculating correction factors to account for a permeable blood–brain barrier in tumor models. However, a well-defined tumor is not the same as a heterogeneous region of ischemia/infarction.
3. In point 6, mixing speed of image acquisition with other parameters such as spatial resolution makes for a very difficult comparison of perfusion techniques. Contrast resolution, with MR leading the other methodologies reviewed, could be added to the equation, too. The message we attempted to convey is that MR permits imaging of the whole brain, whereas current quantitative CT methodologies are limited to a relatively small field of view within the defined region of interest.
T.-Y.L. is the developer of the CT Perfusion software and Robarts Research Institute is the licensor of the software to General Electric Medical Systems.