Abstract
Objectives: To correlate the two types of early ischemic change on noncontrast CT (NCCT) (parenchymal hypoattenuation [PH] and isolated focal swelling [IFS]) with concurrent assessment of cerebral perfusion and to compare their rates of progression to infarction.
Methods: We assessed cortical regions on NCCT for early ischemic change. Quantitative perfusion values were calculated for cortical regions from acute CT perfusion (CTP) maps of cerebral blood volume (CBV), blood flow (CBF), and mean transit time (MTT). Reperfusion and presence of infarction were determined from follow-up MRI.
Results: We studied 40 patients with sub-6 hour anterior circulation ischemic stroke; 19 received IV recombinant tissue plasminogen activator. Of the 202 regions acutely hypoperfused on CTP, 123 were normal on NCCT, 58 had PH, and 21 had IFS. Acute CBV was low in PH regions, and elevated in IFS regions. Acute CBF was reduced in IFS regions, but more so in PH regions. Progression to infarction occurred in virtually all PH regions, but IFS regions had much lower rates of infarction with major reperfusion. Acute CBV in hypoperfused normal NCCT regions ranged from reduced to elevated, with substantially differing risk of infarction.
Conclusions: Isolated focal swelling identifies penumbral tissue and parenchymal hypoattenuation identifies infarct core. Although this has prognostic implications when assessing patient suitability for thrombolytic therapy, the majority of acutely hypoperfused regions appear normal on noncontrast CT. Perfusion CT can stratify the level of risk of subsequent infarction for normal-appearing regions on noncontrast CT.
Noncontrast CT (NCCT) is widely used for acute stroke.1–3 The Alberta Stroke Program Early CT Score (ASPECTS) has improved the reliability of detection of subtle early ischemic changes in the middle cerebral artery (MCA) territory.4,5 The extent of early ischemic change (EIC) on NCCT correlates with poorer clinical outcome, but whether extent of EIC should be used as a selection criterion for thrombolysis is controversial.2,3,6 The most common type of EIC seen on NCCT is parenchymal hypoattenuation (PH).2 This probably indicates severe hypoperfusion of brain tissue and irreversible ischemic injury.3,7 Focal swelling of brain tissue without PH is less common, and its prognostic importance is uncertain.8,9 It has recently been suggested that such isolated focal swelling (IFS) may represent ischemic penumbral or oligemic tissue.9
Multislice CT scanners can assess cerebral perfusion and vessel status, adding only a few minutes to the CT examination.10,11 Perfusion CT imaging (CTP) has the advantage of assessing both reversible and irreversible ischemia by generating parametric maps of cerebral blood volume (CBV), cerebral blood flow (CBF), and contrast mean transit time (MTT).12 With advances in postprocessing software, absolute CBV and CBF values can be generated.13 Thus, the aims of this study were to 1) correlate regional early ischemic change with regional quantitative CT perfusion values, and 2) compare the rates of progression to infarction for these regions, based on presence and type of EIC and quantitative CTP values.
Methods.
Patients.
We prospectively studied consecutive patients with anterior circulation infarction presenting within 6 hours of symptom onset. Only patients who fulfilled clinical criteria for thrombolysis were enrolled.14 Exclusion criteria for CTP/CT angiography (CTA) were contrast medium allergy or renal failure. Patients underwent NCCT/CTP/CTA at baseline and follow-up (day 3) MRI unless there were contraindications to MRI, in which case NCCT/CTA was performed. NIH Stroke Scale (NIHSS) was performed immediately prior to acute CT, day 3 imaging, and at day 90. The study was approved by our institutional Ethics Committee.
Imaging.
CT scans were obtained with a multidetector scanner (16-slice Philips Mx8000). Whole brain noncontrast CT was performed: 120 kV, 170 mA, 2 second scan time, contiguous 6-mm axial slices. This was followed by perfusion CT, comprising two 40-second series. Each series consisted of one image per slice per second, commencing 5 seconds after IV administration of 40 mL of non-ionic iodinated contrast at a rate of 5 mL/sec via a power injector. Acquisition parameters were 80 kVp and 120 mA. Each perfusion series covers a 24 mm axial section acquired as two adjacent 12-mm slices. The first section was at the level of the basal ganglia/internal capsule, and the second was placed directly above, toward the vertex. Thus, the two perfusion CT series allow assessment of two adjacent 24 mm cerebral sections.
CTA was performed after CTP, using the parameters 120 kV, 125 mAs, slice thickness 1.5 mm, pitch 1.5:1, helical scanning mode, IV administration of 70 mL of non-ionic contrast at 4 mL per second. Bolus-tracking software was used to maximize the chance of image acquisition at peak contrast arrival. Data acquisition was from base of skull to the top of lateral ventricles.
Follow-up imaging used a 1.5 T MRI (Siemens Vision), unless there was a contraindication; then NCCT and CTA were repeated using the above protocol. In brief, the stroke MRI protocol included an axial spin-echo T2-weighted series, an axial isotropic diffusion-weighted echoplanar spin-echo sequence (DWI), time of flight MR angiography (MRA), and perfusion-weighted imaging (PWI) with an axial T2*-weighted echoplanar sequence.15
Image analysis.
All images were de-identified and coded so that NCCT could not be linked to CTP or MRI by the assessors (assessors aware of hemisphere affected). First, NCCT images were reviewed digitally at a workstation (Philips MxView). The analysis was restricted to cortical ASPECTS regions (M1–M6), because focal swelling can only be appreciated in cortical gray matter.5,9 Evaluation of each patient's baseline CT scan was performed independently by two stroke neurologists. Each cortical ASPECTS region was recorded as normal, PH, or IFS. Parenchymal hypoattenuation was defined as a region of abnormally decreased attenuation of other parts of the same structures or of contralateral hemisphere.5 Isolated focal swelling was defined by obvious asymmetric sulcal effacement in the absence of cortical hypoattenuation.9 Disagreements between interpreters were decided by consensus.
De-identified CTP data were analyzed by the two stroke neurologists independently. First, parametric maps of CBV, CBF, and MTT were generated with deconvolution analysis using commercial software to generate CBV, CBF, and MTT maps (MIStar, Apollo Medical Imaging, Melbourne, Australia).12,16 To calculate quantitative perfusion values for each cortical ASPECTS region, the two observers then outlined the six ASPECTS regions (M1–6) on the CTP source images for both hemispheres.4,5 However, only cortical areas in the M1–6 regions were included, based on the greater contrast enhancement seen in cortical tissue compared to subcortical white matter.17 A lower Hounsfield unit threshold was then applied to eliminate any remaining white matter (or CSF) pixels from the regions of interest. Acute perfusion values were then automatically determined for the cortical regions from the respective CBV, CBF, and MTT maps (as pixel size and location is identical to the CTP source images). To minimize the contribution of vascular pixels, pixels with CBV >8 mL/100 g were eliminated from the calculation of regional perfusion values.18
To assess for the presence of major reperfusion, the two observers independently measured the acute CTP MTT lesion volumes and day 3 MR-PWI (MTT) lesion volumes. The MR-MTT maps were generated using deconvolution analysis with the same software used to generate the CTP maps.16 The MR-MTT maps were coregistered using this software to obtain the same spatial position and axis orientation as the CTP-MTT maps. Then, the MR-MTT maps were resliced to correspond to the thickness and location of the CTP-MTT maps. Lesions were outlined on both MR and CTP-MTT maps, and then an automated threshold (corresponding to an MTT delay of >2 seconds compared to the contralateral hemisphere) was used to calculate the lesion volume. The acute CTA circle of Willis maximum intensity projection reconstructions were also reviewed and graded as no, partial, or complete vessel occlusion. Day 3 MRA (or CTA) were also graded in the same manner. These assessments were also performed blinded to patient identity and other imaging data.
To determine whether infarction occurred in cortical ASPECTS regions, two different observers not involved in the above assessments of the baseline scans (a stroke neurologist and a neuroradiologist) assessed the de-identified day 3 imaging. The observers were aware of the hemisphere affected. For MRI, relative hyperintensity on DWI was scored as abnormal. Apparent diffusion coefficient (ADC) maps were viewed concurrently to confirm that the hyperintensity on DWI was due to recent infarction and not T2-shine through from chronic infarction or vasogenic edema.19
Statistical analysis.
The following analyses were then performed, applying the Bonferroni correction for multiple comparisons where appropriate, with STATA (Version 7, College Station, TX, 2001).
1. Mean CBV and CBF for cortical ASPECTS regions was compared across the three NCCT categories (normal, PH, and IFS), as well as with contralateral cortical regions, using paired t-tests. In this analysis, ASPECTS regions on the side of the ischemic lesion that were not hypoperfused on acute CTP were treated separately to normal regions on NCCT with hypoperfusion on CTP. The definition of hypoperfusion was a delay in MTT of >2 seconds in the ASPECTS region compared to the respective contralateral ASPECTS region. An MTT threshold was chosen, as selecting a CBF threshold to define hypoperfused regions would directly influence results critical to this study—absolute CBF values in the different NCCT regions. Furthermore, previous perfusion MR and CT studies have indicated that an MTT threshold best identifies tissue that is hypoperfused.13,20,21 According to this definition, normal-appearing regions on NCCT in the affected hemisphere were defined as either hypoperfused (hypoperfused normal) or not.
2. The proportion of cortical ASPECTS regions in each NCCT category (normal, PH, and IFS) progressing to infarction was compared using Fisher test. This was performed separately for the patients with and without major reperfusion at day 3. Major reperfusion was defined as >80% reduction in MTT lesion volume from acute to day 3 or restoration of complete vessel recanalization by day 3.20
3. From the quantitative CTP analysis performed above, regions in the ischemic hemisphere were classified as i) not hypoperfused (no delay in MTT compared to the respective contralateral ASPECTS region). If there was hypoperfusion (delay in MTT >2 seconds compared to contralateral ASPECTS region), then these regions were classified depending on contralateral CBV as ii) reduced CBV (more than 2 standard deviations below mean contralateral CBV), iii) normal CBV (within 2 standard deviations of mean contralateral CBV), or iv) increased CBV (more than 2 standard deviations above mean contralateral CBV). The proportion of cortical ASPECTS regions in each quantitative CTP category progressing to infarction was compared using Fisher test. This was performed separately for the patients with and without major reperfusion at day 3.
4. Logistic regression analysis was used to predict the likelihood of a cortical ASPECTS region progressing to infarction, based upon either presence of early ischemic change on NCCT or quantitative CTP classification. Again, these analyses were performed separately for patients with and without major reperfusion.
Results.
Of 65 consecutive patients with sub-6 hour clinically suspected anterior circulation stroke, five had primary intracerebral hemorrhage. Ten patients with ischemic stroke had clinical exclusions for thrombolysis. Two patients did not undergo CTP/CTA due to contrast allergy and renal failure. Six patients with clinical lacunar syndromes and no lesion seen on CTP were excluded. Two patients with ischemic lesions restricted to the posterior circulation on imaging were excluded. Therefore, a total of 40 ischemic stroke patients were included in the analysis (median age 73 years, interquartile range [IQR] 63 to 79 years; 20 women). Median baseline NIHSS was 16 (IQR 10 to 20). Median time from symptom onset to acute imaging was 2.5 hours (IQR 1.5 to 4.0 hours). Nineteen patients received open-label thrombolytic therapy with standard dose IV recombinant tissue plasminogen activator (rt-PA). Fourteen patients were treated between 3 and 6 hours after stroke onset with standard dose IV rt-PA or placebo in an ongoing randomized double-blind trial (EPITHET).15,21 Six patients were excluded from thrombolysis because early ischemic change on baseline CT was considered too extensive. A further patient was excluded from thrombolysis as baseline CT showed a small acute subdural hematoma (with also a large acute perfusion lesion on CTP).
Twenty-three patients had major reperfusion at day 3. Fifteen of these patients were treated with open-label rt-PA, and eight received blinded rt-PA/placebo. Thirty-three of 40 patients had a visible acute vessel occlusion on CTA (M1 or M2 of middle cerebral artery). Of the 23 patients with major reperfusion at day 3, 19 had an acute vessel occlusion; 17 of these had complete vessel recanalization and 2 had partial recanalization by day 3. In the group without major reperfusion, 14 had an acute vessel occlusion on CTA, and only one had partial recanalization by day 3.
Early ischemic change subtype analysis.
For cortical ASPECTS regions, 15 of 40 patients had normal scans, 17 had a combination of normal and hypodense regions, 2 had hypodense regions only, and 6 had a combination of normal and IFS regions. By region, there were 58 hypoattenuated regions, 21 with isolated focal swelling, and 123 normal regions that were hypoperfused on CTP. There were also a further 38 regions that had no hypoperfusion on CTP (table). These regions were all normal on NCCT. Notably, all IFS regions were also hypoperfused on CTP (MTT delay > 2 seconds compared to contralateral hemisphere).
Table Perfusion values in cortical ASPECTS regions by type of early ischemic change on NCCT
In the affected hemisphere, acute CBV (mean 2.39 ± 0.46 mL/100 g) in hypoperfused but normal NCCT regions was similar to contralateral hemisphere CBV (mean 2.32 ± 0.21 mL/100 g). Compared to contralateral hemisphere regions, CBV in PH regions was significantly lower (mean 0.87 ± 0.37 mL/100 g), and significantly higher in IFS regions (mean CBV 3.47 ± 0.23 mL/100 g).
In the affected hemisphere, acute CBF in hypoperfused but normal NCCT regions (mean 20.55 ± 3.69 mL/100 g/min) was similar to that of IFS regions. Acute CBF in both hypoperfused normal NCCT and IFS regions was significantly lower than contralateral hemisphere CBF (mean 47.90 ± 5.62 mL/100 g/min). Compared to all other regions, CBF was significantly lower in PH regions (mean 10.22 ± 3.79 mL/100 g/min).
Thus, PH regions had markedly reduced CBV and CBF, and regions with IFS had reduced CBF but elevated CBV. Normal appearing regions on NCCT with hypoperfusion on CTP had reduced CBF but relatively normal CBV, and normal regions on NCCT without hypoperfusion on CTP had normal CBF and CBV (figure 1).
Figure 1. Boxplots of cerebral blood volume (CBV) (top) and cerebral blood flow (CBF) (bottom) in the different noncontrast CT (NCCT) groups. Normal refers to hypoperfused but normal NCCT regions in the ischemic hemisphere. Note that isolated focal swelling (IFS) regions have increased CBV and parenchymal hypoattenuation (PH) regions have decreased CBV compared to contralateral regions. Also, all hypoperfused regions (normal, PH, and IFS) have reduced CBF, but PH regions have the lowest CBF.
Progression of NCCT regions to infarction.
For all patients, 65 of 161 (40%) of all normal regions progressed to infarction, but this included the 38 normal regions that were not hypoperfused on CTP. Excluding these regions, none of which progressed to infarction, meant that 65 of 123 hypoperfused but normal NCCT regions subsequently infarcted (53%), vs 57 of 58 (98%) PH regions, and 6 of 21 (29%) IFS regions. For those with major reperfusion at day 3, 35 of 79 (44%) hypoperfused normal NCCT regions progressed to infarction, compared to all 13 PH regions, and only 1 of 14 (7%) IFS regions (figure 2). Both hypoperfused normal and IFS regions were less likely to progress to infarction than PH regions (p < 0.001, Fisher exact). The proportion of IFS regions progressing to infarction was also less than hypoperfused normal regions (p = 0.01, Fisher exact). Therefore, when major reperfusion occurred, rates of progression to infarction were very high for PH regions, moderate for normal NCCT regions that were also hypoperfused on CTP, and low for IFS regions.
Figure 2. Percentage of hypoperfused regions by noncontrast CT (NCCT) group (top) and CT perfusion (CTP) group (bottom) progressing to infarction. Note that for patients with subsequent major reperfusion (left), there is decreasing risk of progression to infarction from parenchymal hypoattenuation (PH) regions, to normal NCCT regions, to isolated focal swelling (IFS) regions. This is also similar for CTP category, with decreasing risk of infarction from hypoperfused reduced cerebral blood volume (CBV) regions, to hypoperfused normal CBV regions, to hypoperfused increased CBV regions. However, with lack of major reperfusion (right), most regions progress to infarction, regardless of NCCT or CTP category.
For those without major reperfusion, 30 of 44 (68%) hypoperfused normal NCCT regions progressed, compared to 44 of 45 PH (98%) regions and 5 of 6 IFS (83%) regions (figure 2). Hypoperfused normal regions were less likely to progress to infarction than PH regions (p < 0.01, Fisher exact), but as opposed to the major reperfusion group, there was not a significant difference seen between hypoperfused normal and IFS regions progressing to infarction. Thus, in the patients without major reperfusion, rates of progression to infarction were high for all NCCT regions that were hypoperfused on CTP.
Quantitative CTP subtype analysis.
Based upon contralateral regional CBV, normal CBV was defined as 1.91 to 2.75 mL/100 g. Of the 240 regions in the affected hemisphere, 38 (16%) were not hypoperfused (no MTT delay compared to unaffected hemisphere). There were 202 regions in the affected hemisphere with hypoperfusion (MTT delay > 2 seconds compared to unaffected hemisphere). Of these, 66 (27%) were hypoperfused with normal CBV, 82 (34%) hypoperfused with reduced CBV, and 54 (23%) hypoperfused with increased CBV. For hypoperfused regions that were normal on NCCT, 66 (54%) had normal CBV, 24 (20%) had reduced CBV, and 33 (26%) had increased CBV. All 58 PH regions had decreased CBV, and the 21 IFS regions all had increased CBV. Thus, in hypoperfused regions, the CBV may actually be in the normal range, increased, or decreased compared to unaffected hemisphere. Notably, in hypoperfused but normal appearing NCCT regions, the same broad range of CBV (from reduced to increased) was seen.
Progression of quantitative CTP regions to infarction.
For all patients, 78 of 82 (95%) hypoperfused decreased CBV regions progressed to infarction, compared to 34 of 66 (52%) hypoperfused normal CBV regions, and 16 of 54 (30%) hypoperfused increased CBV regions. Progression to infarction did not occur in any of the 38 regions that were not hypoperfused on CTP. For the major reperfusion subgroup, 34/35 (97%) hypoperfused decreased CBV regions progressed to infarction, compared to 14 of 34 (41%) hypoperfused normal CBV regions, and 1 of 38 (3%) hypoperfused increased CBV regions (figure 2). There were differences in the proportion of regions progressing to infarction among all three groups (Fisher exact, p < 0.001). Therefore, for hypoperfused regions on CTP, rates of progression to infarction were very high for decreased CBV regions, moderate for normal CBV regions, and low for increased CBV regions, when major reperfusion occurred.
For those without major reperfusion, 44/47 (94%) hypoperfused decreased CBV regions progressed to infarction, vs 20/32 (63%) hypoperfused normal CBV regions, and 15/16 (94%) hypoperfused increased CBV regions (figure 2). In this group, again, more hypoperfused decreased CBV regions progressed to infarction than hypoperfused normal CBV regions (p = 0.001, Fisher exact). There was a small difference between hypoperfused normal CBV and hypoperfused increased CBV regions progressing to infarction (p = 0.04, Fisher exact), but this was not significant when corrected for multiple comparisons. Thus, in the patients without major reperfusion, rates of progression to infarction were high for all hypoperfused regions, regardless of CBV level.
Logistic regression to predict progression to infarction by NCCT or CTP group.
In the patients with major reperfusion, NCCT group predicted progression to infarction, with the odds of infarction being very high for PH compared to normal or IFS regions (OR = 30.6, 95% CI 4.1 to 230.1, p = 0.001). However, the predictive strength of NCCT category was relatively low (R2 = 0.18), reflecting that although hypodensity on NCCT was a predictor of infarction and IFS was protective against infarction, there were also a large number of normal regions on NCCT with extremely variable risk of progression to infarction. Quantitative CTP group in the patients with major reperfusion was also a predictor of progression to infarction, with the odds of infarction being very high for hypoperfused decreased CBV regions compared to hypoperfused normal CBV, hypoperfused increased CBV, or non-hypoperfused regions (OR = 37.1, 95% CI 9.1 to 150.9, p < 0.001). The predictive strength of CTP category was much greater than for NCCT category (R2 = 0.63). When assessing both NCCT group and quantitative CTP group in a combined regression equation, only CTP group was a predictor of progression to infarction, again with very high odds of infarction for hypoperfused decreased CBV regions compared to the other CTP categories (OR = 46.4, 95% CI 6.5 to 333.9, p < 0.001). In contrast, NCCT group was not an independent predictor of infarction in the combined equation (OR = 0.6, 95% CI 0.04 to 9.8, p = 0.72). The predictive strength of the combined regression equation was unchanged from that of CTP category alone (R2 = 0.63). This is because CTP category has the advantage of stratifying the risk of infarction of normal-appearing regions on NCCT. Also, given that all PH regions on NCCT were in the decreased CBV category on CTP (i.e., high risk of infarction), and that all IFS regions were in the increased CBV category (low risk), then it follows that NCCT category does not provide any additional predictive accuracy over CTP category.
In the patients without major reperfusion, similar results to the major reperfusion patients were seen for the individual regression equations, although the predictive strength of the equations and ORs comparing risk of infarction between the categories in each group were much lower. This indicated higher rates of subsequent infarction for normal NCCT (or hypoperfused, normal CBV) and IFS (or hypoperfused, increased CBV) regions. For NCCT group, the OR for PH regions to progress to infarction vs normal or IFS regions was 5.0 (95% CI 2.0 to 12.5, p < 0.001, R2 = 0.13). For CTP group, the OR for hypoperfused decreased CBV regions progressing to infarction compared to the other CTP groups was 3.3 (95% CI 2.0 to 5.5, p < 0.001, R2 = 0.23). Again, in the combined regression equation, CTP group was an independent predictor of progression to infarction (OR = 3.0, 95% CI 1.6 to 5.8, p = 0.001), whereas NCCT group was not (OR = 1.3, 95% CI 0.4 to 4.3, p = 0.67). Similarly, the predictive strength of the combined regression equation was unchanged from that of CTP category alone (R2 = 0.23).
Discussion.
This study correlates early ischemic change on NCCT with immediate assessment of cerebral perfusion (figure 3). We have demonstrated that the pathophysiology of IFS is markedly different from PH. Regions with PH have substantially reduced CBV and CBF, and are irreversibly injured. Regions with IFS have reduced CBF but elevated CBV, and are potentially salvageable from infarction with major reperfusion, indicating penumbral tissue. However, the majority of acutely hypoperfused regions appear normal on NCCT. Normal appearing regions on NCCT when assessed with CTP may be irreversibly ischemic (reduced CBV), or be hypoperfused and at risk of progression to infarction (normal or increased CBV), or not hypoperfused at all. For this reason, perfusion CT provides improved predictive accuracy over NCCT alone.
Figure 3. Patient A: Isolated focal swelling (IFS) on noncontrast CT (NCCT), hypoperfusion on mean transit time (MTT), and increased cerebral blood volume (CBV) on acute CT perfusion (CTP) maps, and no progression to infarction with subsequent major reperfusion. Patient B: IFS on NCCT, hypoperfusion on MTT, and increased CBV on acute CTP maps, but progression to infarction occurred without major reperfusion. Patient C: Hypoperfusion on MTT and increased CBV on acute CTP maps, without any change apparent on acute NCCT. No infarction in cortical regions on follow-up with major reperfusion. Patient D: Hypoperfusion on MTT and decreased CBV on acute CTP maps without any apparent change on NCCT. Subsequent infarction present in reduced CBV regions on follow-up MR. Patient E: Profound decrease in CBV and CBF on acute CTP maps, with associated parenchymal hypoattenuation on NCCT. Extensive infarction on follow-up MR.
The current study extends knowledge about the pathophysiology of early ischemic change.8,9,22 A recent study suggested that IFS was associated with increased CBV on perfusion MR, but there was a considerable delay between NCCT and MRI in most patients.9 This study suggested that IFS could represent benign oligemia, or penumbral tissue, but did not control for subsequent reperfusion and, due to the technique, could only provide relative CBF values. Our results with immediate perfusion imaging confirm that IFS is associated with increased CBV, but show that IFS regions had substantial absolute CBF reduction, likely to be in the penumbral (i.e., 15 to 25 mL/100 g/min), rather than in the oligemic, range (table).23–25 Additionally, all IFS regions had an MTT delay of >2 seconds compared to the normal hemisphere, which is a quite accurate perfusion threshold to distinguish critical hypoperfusion from oligemia.20,21 Nonetheless, we cannot definitively exclude that some IFS regions reflected benign oligemia rather than ischemic penumbra as IFS regions had a low risk of infarction when major reperfusion occurred. However, in the absence of major reperfusion, the majority of IFS regions did progress to infarction, indicating that IFS is generally not a benign phenomenon.
The cerebral perfusion changes occurring with the two types of EIC highlight their markedly differing pathophysiology. The strong correlation seen with increased CBV indicates that IFS is related to cerebral hyperemia due to the compensatory vasodilatation that occurs as an initial response to decreased cerebral perfusion.26 Conversely, PH was associated with a substantial decrease in CBV and CBF. It has been recognized for some time that PH occurs with severe, and probably irreversible, ischemia.27,28 Indeed, it has been accepted that the pathophysiology of PH relates to ischemic edema following severe CBF reduction.27 Our study confirms that PH occurs only with profound CBF reduction, with mean values in these regions below the accepted thresholds for irreversible ischemia (i.e., <15 mL/100 g/min).23–25 However, it has been postulated that ischemic edema may not necessarily be the complete explanation for PH, and that gray matter hypoattenuation may also reflect loss of the normal CBV gradient between gray and white matter.29 Our study lends weight to this intriguing hypothesis, as PH regions had markedly reduced CBV compared to other regions. Therefore, PH and IFS reflect the extremes of CBV alteration in ischemia. IFS appears to directly result from compensatory vasodilatation (which leads to substantial increases in regional blood content), and PH may result from a combination of ischemic edema and relative reduction in regional blood content.
Although NCCT can identify the outer limits of CBV alteration in acute ischemia, substantial but less extreme changes in CBV were not apparent on NCCT. This is highlighted by the greater predictive accuracy for subsequent infarction by quantitative acute CTP classification compared to NCCT classification. Within hypoperfused but normal appearing NCCT regions, the range of CBV was broad compared to contralateral regions (figure 1). Obviously, the other advantage of CTP is to distinguish between hypoperfused but normal-appearing NCCT regions and NCCT regions that are normal because they are not ischemic. Thus, the quantitative CTP assessment allowed stratification of the risk of infarction from almost certain (hypoperfused with reduced CBV) to zero (no hypoperfusion). In between these extremes, there is tissue at risk of progression to infarction, and the level of risk depends both on CBV level and whether subsequent major reperfusion occurs. Hypoperfused regions with increased CBV had a very low probability of infarction with major reperfusion, and hypoperfused regions with normal CBV had a higher risk (albeit much lower than decreased CBV regions) of progression to infarction with major reperfusion. In contrast, both hypoperfused regions with normal CBV and hypoperfused regions with increased CBV had similar and very high rates of progression to infarction without major reperfusion. Thus, hypoperfused regions with reduced CBV are irreversibly injured (ischemic core), and hypoperfused regions with normal or increased CBV are penumbral. Undoubtedly, increased CBV is a good prognostic marker of protection against infarction with major reperfusion. Again, it is possible that using an MTT delay of >2 seconds to define hypoperfusion may have included some oligemic regions with increased CBV. However, 85% of hypoperfused regions with increased CBV progressed to infarction without major reperfusion. Therefore, increased CBV in the presence of hypoperfusion is not a benign finding, and (similar to our IFS findings) these regions are still at considerable risk of infarction without major reperfusion.
It is important to note potential limitations of this study. First, we believe that further work is required to fully validate absolute perfusion values generated with CTP. However, absolute CT perfusion values have been previously shown to be similar to xenon CT and PET values.10,11 Indeed, it is a major advantage of this study that the assessment of cerebral perfusion immediately followed NCCT with a technique that can provide absolute perfusion values rather than relative values (e.g., bolus-tracking MRI). Not knowing precisely when major reperfusion occurs is also a limitation of this study. However, the presence of major reperfusion at day 3 very closely correlates with extensive salvage of penumbra from infarction, and is an excellent predictor of response to thrombolysis.15,20 Another potential limitation is that the treatment allocated to 14 patients (rt-PA or placebo) remains blinded. However, subsequent reperfusion status is much more critical to interpretation of our results than treatment received. Although thrombolytic treatment increases major reperfusion, whether this is going to occur in an individual patient cannot be predicted.15,16,20 Therefore, what a clinician needs to know when making a decision about a treatment that can increase the chance of major reperfusion, but which has considerable risk, is whether there is a substantial difference in tissue fate with and without major reperfusion.
Acknowledgment
The authors thank the CT and MR radiographers at their hospital.
Footnotes
-
Commentary, see page 717
Supported by a National Health and Medical Research Council of Australia Project Grant (ID 351155).
Disclosure: The authors report no conflicts of interest.
Received July 10, 2006. Accepted in final form November 13, 2006.
References
Disputes & Debates: Rapid online correspondence
NOTE: All authors' disclosures must be entered and current in our database before comments can be posted. Enter and update disclosures at http://submit.neurology.org. Exception: replies to comments concerning an article you originally authored do not require updated disclosures.
- Stay timely. Submit only on articles published within the last 8 weeks.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- 200 words maximum.
- 5 references maximum. Reference 1 must be the article on which you are commenting.
- 5 authors maximum. Exception: replies can include all original authors of the article.
- Submitted comments are subject to editing and editor review prior to posting.