(Stroke. 1999;30:1490A-1493.)
© 1999 American Heart Association, Inc.
Letters to the Editor |
SPECT-Derived Relative Perfusion Defect and CT-Derived Hypodense Region in Acute Intracerebral Hemorrhage
Division of Neurology, Department of Medicine, Queen Mary Hospital, Hong Kong
To the Editor:
In their recent, interesting article,1 Mayer and colleagues report that the flow deficit volume (FDV) was greater at the earlier hours after acute intracerebral hemorrhage (ICH) than at the later hours and that the CT-derived hypodense region was larger in the later scans than in the earlier scans. In this study, FDV was obtained by subtracting the CT-derived ICH volume from the volume of SPECT-derived relative perfusion defect. There is a minor typing error in the text: "moderate-to-severe" global blood flow reduction on SPECT study was stated as an exclusion criteria in the "Study Population" subsection, whereas "mild-to-severe" global blood flow reduction on SPECT study was quoted in the "Validation of SPECT Analysis" subsection. Significant midline shift (>5 mm) on CT scan was an exclusion criteria, and the midline shift of the CT scan in Figure 1 in the article 1 appears to be >5 mm. I would raise the following comments.
First, SPECT study can assess perfusion and predict prognosis for recovery.2 3 Nevertheless, SPECT study provides only a relative index of perfusion, and there is no true "zero-flow" region. In contrast to the tracer 99mTc-hexamethylpropyleneamine oxime, which reveals relative blood flow, the tracer 99mTc-ethyl-cysteinate-dimer (ECD) reflects not only relative perfusion but also the metabolic status of the brain tissue. Thus, 99mTc-ECD is more specific for both functioning brain tissue and infarcted brain lesions.3
Second, the FDV was derived from both the CT and SPECT studies, and the time interval between the 2 kinds of imaging studies was not constant.1 The delay from onset of ICH should be based on the SPECT studies rather than the CT studies, because the SPECT-derived relative perfusion changed with time and the CT-derived ICH volume was constant over time. In fact, the time interval after ICH onset for the acute-phase SPECT scans (2 to 36 hours) overlapped with that for the subacute-phase SPECT scans (30 to 136 hours).
Third, univariate analysis and multiple regression analysis were done for the ratios of edema/ICH volume and of FDV/ICH volume, and then ICH volume was found to a significant factor with a negative association. I am interested to know the rationale behind the use of these ratios instead of edema and FDV, and I wonder whether ICH volume was a significant factor because it was used as the denominator.
Fourth, SPECT-derived relative perfusion defects were greater or equal to the corresponding CT-derived ICH volumes in 40 paired measurements, resulting in negative FDV values in 6 measurements. I wonder whether these negative values were used in the statistical analysis. I think negative FDV values should be replaced by zero, since negative FDV values do not have any physiological meaning.
Finally, CT-derived hypodense region in the acute-phase scans is, in theory, a combination of acute interstitial edema and ischemia due to elevated local tissue pressure. In contrast, CT-derived hypodense region in the subacute-phase scans is mainly contributed by vasogenic and inflammatory edema.4 I wonder whether there is any correlation between the CT-derived hypodense region and the FDV in the acute phase.
References
1.
Mayer SA, Lignelli A, Fink ME, Kessler DB,
Thomas CE, Swarup R, Van Heertum RL. Perilesional blood flow and
edema formation in acute intracerebral
hemorrhage: a SPECT study. Stroke.. 1998;29:1791–1798.
2.
Alexandrov AV, Black SE, Ehrlich LE, Bladin CF,
Smurawska LT, Pirisi A, Caldwell CB. Simple visual analysis of
brain perfusion on HMPAO SPECT predicts early outcome in acute stroke.
Stroke.. 1996;27:1537–1542.
3.
Berrouschot J, Barthel H, von Kummer R, Knapp WH, Hesse
S, Schneider D.
99mTechnetium-ethyl-cysteinate-dimer
single-photon emission CT can predict fatal ischemic brain
edema. Stroke.. 1998;29:2556–2562.
4.
Wagner KR, Xi G, Hua YM, de Courten-Myers GM, Myers RE,
Broderick JP, Brott TG. Lobar intracerebral
hemorrhage model in pigs: rapid edema development in
perihematomal white matter. Stroke.. 1996;27:490–497.
Response
Department of Neurology
Department of Radiology, Columbia University College of Physicians and Surgeons, New York, NY
Key Words: intracerebral hemorrhage • cerebral blood flow
We thank Dr Cheung for his thoughtful comments regarding our article. When we designed our study, we accepted the inaccuracies inherent in analyzing SPECT and CT data obtained at different times and understood that no method of analysis would be completely satisfactory. Without image coregistration, we recognized that any attempt to quantify cerebral blood flow (CBF) within a small perihematoma region of interest (ROI) would be impossible, because there would be no way to know whether the ROI was precisely on the border of the clot. Instead, we developed a novel volumetric approach for measuring perfusion around each hematoma by subtracting CT-derived ICH volumes from theoretical SPECT-derived volumes of brain tissue with zero flow (based on the percent reduction in tracer counts compared with a mirror ROI on the contralateral side) to yield a "flow deficit volume." This technique was originally developed and validated by Mountz1 as a method for quantifying perfusion in ischemic stroke. Obviously, these FDVs are a theoretical construct, because it is almost certain that CBF decreases gradually around a hemorrhage. Despite the inherent drawbacks of this method, including the potential for inaccuracy due to anatomic brain distortion or contralateral CBF reduction, the huge advantage was that we were able to obtain FDVs with good reliability regardless of how the ROI was drawn, as long as it was drawn widely. Imprecision may have also resulted from comparison of volumetric measurements obtained by different modalities (CT and SPECT), but we established the validity of this approach by demonstrating that both methods were able to estimate the volume of a brain phantom with reasonable accuracy (<2% error).
Obviously, the CT and SPECT studies could not be performed at exactly the same time, and this also may have led to inaccuracy. We attempted to obtain the acute-phase studies between 0 and 24 hours after ICH and the subacute-phase studies between 48 and 72 hours. The average interval between the CT and SPECT examination was 6.7 hours during the acute phase and 2.7 hours during the subacute phase. Although significant changes in ICH volume, edema volume, or perihematoma perfusion may have occurred during these intervals, we feel that this is unlikely. Because SPECT was not available on weekends, we allowed for protocol violations, including an initial SPECT as late as 36 hours after ICH and follow-up scans as early as 24 hours after the initial scan. One patient underwent SPECT 31.5 and 103.5 hours after ICH, and hence this "acute" SPECT was performed later than the earliest "subacute" study, which was obtained at 30 hours. We allowed this discrepancy because our intent was to measure trends in perfusion and because the degree of overlap was relatively small.
We analyzed FDV/ICH volume and edema/ICH volume to assess for factors that might influence the relative extent of perfusion and edema around a clot. These ratios were used to correct for ICH volume, which correlated strongly with both of these measurements. Given the variable timing of the SPECT scans, this analysis also provided us with an opportunity to confirm our main findings by looking at the effect of time after ICH on these ratios. In fact, time after ICH was the only variable found to influence the relative extent of perfusion around a hematoma, with longer intervals correlating with smaller FDV/ICH ratios, as expected. In retrospect, we agree with Dr Cheung that the observed negative correlation between ICH volume and edema/ICH volume is probably explained by inclusion of this variable in the denominator. We did not analyze FDV/edema volume because we felt the meaning of this ratio would be hard to interpret.
We used 99mTc-hexamethylpropyleneamine oxime because we sought to assess perfusion. Use of 99mTc-ECD would have confounded our analysis, because its uptake is related to perfusion as well as metabolism (eg, uptake is reduced in perfused but hypometabolic tissue). If the cortical hyperperfusion that we observed in some cases was not related to an increase in metabolic activity, which seems likely, ECD might not have detected it.
In 6 of 46 paired measurements, the SPECT-derived zero-flow volume was slightly smaller than the CT-derived ICH volume, yielding a negative FDV value. Viewed another way, however, with just one exception all FDV values exceeded or equaled the corresponding CT-derived ICH volume within a small margin of error (±2 mL), which we feel confirms the validity of our approach. We used these negative FDV values rather than zero in our analysis, because conceptually they indicate hyperperfusion. In fact, in the only instance of a large negative FDV (-8 ml) there was visible hyperperfusion in the ROI around the hematoma.
The pathophysiology of perihematoma brain injury in ICH is complex. We agree with Dr Cheung that on acute CT scans, hypodensity probably represents interstitial edema from leakage of plasma, and that during the subacute phase there is an additional component of inflammatory vasogenic edema. There was only a weak, nonsignificant correlation between FDV/ICH volume and edema/ICH volume during the acute phase (r=0.36, P=0.09), suggesting that the initial extent of perfusion deficit is not a particularly important determinant of edema in the acute stage of ICH.
References
1. Mountz JM. A method of analysis of SPECT blood flow image data for comparison with computed tomography. Clin Nucl Med.. 1989;14:192–196.[Medline] [Order article via Infotrieve]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||