Measurement of Regional Cerebral Blood Volume by Emission Tomography Robert L. Grubb, Jr, MD, Marcus E. Raichle, MD, Carol S. Higgins, BA, and John 0. Eichling, PhD
The technique of positron emission tomography was used to measure cerebral blood volume (CBV) in 10 normal right-handed human volunteers following inhalation of trace quantities of cyclotron-produced, "C-labeled carbon monoxide. In scans obtained 4 cm above the orbitomeatal line, CBV was 4.3 ml per 100 gm of tissue, whereas in scans obtained 8 cm above the orbitomeatal line, CBV was significantly less (3.3 ml per 1 0 0 gm; p < 0.001). This difference reflects the greater proportion of gray matter in the lower scan. Furthermore, the CBV was significantly larger Cp < 0.001) in the left cerebral hemisphere in the tomographic scans obtained 4 cm above the orbitomeatal line. These scans include the region of the superior surface of the temporal lobe (planum temporale), which is thought to be larger in individuals with left cerebral dominance for speech. This observation is the first in vivo demonstration of a structural correlate of a known functional difference in the cerebral hemispheres of man. Grubb RL Jr, Raichle ME, HigRins CS, et al: Measurement of regional cerebral blood volume by emission tomography. Ann Neurol 4:322-328, 1978 Present methods of studying brain hemodynamics (i.e., blood flow and blood volume) in vivo have serious shortcomings, particularly when applied to studies in humans [ 2 , 171. The appearance of linear accelerators and cyclotrons in the medical environment, the parallel development of rapid chemical techniques for incorporating their products (i.e., short-lived, positron-emitting radionuclides lSO,I3N, I1C, ISF) into compounds [23], plus recent developments in imaging systems employing the concept of positron emission tomography offer an opportunity to circumvent most of these shortcomings. Emission tomography is a nuclear medicine visualization technique that yields an image of the distribution of a previously administered radionuclide in any desired transverse section of the body. Positron emission tomography utilizes the unique properties of the annihilation radiation generated when positrons are absorbed in matter. It is characterized by the fact that an image reconstructed from the radioactive counting data is a highly faithful representation of the spatial distribution of a radionuclide in the chosen section. The resulting data make it possible to calculate values for parameters of physiological and biochemical significance when used with appropriate mathematical models. This approach is analogous to quantitative tissue autoradiography with the added advantage of allowing in vivo studies. In this report we describe the basis for a low-risk, quantitative, in vivo method of measuring cerebral
blood volume (CBV) regionally in the human brain. Our method employs emission tomography to monitor externally the spatial distribution in brain of carboxyhemoglobin labeled with the short-lived (20-minute half-life), cyclotron-produced, positronemitting radionuclide carbon 11. The data for our report were obtained from I 0 normal human volunteers.
From the Department of Neurology and Neurological Surgery and the Division of Radiation Sciences, The Edward Mallinckrodt Institute of Radiology, Washington University School of Medicine, St. Louis, MO.
Accepted for publication Mar 30, 1978.
Methods Cerebral blood volume was measured in 10 right-handed normal male volunteers. Informed consent for these studies was obtained in accordance with the regulations of the Washington University School of Medicine Human Use Committee. Emission tomographic scans of the head were performed following inhalation of "C-labeled carbon monoxide. A positron emission tomograph developed in our laboratory was used to quantitatively image the cross-sectional blood volume. The basic principles, design, and performance characteristics of the positron emission tomograph used in this study have been previously described [ 3 , 7, 14-16, 2 11.
The subjects for this study were in a resting normocapnic state. End-tidal carbon dioxide tension was monitored with a capnograph. An estimate of arterial carbon dioxide tension (P+O,) was calculated by multiplying the end-tidal carbon dioxide tension by atmospheric pressure with a correction for water vapor pressure. Each subject was positioned in the tomograph with the aid of a low-power laser beam of light across the center plane of the tomograph. By having the subject breathe approximately 10 to 15
Address reprint requests tO Dr Grubb, of Neurology and Neurological Surgery, Barnes Hospital Plaza, St. Louis, MO 631 10.
322 0364-5134/78/0004-0406S01.25 @ 1978 by Robert L. Grubb, Jr
mCi of high-specific-activity "C-labeled carbon monoxide contained in 800 fil of helium and 1,000 nil of room air from an Ambu bag, the blood was labeled with "Ccarboxyhemoglobin. Following inhalation of the gas, 6 to 7 minutes were allowed for the labeled blood to reach an equilibrium state. Two emission tomographic scans were then obtained sequentially 4 and 8 cm above the orbitomeatal (0-M) line. Data were collected for a period long enough to ensure sufficient counts in the image for an accurate image reconstruction (range, 370,000 to 1,000,000 counts; mean, 600,000 counts). This usually required 6 minutes for the first scan and 10 minutes for the subsequent one. Scans of the brain were computer processed in order to calculate a mean value of the scan for each region of interest. During each emission scan, several blood samples were drawn from a peripheral vein. These blood samples were weighed and counted in a 7.5 x 7.5 cm thallium-activated sodium iodide well detector to obtain the activity of the labeled blood (counts per second per gram of blood). A curve of the activity of "C-tagged carbon monoxide in the blood, corrected for the physical decay of "C from the beginning of the emission scan, was constructed as a function of time. The value for blood "C carbon monoxide activity used in the calculation of CBV (see the equation) was obtained from this curve by selecting the point in time corresponding to the midpoint of the emission scan. Calibration of the emission tomograph in order to determine the actual concentration of "C-labeled carbon monoxide in brain from the emission scan (counts per second per cubic centimeter of tissue), was performed by iniaging an appropriately designed phantom. This phantom, a cylinder 18 cm in diameter divided into five pie-shaped chambers of equal size, was filled with varying concentrations of "C-labeled bicarbonate in saline (typically a 1:2:3:4:5ratio). T h e phantom was then scanned twice. Because of the short half-life of " C (approximately 20 minutes), we obtained ten concentrations of radioactivity which bracketed the radioactivity concentration of "Clabeled carbon monoxide in the subject's scan. Weighted aliquots were then taken from each phantom chamber and counted in the same well detector used to count the blood samples. Scans of the phantom were computer processed in order to calculate a mean relative value of the scan for each chamber. A regression equation was then obtained for these data comparing the relative scan data and the directly measured activity from the phantom. This equation was used to determine the actual concentration of "C-labeled carbon monoxide in the subject's scan. Subject and phantom scan data from the emission tomograph were acquited to yield a spatial resolution of 15 mm full width at half maximum. Corrections for detector nonuniformity, physical decay of the radionuclide, and attenuation were made, and images were reconstructed using an Interdata Model 70 computer. T h e attenuation correction factors were computed using an experimentally determined value for the attenuation coefficient and the physical dimensions of the scanned object. Each data point is where x is the corrected for attenuation by the factor efiXx, thickness of the object along the coincidence line a n d F is the experimentally derived average linear attenuation
coefficient. The validity of this method has been previously reported [ 1 5 ] . The scan reconstruction algorithm was a convolution-based, filtered back projection that resulted in a 100 x 100 array representing the 48 cm diameter field; thus, each picture element size is 0.48 cm. Images were displayed and regions of interest sclected o n a 256 X 240 imaging system with 64 gray levels. Cerebral blood volume (milliliters per 100 g m of brain tissue) was calculated by the equation: CBV =
Cbr
Cb, * f *P h , ' P,
x 100
where Cbr is the concentration of "C-labeled carbon monoxide (counts per second per cubic centimeter of tissue) in the brain determined by the emission tomographic scan; Cbl is the concentration of "C-tagged carbon monoxide (counts per second per gram of blood) in the peripheral venous blood; f is the ratio of the mean cerebral small vessel hematocrit to large vessel hematocrit; Pblis the density of blood (gm per milliliter), and Pt is the density of brain tissue (gm per milliliter). An average value of 0.85 for the ratio of cerebral small vessel hematocrit to large vessel hematocrit was used in these studies [5]. A value of 1.05 g m per milliliter was used for both the density of blood and the density of brain tissue. The significance of differences between values of CBV obtained at 4 cm and 8 cm above the 0 - M line was tested by t tests, and paired t tests were used to test the significance of differences in CBV values obtained in the left and right cerebral hemispheres
Results Figures 1 and 2 show horizontal sections of the brain and "C carbon monoxide scans 4 and 8 cm, respectively, above the 0 - M line. The values of CBV and estimated P k o Z are contained in Table 1. In scans obtained 4 cm above the 0 - M line, an average CBV value of 4.3 (k0.4 SD) ml per 100 gm of tissue was found. At 8 cm above the 0 - M line the average value was 3.3 (k0.5 SD). The differences in CBV value at the two different levels scanned is significant ( p < 0.001). The mean estimated Paco2 in scans 4 cm above the 0 - M line was 4 1 (25 SD);8 cm above the 0 - M line the mean estimated Paco, was 38 ( - t 2 SD). The differences in Paco, were not significant. At 4 cm above the 0 - M line the CBV in the left cerebral hemisphere was 4.3 (k0.5 SD) ml per 100 gm of tissue while the CBV in the right cerebral hemisphere was 4.1 ( k 0 . 4 SD) ml per 100 gm.These left-right cerebral hemisphere differences in CBV 4 cm above the 0 - M line were significant ( p < 0.001). CBV was slightly greater in the left than in the right cerebral hemisphere 8 cm above the 0 - M line, but these differences were not significant. Figure 3 illustrates the typical area of a scan used to calculate whole-brain CBV. Figure 4 shows the typical areas of a scan used t o compute CBV for the left and right cerebral hemispheres.
Grubb et al: Cerebral Blood Volume in Humans 323
F i g 1 . (A) Horizontulsectiom of u huniun bruin. ( B , Emi.r.rivn t oniopcrphi(.srun after inhulution lif C-labrhd 1.urho in nionoxide. Brain slice and totnogruphic slice were both obtcrined 4 cni ubove the orbitomeutulline. The “hot.spot” in the posterior portion of the emission tomogruphir scavi vepre.sents the superior .wgittcrl.rinu.r.
’‘
Fzg 2. (A)Horizontalsection of a human brain. ( B )EmiJsion
tomographicscan afrer inhalation of ’%labeled t arbon nionoxide. Brain dice and tomographic slice were both obtained 8 cm about the orbitomru~dline.
324
Annals of Ncurology
Vol 4
No 4
October 1378
Table 1 . Cerebral Blood Volume und Estimated Arterial Carbon Dioxide Tension i n 10 Subjects
CBV, Whole Brain Subject
(m11100 gm tissue)
A Scan 4 cm above 0 - M line 1 4.0
2 3
4 5 6 7
8 Mean
4.1 5.0 4.8 4.8 4.7 4.4 4.2 4.1 4.2 4.3 4.0 4.0 3.8 3.6 4.3 (20.4 SD) (N = 15)
B Scan 8 cm above 0 - M line 8 3.1 2.9 9 3.1 3.1 10 4.0 3.8 Mean 3.3 (20.5 SD) ( N = 6)
CHV
CBV,Lcft Cerebral Hemisphere (m11100 gm tissue)
CBV,Right Cerebral (mV100 gm tissue)
(mm kg)
4.2 4.3 5.0 4.8 4.9 4.6 4.6 4.3
3.9 4.0 4.8 4.7 4.4 4.5 4.3 4.0 3.8 4.0 4.4 4.0 3.7 3.5 i .4 4.1 (20.4 SD) ( N = 15)
37 36 40 41 48 48 40 41
35 41 ( 2 5 SD) ( N = 9)
3.1 2.6 3.0 2.9 3.6 3.1 3.1 (?0.3 SD) ( N = 6)
36 39 37 36 41 40 38 ( ? 2 SD) ( N 6)
4.2
4.3 4.6 4.2 3.9 3.6 3.2 4.3 ( 2 0 . 5 SD) ( N = 15) 3.1 2.8 3.0 3.0
3.9 3.6 3.2 (20.4 SD) (N = 6)
Hemisphere
Estimated
Pace.
... ... ... ... ... ...
= cerebral blood volume.
Discussion Our results demonstrate the feasibility of quantitatively measuring regional CBV in vivo using positron emission tomography to image circulating "Clabeled carboxyhemoglobin. Because blood can be easily labeled by inhalation of trace quantities of the "C-labeled gas carbon monoxide, and quantitation requires only sampling of peripheral venous blood, this measurement is associated with minimal risk and discomfort to the subject. Furthermore, we consider the risk from exposure to ionizing radiation acceptable. For example, in a typical study in which the subject inhales 10 to 15 mCi of "C-labeled carbon monoxide, approximately 3 to 5 mCi actually labels the blood, and the remainder is quickly ventilated. Under these circumstances the blood receives approximately 50 mR per millicurie and the wholebody radiation dose is approximately 10 m R per millicurie. In addition to its ease of administration and the low radiation dose to the 'subject, our tracer, "C-labeled
carboxyhemoglobin, has the additional advantage of selectively labeling red blood cells. The use of labeled red cells to measure CBV circumvents the potential problem of extravascular migration of the tracer when damage has occurred to the blood-brain barrier. This is always a potential problem that cannot be anticipated when diseased tissue is studied. Data based on studies employing a plasma tracer must, as pointed out by others [2, 171, be viewed with caution when damage to the blood-brain barrier is suspected. The use of a red cell tracer-or, for that matter, a plasma tracer-for measurement of CBV necessitates knowledge of the local tissue hematocrit (see the equation). As pointed out by others, this may differ from the large vessel hematocrit. Thus, a correction must be inserted in the calculation of CBV for this discrepancy. In normal brain, the limited data currently available suggest a value of 0.85 for the ratio of small vessel or tissue hematocrit to large vessel hematocrit [51. Rosenblum [ 191 has cautioned that the local tissue hematocrit may change in acute dis-
Grubb
et
al: Cerebral Blood Volume in Humans 325
Fig 3. Dortedlinesurroundr typicalarea of "C carbon monoxide uitivity in emission tomographicscan used t o calculate wholebrain CBV.
Fig4. Dottedlinessurraundtypicalareasof "C carbon monoxide aitivity in emission tomngraphicsi-anused to calculate CBV i n left and right cerebral hemispheres.
eases of the brain such as ischemic infarction. Under such circumstances, errors in the computation of CBV couid be introduced by selection of an arbitrary value. In the future it should be possible to characterize the local tissue hematocrit more thoroughly in normal as well as diseased brain by sequential measurement of plasma and red cell volume in the same patient by emission tomography. This could be accomplished by the use of "C carbon monoxidelabeled red blood cells followed by intravenous administration of "C-labeled methylalbumin [20]. The precise measurement of CBV has proved to be difficult, and few studies in humans have been done [ 5 , 8-1 1 , 131. Table 2 lists values of CBV found in normal humans utilizing several different merhods. With the exceprion of one srudy [ I I ] employing phosphorus 32-labeled red blood cells, all these studies are in fairly close agreement. Three observations in our data are worthy of note. First, the average of CBV in normal humans was higher in emission tomographic scans obtained 4 cm above the 0 - M line compared with scans obtained 8
cm above this reference point. Although the mean estimated PacoZ was slightly higher in the subjects scanned 4 cm above the 0 - M line, this would not account for the differences in CBV seen at the two levels [ b ] .That different values of CBV were obtained at these two levels is not surprising, as different cross-sectional levels of the brain contain different ratios of gray to white matter. Gray matter is known to have a higher capillary density [l] and a higher resting cerebral blood flow [12]. As can be seen in Figures 1 and 2, emission tomographic scans obtained 4 cm above the 0 - M line contain a larger amount of deeper gray structures compared to scans 8 cm above this point. Second, in scans obtained 4 cm above the 0 - M line, the CBV was found to be significantly greater in the left cerebral hemisphere. This was, by history, the dominant hemisphere in all the subjects studied. Functional differences between the cerebral hemispheres of man have been noted for many years, but until recently, structural differences between the hemispheres have not been observed. However, in
326
Annals of Neurology
Vol 4
No 4
October 1078
Table 2 . Previously Determined Values far CBV in Normal Humans Average Value
of CBV (mli 100 Method
g m tissue)
Positron emission tomographic scan Radionuclide emission tomographic scan [81 X-ray computed tomographic scan
3.3-4.3 2-4
3
[I31 X-ray fluorescence [ 5 ] G a m m a camera mean transit time (gYmTc) [9, 101 Dye-dilution technique m e a n transit time (3LPP-labeledR B C s ) [ 1 I ]
3.2 4.05
9
1968, Geschwind and Levitsky [4]found the superior surface of the temporal lobe (planum temporale) to be larger in the left hemisphere in 65% of brains examined at autopsy. This observation was extended by Witelson and Pallie [22] in 1973 by their demonstration that this asymmetry is present in neonates as well as adults. Our finding of an increase in CBV in this area is further evidence of a structural asymmetry underlying functional differences between the two hemispheres. Our finding assumes special importance when it is realized that for the first time, such differences can be determined safely in vivo. We recognize the need for additional work in this area. Of particular interest would be the demonstration of an increase in regional CBV in the right hemisphere in patients known to have right cerebral dominance for language. Finally, our scans clearly reveal regional differences in "C activity, and hence CBV, within each cerebral hemisphere, corresponding to anticipated differences in CBV in gray and white matter structures. We are reluctant to quantitate these differences at this time because the emission tomograph used in these studies has a minimum resolution of 1.5 cm3. We do not consider this resolution sufficient t o allow quantitative independent determination of CBV regionally in gray and white matter. Anticipated advances in the design and resolution of emission tomographs, however, should make such independent measurements available in the near future. In summary, our studies demonstrate the feasihility of safely making quantitative in vivo measurements of regional CBV in humans. Such measurements should greatly facilitate studies designed to understand cerebral hemodynamics in the normal as well as diseased human central nervous system.
Furthermore, they should complement measurements of regional cerebral metabolism by emission tomography [ 181. Supported by US Public Health Service Grants HL 1385 1 and NSO 6833 (NINCDS) and by Teacher-Investigator Award NS 11059 from NINCDS (Dr Raichle). The authors wish to thank the staff of the Washington University School of Medicine cyclotron for their invaluable technical assistance in these experiments. We are grateful to Dr Michel M. TerPogossian for his generous support and helpful suggestions during the course of this work.
References I . Blinkov SM, Glezer 11: The Human Brain in Figures and Tables. New York, Plenum, 1968, p 267 2. Eichling JO, Gado MH, Grubb RL Jr. et al: Potential pitfalls in the measurement of regional cerebral blood volume, in Harper M, Jennett B, Miller D , et a1 (eds): Blood Flow and Metabolism in the Brain, Proceedings of the 7th International Symposium on Cerebral Blood Flow and Metabolism. Edinburgh and London, Churchill/Livingstone, 1975, pp 7.157.19 3. Eichling JO, Higgins CS, Ter-Pogossian MM: Determination of radionuclide concentrations with positron CT scanning (PETT). J NLKI Med 18:845-847, 1977 4. Geschwind N , Levitsky W: Human brain: left-right asymmetries in temporal speech region. Science 161: 186-1 8 7 , 1968 5 . Grubb RL Jr, Phelps ME, Ter-Pogossian MM: Regional cerebral blood volume in humans. X-ray fluorescence studies. Arch Neurol 28:18-44, 1973 6. Grubb RL Jr, Raichle ME, Eichling JO, et al: The effects of changes in Pa,,,, on cerebral blood volume, blood flow, and vascular mean transit time. Stroke 5:630-639, 1974 7. Hoffman EJ, Phelps ME, Mullani NA, et al: Design and performance characteristics of a whole body positron transaxial tomograph. J Nucl Med 17:493-502, 1976 8. Kuhl DE, Reivich M, Alavi A, et al: Local cerebral blood volume determined by three-dimensional reconstruction of radionuclide scan data. Circ Res 36:610-619, 1975 9. Mathew NT, Meyer JS, Bell RL, et al: Regional cerebral blood flow and blood volume measured with the gamma camera. Neuroradiology 4:133-140, 1972 10. Mathew NT, Meyer JS, Hartmann A, et al: Abnormal cerebrospinal fluid-blood flow dynamics. Arch Neurol 32:657664, 1975 Cerebral circulation stud11. Nylin G, Hedlund S, Regnstrom 0: ied with labeled red cells in healthy males. Circ Res 9666674, 1961 12. Olesen J: Cerebral blood flow, methods for measurement, regulation, effects of drugs and changes in disease. Acta Neurol Scand 5O:Suppl 57:l-134, 1974 13 Penn RD, Walser R, Ackerman L: Cerebral blood volume in man. Computer analysis of a computerized brain scan. JAMA 234:1154-1155, 1975 14 Phelps ME, Hoffman EJ, Coleman RE, et al: Tomographic images of blood pool and perfusion in brain and heart. J Nucl Med 17:603-612, 1976 15 Phelps ME, Hoffman EJ, Mullani NA, et al: Some performance and design characteristics of PETT 111, in TerPogossian MM, Phelps ME, Brownell GL, et a1 (eds): Reconstruction Tomography in Diagnostic Radiology and Nuclear
G r u b b e t al: Cerebral Blood V o l u m e in H u m a n s
327
16.
17.
18.
19.
Medicine. Baltimore, University Park Press, 1977, pp 371392 Phelps ME, Hoffman EJ, Mullani NA, et al: Application of annihilation coincidence detection to transaxial reconstruction tomography. J Nucl Med 16210-224, 1976 Phelps ME, Kuhl DE: Pitfalls in the measurement of cerebral blood volume with computed tomography. Radiology 121:375-377, 1976 Raichle ME, Welch MJ, Grubb RL Jr, et al: Measurement of regional substrate utilization rates using emission tomography. Science 199:986-987, 1978 Rosenblum WI: Can plasma skimming or inconstancy of re-
328 Annals of Neurology Vol 4 No 4 October 1978
20. 21.
22. 23.
gional hematocrit introduce serious error in regional cerebral blood flow measurements or their interpretation? Stroke 3~248-254, 1972 Straatmann MG, Welch MJ: A general method for labeling protein with "C. J Nucl Med 16:425-428, 1975 Ter-Pogossian MM, Phelps ME, Hoffman EJ, et al: A positron-emission transaxial tomograph for nuclear imaging (PE'IT). Radiology 114:89-98, 1975 Witelson SF, Pallie W: Left hemisphere specialization for language in the newborn. Brain 96:641-646, 1071 Wolf AP, Redvanly CS: Carbon-l 1 and radiopharmaceuticals. Int J Appl Radiat Isot 28:29-48, 1977
The technique of positron emission tomography was used to measure cerebral blood volume (CBV) in 10 normal right-handed human volunteers following inhalation of trace quantities of cyclotron-produced, "C-labeled carbon monoxide. In scans obtained 4 cm above the orbitomeatal line, CBV was 4.3 ml per 100 gm of tissue, whereas in scans obtained 8 cm above the orbitomeatal line, CBV was significantly less (3.3 ml per 1 0 0 gm; p < 0.001). This difference reflects the greater proportion of gray matter in the lower scan. Furthermore, the CBV was significantly larger Cp < 0.001) in the left cerebral hemisphere in the tomographic scans obtained 4 cm above the orbitomeatal line. These scans include the region of the superior surface of the temporal lobe (planum temporale), which is thought to be larger in individuals with left cerebral dominance for speech. This observation is the first in vivo demonstration of a structural correlate of a known functional difference in the cerebral hemispheres of man. Grubb RL Jr, Raichle ME, HigRins CS, et al: Measurement of regional cerebral blood volume by emission tomography. Ann Neurol 4:322-328, 1978 Present methods of studying brain hemodynamics (i.e., blood flow and blood volume) in vivo have serious shortcomings, particularly when applied to studies in humans [ 2 , 171. The appearance of linear accelerators and cyclotrons in the medical environment, the parallel development of rapid chemical techniques for incorporating their products (i.e., short-lived, positron-emitting radionuclides lSO,I3N, I1C, ISF) into compounds [23], plus recent developments in imaging systems employing the concept of positron emission tomography offer an opportunity to circumvent most of these shortcomings. Emission tomography is a nuclear medicine visualization technique that yields an image of the distribution of a previously administered radionuclide in any desired transverse section of the body. Positron emission tomography utilizes the unique properties of the annihilation radiation generated when positrons are absorbed in matter. It is characterized by the fact that an image reconstructed from the radioactive counting data is a highly faithful representation of the spatial distribution of a radionuclide in the chosen section. The resulting data make it possible to calculate values for parameters of physiological and biochemical significance when used with appropriate mathematical models. This approach is analogous to quantitative tissue autoradiography with the added advantage of allowing in vivo studies. In this report we describe the basis for a low-risk, quantitative, in vivo method of measuring cerebral
blood volume (CBV) regionally in the human brain. Our method employs emission tomography to monitor externally the spatial distribution in brain of carboxyhemoglobin labeled with the short-lived (20-minute half-life), cyclotron-produced, positronemitting radionuclide carbon 11. The data for our report were obtained from I 0 normal human volunteers.
From the Department of Neurology and Neurological Surgery and the Division of Radiation Sciences, The Edward Mallinckrodt Institute of Radiology, Washington University School of Medicine, St. Louis, MO.
Accepted for publication Mar 30, 1978.
Methods Cerebral blood volume was measured in 10 right-handed normal male volunteers. Informed consent for these studies was obtained in accordance with the regulations of the Washington University School of Medicine Human Use Committee. Emission tomographic scans of the head were performed following inhalation of "C-labeled carbon monoxide. A positron emission tomograph developed in our laboratory was used to quantitatively image the cross-sectional blood volume. The basic principles, design, and performance characteristics of the positron emission tomograph used in this study have been previously described [ 3 , 7, 14-16, 2 11.
The subjects for this study were in a resting normocapnic state. End-tidal carbon dioxide tension was monitored with a capnograph. An estimate of arterial carbon dioxide tension (P+O,) was calculated by multiplying the end-tidal carbon dioxide tension by atmospheric pressure with a correction for water vapor pressure. Each subject was positioned in the tomograph with the aid of a low-power laser beam of light across the center plane of the tomograph. By having the subject breathe approximately 10 to 15
Address reprint requests tO Dr Grubb, of Neurology and Neurological Surgery, Barnes Hospital Plaza, St. Louis, MO 631 10.
322 0364-5134/78/0004-0406S01.25 @ 1978 by Robert L. Grubb, Jr
mCi of high-specific-activity "C-labeled carbon monoxide contained in 800 fil of helium and 1,000 nil of room air from an Ambu bag, the blood was labeled with "Ccarboxyhemoglobin. Following inhalation of the gas, 6 to 7 minutes were allowed for the labeled blood to reach an equilibrium state. Two emission tomographic scans were then obtained sequentially 4 and 8 cm above the orbitomeatal (0-M) line. Data were collected for a period long enough to ensure sufficient counts in the image for an accurate image reconstruction (range, 370,000 to 1,000,000 counts; mean, 600,000 counts). This usually required 6 minutes for the first scan and 10 minutes for the subsequent one. Scans of the brain were computer processed in order to calculate a mean value of the scan for each region of interest. During each emission scan, several blood samples were drawn from a peripheral vein. These blood samples were weighed and counted in a 7.5 x 7.5 cm thallium-activated sodium iodide well detector to obtain the activity of the labeled blood (counts per second per gram of blood). A curve of the activity of "C-tagged carbon monoxide in the blood, corrected for the physical decay of "C from the beginning of the emission scan, was constructed as a function of time. The value for blood "C carbon monoxide activity used in the calculation of CBV (see the equation) was obtained from this curve by selecting the point in time corresponding to the midpoint of the emission scan. Calibration of the emission tomograph in order to determine the actual concentration of "C-labeled carbon monoxide in brain from the emission scan (counts per second per cubic centimeter of tissue), was performed by iniaging an appropriately designed phantom. This phantom, a cylinder 18 cm in diameter divided into five pie-shaped chambers of equal size, was filled with varying concentrations of "C-labeled bicarbonate in saline (typically a 1:2:3:4:5ratio). T h e phantom was then scanned twice. Because of the short half-life of " C (approximately 20 minutes), we obtained ten concentrations of radioactivity which bracketed the radioactivity concentration of "Clabeled carbon monoxide in the subject's scan. Weighted aliquots were then taken from each phantom chamber and counted in the same well detector used to count the blood samples. Scans of the phantom were computer processed in order to calculate a mean relative value of the scan for each chamber. A regression equation was then obtained for these data comparing the relative scan data and the directly measured activity from the phantom. This equation was used to determine the actual concentration of "C-labeled carbon monoxide in the subject's scan. Subject and phantom scan data from the emission tomograph were acquited to yield a spatial resolution of 15 mm full width at half maximum. Corrections for detector nonuniformity, physical decay of the radionuclide, and attenuation were made, and images were reconstructed using an Interdata Model 70 computer. T h e attenuation correction factors were computed using an experimentally determined value for the attenuation coefficient and the physical dimensions of the scanned object. Each data point is where x is the corrected for attenuation by the factor efiXx, thickness of the object along the coincidence line a n d F is the experimentally derived average linear attenuation
coefficient. The validity of this method has been previously reported [ 1 5 ] . The scan reconstruction algorithm was a convolution-based, filtered back projection that resulted in a 100 x 100 array representing the 48 cm diameter field; thus, each picture element size is 0.48 cm. Images were displayed and regions of interest sclected o n a 256 X 240 imaging system with 64 gray levels. Cerebral blood volume (milliliters per 100 g m of brain tissue) was calculated by the equation: CBV =
Cbr
Cb, * f *P h , ' P,
x 100
where Cbr is the concentration of "C-labeled carbon monoxide (counts per second per cubic centimeter of tissue) in the brain determined by the emission tomographic scan; Cbl is the concentration of "C-tagged carbon monoxide (counts per second per gram of blood) in the peripheral venous blood; f is the ratio of the mean cerebral small vessel hematocrit to large vessel hematocrit; Pblis the density of blood (gm per milliliter), and Pt is the density of brain tissue (gm per milliliter). An average value of 0.85 for the ratio of cerebral small vessel hematocrit to large vessel hematocrit was used in these studies [5]. A value of 1.05 g m per milliliter was used for both the density of blood and the density of brain tissue. The significance of differences between values of CBV obtained at 4 cm and 8 cm above the 0 - M line was tested by t tests, and paired t tests were used to test the significance of differences in CBV values obtained in the left and right cerebral hemispheres
Results Figures 1 and 2 show horizontal sections of the brain and "C carbon monoxide scans 4 and 8 cm, respectively, above the 0 - M line. The values of CBV and estimated P k o Z are contained in Table 1. In scans obtained 4 cm above the 0 - M line, an average CBV value of 4.3 (k0.4 SD) ml per 100 gm of tissue was found. At 8 cm above the 0 - M line the average value was 3.3 (k0.5 SD). The differences in CBV value at the two different levels scanned is significant ( p < 0.001). The mean estimated Paco2 in scans 4 cm above the 0 - M line was 4 1 (25 SD);8 cm above the 0 - M line the mean estimated Paco, was 38 ( - t 2 SD). The differences in Paco, were not significant. At 4 cm above the 0 - M line the CBV in the left cerebral hemisphere was 4.3 (k0.5 SD) ml per 100 gm of tissue while the CBV in the right cerebral hemisphere was 4.1 ( k 0 . 4 SD) ml per 100 gm.These left-right cerebral hemisphere differences in CBV 4 cm above the 0 - M line were significant ( p < 0.001). CBV was slightly greater in the left than in the right cerebral hemisphere 8 cm above the 0 - M line, but these differences were not significant. Figure 3 illustrates the typical area of a scan used to calculate whole-brain CBV. Figure 4 shows the typical areas of a scan used t o compute CBV for the left and right cerebral hemispheres.
Grubb et al: Cerebral Blood Volume in Humans 323
F i g 1 . (A) Horizontulsectiom of u huniun bruin. ( B , Emi.r.rivn t oniopcrphi(.srun after inhulution lif C-labrhd 1.urho in nionoxide. Brain slice and totnogruphic slice were both obtcrined 4 cni ubove the orbitomeutulline. The “hot.spot” in the posterior portion of the emission tomogruphir scavi vepre.sents the superior .wgittcrl.rinu.r.
’‘
Fzg 2. (A)Horizontalsection of a human brain. ( B )EmiJsion
tomographicscan afrer inhalation of ’%labeled t arbon nionoxide. Brain dice and tomographic slice were both obtained 8 cm about the orbitomru~dline.
324
Annals of Ncurology
Vol 4
No 4
October 1378
Table 1 . Cerebral Blood Volume und Estimated Arterial Carbon Dioxide Tension i n 10 Subjects
CBV, Whole Brain Subject
(m11100 gm tissue)
A Scan 4 cm above 0 - M line 1 4.0
2 3
4 5 6 7
8 Mean
4.1 5.0 4.8 4.8 4.7 4.4 4.2 4.1 4.2 4.3 4.0 4.0 3.8 3.6 4.3 (20.4 SD) (N = 15)
B Scan 8 cm above 0 - M line 8 3.1 2.9 9 3.1 3.1 10 4.0 3.8 Mean 3.3 (20.5 SD) ( N = 6)
CHV
CBV,Lcft Cerebral Hemisphere (m11100 gm tissue)
CBV,Right Cerebral (mV100 gm tissue)
(mm kg)
4.2 4.3 5.0 4.8 4.9 4.6 4.6 4.3
3.9 4.0 4.8 4.7 4.4 4.5 4.3 4.0 3.8 4.0 4.4 4.0 3.7 3.5 i .4 4.1 (20.4 SD) ( N = 15)
37 36 40 41 48 48 40 41
35 41 ( 2 5 SD) ( N = 9)
3.1 2.6 3.0 2.9 3.6 3.1 3.1 (?0.3 SD) ( N = 6)
36 39 37 36 41 40 38 ( ? 2 SD) ( N 6)
4.2
4.3 4.6 4.2 3.9 3.6 3.2 4.3 ( 2 0 . 5 SD) ( N = 15) 3.1 2.8 3.0 3.0
3.9 3.6 3.2 (20.4 SD) (N = 6)
Hemisphere
Estimated
Pace.
... ... ... ... ... ...
= cerebral blood volume.
Discussion Our results demonstrate the feasibility of quantitatively measuring regional CBV in vivo using positron emission tomography to image circulating "Clabeled carboxyhemoglobin. Because blood can be easily labeled by inhalation of trace quantities of the "C-labeled gas carbon monoxide, and quantitation requires only sampling of peripheral venous blood, this measurement is associated with minimal risk and discomfort to the subject. Furthermore, we consider the risk from exposure to ionizing radiation acceptable. For example, in a typical study in which the subject inhales 10 to 15 mCi of "C-labeled carbon monoxide, approximately 3 to 5 mCi actually labels the blood, and the remainder is quickly ventilated. Under these circumstances the blood receives approximately 50 mR per millicurie and the wholebody radiation dose is approximately 10 m R per millicurie. In addition to its ease of administration and the low radiation dose to the 'subject, our tracer, "C-labeled
carboxyhemoglobin, has the additional advantage of selectively labeling red blood cells. The use of labeled red cells to measure CBV circumvents the potential problem of extravascular migration of the tracer when damage has occurred to the blood-brain barrier. This is always a potential problem that cannot be anticipated when diseased tissue is studied. Data based on studies employing a plasma tracer must, as pointed out by others [2, 171, be viewed with caution when damage to the blood-brain barrier is suspected. The use of a red cell tracer-or, for that matter, a plasma tracer-for measurement of CBV necessitates knowledge of the local tissue hematocrit (see the equation). As pointed out by others, this may differ from the large vessel hematocrit. Thus, a correction must be inserted in the calculation of CBV for this discrepancy. In normal brain, the limited data currently available suggest a value of 0.85 for the ratio of small vessel or tissue hematocrit to large vessel hematocrit [51. Rosenblum [ 191 has cautioned that the local tissue hematocrit may change in acute dis-
Grubb
et
al: Cerebral Blood Volume in Humans 325
Fig 3. Dortedlinesurroundr typicalarea of "C carbon monoxide uitivity in emission tomographicscan used t o calculate wholebrain CBV.
Fig4. Dottedlinessurraundtypicalareasof "C carbon monoxide aitivity in emission tomngraphicsi-anused to calculate CBV i n left and right cerebral hemispheres.
eases of the brain such as ischemic infarction. Under such circumstances, errors in the computation of CBV couid be introduced by selection of an arbitrary value. In the future it should be possible to characterize the local tissue hematocrit more thoroughly in normal as well as diseased brain by sequential measurement of plasma and red cell volume in the same patient by emission tomography. This could be accomplished by the use of "C carbon monoxidelabeled red blood cells followed by intravenous administration of "C-labeled methylalbumin [20]. The precise measurement of CBV has proved to be difficult, and few studies in humans have been done [ 5 , 8-1 1 , 131. Table 2 lists values of CBV found in normal humans utilizing several different merhods. With the exceprion of one srudy [ I I ] employing phosphorus 32-labeled red blood cells, all these studies are in fairly close agreement. Three observations in our data are worthy of note. First, the average of CBV in normal humans was higher in emission tomographic scans obtained 4 cm above the 0 - M line compared with scans obtained 8
cm above this reference point. Although the mean estimated PacoZ was slightly higher in the subjects scanned 4 cm above the 0 - M line, this would not account for the differences in CBV seen at the two levels [ b ] .That different values of CBV were obtained at these two levels is not surprising, as different cross-sectional levels of the brain contain different ratios of gray to white matter. Gray matter is known to have a higher capillary density [l] and a higher resting cerebral blood flow [12]. As can be seen in Figures 1 and 2, emission tomographic scans obtained 4 cm above the 0 - M line contain a larger amount of deeper gray structures compared to scans 8 cm above this point. Second, in scans obtained 4 cm above the 0 - M line, the CBV was found to be significantly greater in the left cerebral hemisphere. This was, by history, the dominant hemisphere in all the subjects studied. Functional differences between the cerebral hemispheres of man have been noted for many years, but until recently, structural differences between the hemispheres have not been observed. However, in
326
Annals of Neurology
Vol 4
No 4
October 1078
Table 2 . Previously Determined Values far CBV in Normal Humans Average Value
of CBV (mli 100 Method
g m tissue)
Positron emission tomographic scan Radionuclide emission tomographic scan [81 X-ray computed tomographic scan
3.3-4.3 2-4
3
[I31 X-ray fluorescence [ 5 ] G a m m a camera mean transit time (gYmTc) [9, 101 Dye-dilution technique m e a n transit time (3LPP-labeledR B C s ) [ 1 I ]
3.2 4.05
9
1968, Geschwind and Levitsky [4]found the superior surface of the temporal lobe (planum temporale) to be larger in the left hemisphere in 65% of brains examined at autopsy. This observation was extended by Witelson and Pallie [22] in 1973 by their demonstration that this asymmetry is present in neonates as well as adults. Our finding of an increase in CBV in this area is further evidence of a structural asymmetry underlying functional differences between the two hemispheres. Our finding assumes special importance when it is realized that for the first time, such differences can be determined safely in vivo. We recognize the need for additional work in this area. Of particular interest would be the demonstration of an increase in regional CBV in the right hemisphere in patients known to have right cerebral dominance for language. Finally, our scans clearly reveal regional differences in "C activity, and hence CBV, within each cerebral hemisphere, corresponding to anticipated differences in CBV in gray and white matter structures. We are reluctant to quantitate these differences at this time because the emission tomograph used in these studies has a minimum resolution of 1.5 cm3. We do not consider this resolution sufficient t o allow quantitative independent determination of CBV regionally in gray and white matter. Anticipated advances in the design and resolution of emission tomographs, however, should make such independent measurements available in the near future. In summary, our studies demonstrate the feasihility of safely making quantitative in vivo measurements of regional CBV in humans. Such measurements should greatly facilitate studies designed to understand cerebral hemodynamics in the normal as well as diseased human central nervous system.
Furthermore, they should complement measurements of regional cerebral metabolism by emission tomography [ 181. Supported by US Public Health Service Grants HL 1385 1 and NSO 6833 (NINCDS) and by Teacher-Investigator Award NS 11059 from NINCDS (Dr Raichle). The authors wish to thank the staff of the Washington University School of Medicine cyclotron for their invaluable technical assistance in these experiments. We are grateful to Dr Michel M. TerPogossian for his generous support and helpful suggestions during the course of this work.
References I . Blinkov SM, Glezer 11: The Human Brain in Figures and Tables. New York, Plenum, 1968, p 267 2. Eichling JO, Gado MH, Grubb RL Jr. et al: Potential pitfalls in the measurement of regional cerebral blood volume, in Harper M, Jennett B, Miller D , et a1 (eds): Blood Flow and Metabolism in the Brain, Proceedings of the 7th International Symposium on Cerebral Blood Flow and Metabolism. Edinburgh and London, Churchill/Livingstone, 1975, pp 7.157.19 3. Eichling JO, Higgins CS, Ter-Pogossian MM: Determination of radionuclide concentrations with positron CT scanning (PETT). J NLKI Med 18:845-847, 1977 4. Geschwind N , Levitsky W: Human brain: left-right asymmetries in temporal speech region. Science 161: 186-1 8 7 , 1968 5 . Grubb RL Jr, Phelps ME, Ter-Pogossian MM: Regional cerebral blood volume in humans. X-ray fluorescence studies. Arch Neurol 28:18-44, 1973 6. Grubb RL Jr, Raichle ME, Eichling JO, et al: The effects of changes in Pa,,,, on cerebral blood volume, blood flow, and vascular mean transit time. Stroke 5:630-639, 1974 7. Hoffman EJ, Phelps ME, Mullani NA, et al: Design and performance characteristics of a whole body positron transaxial tomograph. J Nucl Med 17:493-502, 1976 8. Kuhl DE, Reivich M, Alavi A, et al: Local cerebral blood volume determined by three-dimensional reconstruction of radionuclide scan data. Circ Res 36:610-619, 1975 9. Mathew NT, Meyer JS, Bell RL, et al: Regional cerebral blood flow and blood volume measured with the gamma camera. Neuroradiology 4:133-140, 1972 10. Mathew NT, Meyer JS, Hartmann A, et al: Abnormal cerebrospinal fluid-blood flow dynamics. Arch Neurol 32:657664, 1975 Cerebral circulation stud11. Nylin G, Hedlund S, Regnstrom 0: ied with labeled red cells in healthy males. Circ Res 9666674, 1961 12. Olesen J: Cerebral blood flow, methods for measurement, regulation, effects of drugs and changes in disease. Acta Neurol Scand 5O:Suppl 57:l-134, 1974 13 Penn RD, Walser R, Ackerman L: Cerebral blood volume in man. Computer analysis of a computerized brain scan. JAMA 234:1154-1155, 1975 14 Phelps ME, Hoffman EJ, Coleman RE, et al: Tomographic images of blood pool and perfusion in brain and heart. J Nucl Med 17:603-612, 1976 15 Phelps ME, Hoffman EJ, Mullani NA, et al: Some performance and design characteristics of PETT 111, in TerPogossian MM, Phelps ME, Brownell GL, et a1 (eds): Reconstruction Tomography in Diagnostic Radiology and Nuclear
G r u b b e t al: Cerebral Blood V o l u m e in H u m a n s
327
16.
17.
18.
19.
Medicine. Baltimore, University Park Press, 1977, pp 371392 Phelps ME, Hoffman EJ, Mullani NA, et al: Application of annihilation coincidence detection to transaxial reconstruction tomography. J Nucl Med 16210-224, 1976 Phelps ME, Kuhl DE: Pitfalls in the measurement of cerebral blood volume with computed tomography. Radiology 121:375-377, 1976 Raichle ME, Welch MJ, Grubb RL Jr, et al: Measurement of regional substrate utilization rates using emission tomography. Science 199:986-987, 1978 Rosenblum WI: Can plasma skimming or inconstancy of re-
328 Annals of Neurology Vol 4 No 4 October 1978
20. 21.
22. 23.
gional hematocrit introduce serious error in regional cerebral blood flow measurements or their interpretation? Stroke 3~248-254, 1972 Straatmann MG, Welch MJ: A general method for labeling protein with "C. J Nucl Med 16:425-428, 1975 Ter-Pogossian MM, Phelps ME, Hoffman EJ, et al: A positron-emission transaxial tomograph for nuclear imaging (PE'IT). Radiology 114:89-98, 1975 Witelson SF, Pallie W: Left hemisphere specialization for language in the newborn. Brain 96:641-646, 1071 Wolf AP, Redvanly CS: Carbon-l 1 and radiopharmaceuticals. Int J Appl Radiat Isot 28:29-48, 1977
