Technical considerations of CTP

Comparison with MR-PWI

Advantages

Quantitation and resolution

While CTP and MR-PWI both attempt to evaluate capillary-level hemodynamics, their differences should be considered. Dynamic susceptibility contrast (DSC) MR-PWI techniques rely on the T2* effect induced in adjacent tissues by high concentrations of intravenous gadolinium; CTP relies on direct visualization of intravascular contrast material. The linear relationship between contrast concentration and CT attenuation more readily lends itself to quantitation, which is not easily achieved with MR-PWI. In addition, CTP has greater spatial (but not contrast) resolution compared with MR-PWI. These factors contribute to the possibility that visual evaluation of core/penumbra mismatch may be more quantitatively reliable with CTP than with MR perfusion (MRP) ( Fig. 2 ).

Correlation between CT and MR perfusion maps, performed approximately 40 minutes apart. Perfusion images in an 88-year-old man who was found unresponsive. CTP images (top row, from left: cerebral blood flow [CBF], cerebral blood volume [CBV], mean transit time [MTT]) and MRP images (bottom row, from left: CBF, CBV, MTT) demonstrate a large perfusion deficit in the right MCA distribution.

Availability and safety

CT also benefits from the practical availability and relative ease of scanning, particularly when dealing with critically ill patients and the attendant monitors or ventilators. CT may also be the only option for a subgroup of patients with an absolute contraindication to MR scanning, such as a pacemaker, and is a safe option when the patient cannot be screened for MR safety. For patients with borderline glomerular filtration rate (GFR 45–60) in whom perfusion data is important to patient management, the risk-to-benefit ratio of administering iodinated contrast is likely less than that of gadolinium.

Comparison with MR-PWI

Advantages

Quantitation and resolution

While CTP and MR-PWI both attempt to evaluate capillary-level hemodynamics, their differences should be considered. Dynamic susceptibility contrast (DSC) MR-PWI techniques rely on the T2* effect induced in adjacent tissues by high concentrations of intravenous gadolinium; CTP relies on direct visualization of intravascular contrast material. The linear relationship between contrast concentration and CT attenuation more readily lends itself to quantitation, which is not easily achieved with MR-PWI. In addition, CTP has greater spatial (but not contrast) resolution compared with MR-PWI. These factors contribute to the possibility that visual evaluation of core/penumbra mismatch may be more quantitatively reliable with CTP than with MR perfusion (MRP) ( Fig. 2 ).

Correlation between CT and MR perfusion maps, performed approximately 40 minutes apart. Perfusion images in an 88-year-old man who was found unresponsive. CTP images (top row, from left: cerebral blood flow [CBF], cerebral blood volume [CBV], mean transit time [MTT]) and MRP images (bottom row, from left: CBF, CBV, MTT) demonstrate a large perfusion deficit in the right MCA distribution.

Availability and safety

CT also benefits from the practical availability and relative ease of scanning, particularly when dealing with critically ill patients and the attendant monitors or ventilators. CT may also be the only option for a subgroup of patients with an absolute contraindication to MR scanning, such as a pacemaker, and is a safe option when the patient cannot be screened for MR safety. For patients with borderline glomerular filtration rate (GFR 45–60) in whom perfusion data is important to patient management, the risk-to-benefit ratio of administering iodinated contrast is likely less than that of gadolinium.

Radiation dose in CTP: facts, publicity, and response

On October 8, 2009, the US FDA issued an initial notification regarding a safety investigation of facilities performing CTP scans. Because of incorrect settings on the CT scanner console of a single facility, more than 200 patients over a period of 18 months received radiation doses that were approximately 8 times the expected level. Approximately 40% of the patients lost patches of hair as a result of the overdoses. These incidents followed another unrelated incident in a community hospital in Arcata, California, in 2008, where a CT technologist activated a CT scan 151 times over the same region of the head of a 2-year-old.

These incidences highlight the importance of CT quality assurance programs. Such programs should include regular reviews of CT protocols by a specialized CT physicist, testing protocols on dose phantoms, and monitoring of actual doses received by patients for each type of CT protocol. Radiologists and technologists should be familiar with the dose indices displayed on the CT scanner console and should include the volumetric CT dose index (CTDI _vol ) and the dose-length product (DLP). The CTDI _vol , represents the average dose delivered within the reconstructed section. The DLP is the CTDI _vol multiplied by the scan length expressed in centimeters. DLP can be used to assess the overall radiation burden in a CT study. In response to the Arcata incident, the state of California passed a landmark dose reporting law in November 2010; it mandates the report of CTDI/DLP doses in all CT studies, and that overdoses be reported to the patient, the referring MD, and the state of California. FDA initiatives during 2010 and endorsed by MITA (the Medical Industry and Technology Alliance) will likely result in dose safeguards built into CT scanner designs, dose recording for all CT studies, and tools to track the imaging history of each patient. Moreover, the American College of Radiology Dose Index Registry will monitor the cumulative dose information for each patient.

All CT protocols should respect the ALARA (As Low As Reasonably Achievable) principle. Expert consensus and over a decade of experience suggest that CTP studies should be performed at 80 kVp and no more than 200 mAs. When using such parameters, the effective radiation dose associated with a single slab CTP study is approximately equal to that of an unenhanced head CT, roughly 2 to 3 mSv. For comparison, the background radiation for a person living in Boston for a year is approximately 3 mSv. The authors’ current shuttle mode CTP scanning protocol at Massachusetts General Hospital results in a radiation exposure, CTDI _vol , of 0.35 Gy, a value that is substantially below the FDA recommended maximum recommended radiation dose of 0.5 Gy. Note that this kVp setting is lower than that used for standard CT studies (120–140 kVp), and was selected primarily because the conspicuity of the iodine contrast agent is much greater at 80 kVp than at 140 kVp, resulting in improved CTP map image quality at a markedly lower radiation dose.

CTP methodology: general principles

Perfusion-weighted CT and MR, in distinction to those of MR and CTA which detect bulk vessel flow, are sensitive to capillary, tissue-level blood flow. The generic term “cerebral perfusion” refers to tissue level blood flow in the brain. This flow can be described using a variety of parameters, which primarily include cerebral blood flow (CBF), cerebral blood volume (CBV), and mean transit time (MTT). Definitions of these parameters are as follows.

CBV is defined as the total volume of blood in a given unit volume of the brain. This definition includes blood in the tissues, as well as blood in the large capacitance vessels such as arteries, arterioles, capillaries, venules, and veins. CBV has units of milliliters of blood per 100 g of brain tissue (ml/100 g).

CBF is defined as the volume of blood moving through a given unit volume of brain per unit time. CBF has units of ml of blood per 100 g of brain tissue per minute (ml/100 g/min).

MTT is defined as the average of the transit time of blood through a given brain region. The transit time of blood through the brain parenchyma varies depending on the distance traveled between arterial inflow and venous outflow, and is measured in seconds. Mathematically, mean transit time is related to both CBV and CBF according to the central volume principle, which states that MTT = CBV/CBF.

“Core” is typically operationally defined as brain likely to be irreversibly infarcted at presentation, despite early recanalization (and which indeed may be at increased risk for hemorrhagic transformation with early robust reperfusion), and “penumbra” as the functionally ischemic but potentially salvageable “at-risk” brain parenchyma ; the latter is to be distinguished from “benign oligemia” reflecting hypoperfused but functionally viable tissue. “Mismatch” is defined as the difference in volume and location between core and penumbra, although as currently measured in routine practice, “penumbra” often includes regions of benign oligemia as defined by other transit time measures such as TTP or Tmax (time to peak, or maximal, enhancement). A mismatch of greater than 20% is typically, arbitrarily, considered to represent a clinically significant penumbra for both research and clinical management; major trials using this definition include the Diffusion-weighted imaging Evaluation For Understanding Stroke Evolution (DEFUSE) trial and the Echoplanar Imaging Thrombolytic Evaluation Trial (EPITHET).

CTP theory and modeling

CTP postprocessing

General Principles

In urgent clinical cases, perfusion changes can often be observed immediately following scanning by direct visual inspection of the axial source images at the CT scanner console. Soft copy review at a workstation using “movie” or “cine” mode can reveal relative perfusion changes over time. It has been suggested that arterial-phase CTP source images (CTP-SI) closely correlate with CTP-CBF and venous-phase CTP-SI with CBV; hence a visual inspection of the CTP-SI can determine penumbra, core, and mismatch. However, advanced postprocessing is required to appreciate subtle changes and to obtain quantification.

Axial source images acquired from a cine CT perfusion study are networked to a free-standing workstation for detailed analysis, including construction of CBF, CBV, and MTT maps. Postprocessing can be manual, semi-automated, or fully automated.

The computation of quantitative first-pass cine cerebral perfusion maps typically requires some combination of the following user inputs ( Fig. 3 ).

• •
  

  Arterial Input Function ( AIF ): A small ROI is placed over the central portion of a large intracranial artery, preferably an artery orthogonal to the imaging plane to minimize “dilutional” effects from volume averaging. An attempt should be made to select an arterial ROI with maximal peak contrast intensity.

  
  
• •
  

  Venous Outflow Function ( VOF ): A small venous ROI with similar attributes is selected, most commonly at the superior sagittal sinus. The purpose of the VOF is to serve as a reference for normalization of the quantitative CTP parameter values, relative to an averaged maximum intravascular contrast measurement. Because veins (particularly the posterior superior sagittal sinus) are typically larger than arteries, the venous TDC is not as subject to partial volume average as is the arterial TDC.

  
  
• •
  

  Baseline : The baseline is the “flat” portion of the arterial time density curve, prior to the upward sloping of the curve caused by contrast enhancement. The baseline typically begins to increase after 4 to 6 seconds.

  
  
• •
  

  Post-Enhancement Cutoff : This refers to the “tail” portion of the TDC, which may slope upwards toward a second peak value if recirculation effects are present. When such upward sloping at the “tail” of the TDC is noted, the data should be truncated to avoid including the recirculation of contrast. The perfusion analysis program will subsequently ignore data from slices beyond the cutoff.

CTP post processing. ( Left ) Appropriate ROI placement on an artery (anterior cerebral artery, running perpendicular to the plane of section to avoid volume averaging) and on a vein (the superior sagittal sinus, also running perpendicular to the plane of section and placed to avoid the inner table of the skull). ( Right ) the time-density curves (TDC) generated from this artery (A) and vein (B) show the arrival, peak, and passage of the contrast bolus over time. These TDCs serve as the arterial input function and the venous output for the subsequent deconvolution calculation.

It is worth noting that major variations in the input values described may not only result in perfusion maps of differing image quality, but potentially in perfusion maps with variation in their quantitative values for CBF, CBV, and MTT. As previously noted, special care must be taken in choosing an optimal VOF, because that VOF value may be used to normalize the quantitative parameters.

Although the precise choice of CTP scanning level is dependent on both the clinical question being asked and on other available imaging findings, an essential caveat in selecting a CTP slice is that the imaged level must contain a major intracranial artery. This caveat is necessary to assure the availability of an AIF, to be used for the computation of perfusion maps using the deconvolution software.

Correction for delay and dispersion of the AIF: why it matters

Calculation of CBF requires knowledge of the AIF, which in practice is estimated from a major artery, assuming that it represents the exact and only input to the tissue voxel of interest, with neither “delay” nor “dispersion.” There are several clinical situations, however, where the AIF TDC will lag, and the tissue TDC will lag behind the AIF curve (“delay”). AIF delay can occur due to extracranial causes (atrial fibrillation, severe carotid stenosis, poor left ventricular ejection fraction) or intracranial causes (proximal intracranial obstructive thrombus with poor collaterals). Moreover, in such cases the contrast bolus forming the AIF can spread out over multiple pathways proximal to the tissue ROI (“dispersion”).

Delay and dispersion do not have the same hemodynamic significance or clinical implications in acute versus chronic vascular conditions. In chronic vascular conditions such as an extracranial stenosis, delay and dispersion can result in grossly underestimated CBF and overestimated MTT, most notably in regions of nonischemic hypoperfusion (benign oligemia). These changes indicate the hemodynamic significance of the stenosis, but can be misleading clinically in that inexperienced readers may interpret them as the hallmark of an acute ischemic stroke (especially when the distinction between true ischemia and benign oligemia is unclear). As a result, they need to be corrected for. In acute ischemic stroke, however, delay and dispersion are integral to the underlying pathophysiological hemodynamics, and therefore need not necessarily be corrected for so that appropriate clinical interpretation is possible. In practice, many acute stroke patients have associated chronic carotid stenosis, and hence the “truth” lies somewhere in between. Although it can be argued that delay and dispersion should be corrected for so that MTT and CBF maps more accurately reflect physiologic ground truth, such correction can make MTT and CBF maps look too “normal.” In any case, commercial software packages with delay-sensitive deconvolution algorithms can overestimate penumbra (ie, MTT or CBF lesion volume) and consequently final infarct volume, whereas penumbra estimated with delay-insensitive software generally correlates better with final infarct volume. Delay-sensitive methods might also overestimate CBV lesion size in some patients with concomitant intra- or extracranial severe hemodynamic delay ( Fig. 4 ).

Example of a false-positive CBV map “core” lesion, which was not present in maps that were postprocessed using newer generation, delay-insensitive software. Patient is a 61-year-old with atrial fibrillation, congestive heart failure, and an ICA thrombus that recanalized following intra-arterial thrombolysis. ( A ) Admission CT-CBV map processed using “first-generation” delay-sensitive software shows a large right MCA territory low blood volume lesion (white outline). ( B ) CT-CBV map appears normal when postprocessed using newer generation, delay-insensitive software. ( C ) Follow-up CT appears normal. ( D ) ROIs for the TDCs shown in E and F ( arrow ). ( E ) Normal arterial input and venous outflow curves. ( F ) Delay in tissue TDCs on the ischemic compared with the contralateral side, due to proximal vascular obstruction and flow through collateral pathways in the setting of atrial fibrillation and low cardiac output.

( Reproduced from Schaefer PW, Mui K, Kamalian S, et al. Avoiding “pseudo-reversibility” of CT-CBV infarct core lesions in acute stroke patients after thrombolytic therapy. The need for algorithmically “delay-corrected” CT perfusion map postprocessing software. Stroke 2009;40(8):2875–8; with permission.)

Several approaches have been used to minimize the effects of delay and dispersion in deconvolution methods. A detailed description is beyond the scope of this article. (For a review of the methods used for correcting for delay, see Konstas and colleagues. )

Clinical applications of CTP

Current indications for performing CTP imaging of acute stroke include: (1) when DWI is unavailable, (2) when the (a) diagnosis, (b) type of stroke, or (c) extent of ischemia is uncertain based on the clinical examination and other imaging, or (3) for postaneurysmal subarachnoid hemorrhage (SAH) vasospasm risk assessment (given cumulative radiation dose issues, performing more than 2 or 3 serial CTPs per admission is ill advised). Although to date there is no high level of evidence to suggest that perfusion scanning has a direct role in the decision to administer IV or intra-arterial (IA) thrombolytic therapy, potential future indications for perfusion imaging in the first 9 to 12 hours after stroke onset include: (1) exclusion of patients most likely to hemorrhage and inclusion of patients most likely to benefit from thrombolysis, (2) extension of the time window beyond 4.5 hours for IV and 6 hours for anterior circulation IA thrombolysis, (3) triage to other available therapies, such as hypertension or hyperoxia administration, (4) disposition decisions regarding neurologic intensive care unit (NICU) admission or emergency department discharge, and (5) rational management of “wake-up” strokes, for which precise time of onset is unknown.

The results of the 2008 ECASS expanded the 3-hour time window for intravenous thrombolysis and revealed that although safe and effective up to 4.5 hours after stroke onset, treatment benefits roughly half as many patients as those treated within 3 hours. Hence, the ratio between the hemorrhagic risks of treatment versus the potential clinical benefits becomes a more critical consideration as the time window for therapy is expanded with newer intravenous and IA techniques. It is the mismatch between the size of the infarct core (proportional to hemorrhagic risk) and the size of the ischemic penumbra (proportional to potentially salvageable tissue), as determined by CTP, that provides an imaging measure of this risk-to-benefit ratio. Evidence suggests that core/penumbra mismatch may persist for up to 24 hours in some patients.

Despite this evidence, some large clinical trials that used mismatch as patient selection criteria, such as the Desmoteplase In Acute ischemic Stroke (DIAS)-2 trial, which used both CTP and MRP up to 9 hours after stroke onset, have failed to show a benefit of treatment. The failure of DIAS-2 to demonstrate a role of mismatch imaging for patient selection in IV thrombolysis may have been, in part, related to technical differences between CTP and MRP, and the lack of standardization and validation of these methods (see later discussion).

Image interpretation: infarct detection with CTA-SI

Several groups have suggested that CTA source images, like DWI, may sensitively delineate tissue destined to infarct despite successful recanalization. The superior accuracy of CTA-SI in identifying infarct “core” compared with unenhanced CT has been unequivocally established in multiple studies. Theoretical modeling indicates that early generation CTA-SI—assuming an approximately steady state of contrast in the brain arteries and parenchyma during image acquisition — are predominantly blood volume weighted rather than blood flow weighted. However, newer, faster MDCT CTA-SI protocols, such as those used in 64-slice scanners, with injection rates up to 7 mL/s and short preparation delay times of 15 to 20 seconds, change the temporal shape of the “time-density” infusion curve, eliminating a near-steady state during the timing of the CTA-SI acquisition. Hence, with the current generation of CTA protocols on faster state-of-the art MDCT scanners, the CTA-SI maps have become significantly more flow weighted.

With early-generation CTA protocols, in the absence of early recanalization, CTA-SI typically defines minimal final infarct size and, hence, like DWI, can be used to identify “infarct core” in the acute setting ( Fig. 5 ). Coregistration and subtraction of the conventional, unenhanced CT brain images from the axial, postcontrast CTA-SI should result in quantitative blood volume maps of the entire brain. CTA-SI subtraction maps, obtained by coregistration and subtraction of the unenhanced head CT from the CTA-SI, are particularly appealing for clinical use because, unlike quantitative first-pass CT perfusion maps, they provide whole brain coverage. A pilot study from the authors’ group, on 22 consecutive patients with MCA stem occlusion who underwent IA thrombolysis following imaging, demonstrated that CTA-SI and CTA-SI subtraction maps improve infarct conspicuity over that of unenhanced CT in patients with hyperacute stroke. Concurrent review of unenhanced CT, CTA-SI, and CTA-SI subtraction images may be indicated for optimal CT assessment of hyperacute MCA stroke.

An acute right MCA distribution basal ganglia infarction is more conspicuous on the CTA source images (CTA-SI, top right, arrow ) than in the unenhanced CT ( top left ). Subsequent DWI ( bottom left ) and unenhanced CT ( bottom right ) confirm the presence of infarction seen on the CTA-SI. In the setting of a proximal large vessel occlusion without early robust recanalization, low blood pool and poor collateralization on CTA-SI are predictive of infarction.

( From Schaefer PW, Mui K, Kamalian S, et al. Avoiding “pseudo-reversibility” of CT-CBV Infarct core lesions in acute stroke patients after thrombolytic therapy. The need for algorithmically “delay-corrected” CT perfusion map postprocessing software. Stroke 2009;40(8):2875–8; with permission.)

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related posts:

Imaging Acute Stroke is Rapidly Evolving MR Perfusion Imaging in Acute Ischemic Stroke Intra-arterial Stroke Therapy: Recanalization Strategies, Patient Selection and Imaging Measuring Permeability in Acute Ischemic Stroke Noninvasive Carotid Artery Imaging with a Focus on the Vulnerable Plaque ASPECTS and Other Neuroimaging Scores in the Triage and Prediction of Outcome in Acute Stroke Patients

Stay updated, free articles. Join our Telegram channel

Join