Dynamic monitoring of [11C]diprenorphine in rat brain using a prototype positron imaging device

Journal of Neuroscience Methods, 40 (1991) 223-232

223

© 1991 Elsevier Science Publishers B.V. All rights reserved 0165-0270/91/$03.50 NSM 01315

Dynamic monitoring of [11C]diprenorphine in rat brain using a prototype positron imaging device Suren Rajeswaran 1, Susan P. H u m e 1, Jill E. C r e m e r 1, John Young 2, Dale L. Bailey John A s h b u r n e r 1, Sajinder K. Luthra 1, A n t h o n y K.P. Jones 1 and Terry Jones 1

1

1MRC Cyclotron Unit, Hamrnersmith Hospital, London W12 OHS (U.K.) and 2 CTI Group Inc., Knoxville, TN37922 (U.S.A.) (Received 12 June 1991) (Revised version received 17 September 1991) (Accepted 20 September 1991)

Key words: Positron emission tomography; Small Animal; [11C]diprenorphine; Radioligands The present work tests the feasibility of using the most recently developed positron emission tomograph detector technology to image positron-emitting radioligands in small experimental animals. A prototype imaging device, using two opposing multicrystal, high-resolution ( ~ 4 ram) block detectors of bismuth germanate to produce a 2-dimensional image in the centre of the field of view, is described. To evaluate the probe's potential as a non-invasive experimental tool, the dynamic regional distribution of the established opiate receptor ligand, [llC]diprenorphine was determined in rat brain following intravenous injection. The distribution of counts in the images was consistent with the localisation of diprenorphine binding sites and the specificity of the signal obtained was confirmed by administration of non-radioactive diprenorphine and naloxone. Although the signal-to-noise ratio was reduced compared with data obtained by post mortem dissection, the dynamic data acquisition capabilities of the system demonstrate the feasibility of monitoring the kinetics of ligand binding in individual animals and encourages further design of a small-diameter detector system with tomographic capabilities.

Introduction

Positron emission tomography (PET) is now an established method for the in vivo measurement of the regional distributions of tracer compounds labelled with cyclotron-produced positron emitting radionuclides. Its ability to generate quantitative, dynamic, functional images provides unique insights into tissue physiology, biochemistry and drug kinetics. PET tracers currently in use in man enable the measurement of blood flow and metabolism and the characterisation of some neurotransmitter systems (Eriksson et al.,

Correspondence: Mr. S. Rajeswaran, MRC Cyclotron Unit, Hammersmith Hospital, Ducane Road, London W12 OHS, U.K. Tel.: 081-740-3162. Fax: 081-743-3987.

1990). The latter studies involve the use of specific and selective radiolabelled ligands whose dynamic properties and binding characteristics have been studied experimentally prior to their use in PET. Data-sets from rodents have generally been obtained post-mortem by measuring the radioactivity in brain regions at selected times after intravenous (i.v.) injection of the labelled compound. In addition to this being a time and animal intensive procedure, errors arise due to inter-animal variation. The in vivo determination of kinetic rate constants in small animals using external coincidence detection has been previously reported (Lockwood and Kenny, 1981; Tagaki et al., 1984; Nakai et al., 1987; Redies et al., 1987) but, to date, the method has used only a pair of single-crystal coincidence detectors. By using the latest genera-

224

tion of multicrystal, high-resolution PET detectors, the potential exists for the non-invasive acquisition of regional, differential tracer kinetics from individual animals. This paper describes data from a probe system which incorporates two such detectors, producing single-plane information. The initial aim of the study was to assess the system's capabilities in monitoring the uptake and retention of positronemitting ligands in small regions of rat brain by making use of prior knowledge of the biodistribution and binding kinetics of an established PET tracer, the opioid receptor ligand [ltC]diprenorphine (Jones et al., 1988; Frost et al., 1989). In the rat, diprenorphine binds specifically to the majority of brain regions except the cerebellum (Perry et al., 1980; Seeger et at., 1984), so that by 1 h after i.v. injection the ratio of counts in selected brain regions compared with cerebellum is of the order of ten. The cerebellum can therefore be used as a reference tissue for monitoring mainly non-specific binding (Cunningham et al., 1991).

Materials and methods

Detector system The dual probe system developed by CTI Inc., Knoxville, TN, U.S.A., uses 2 blocks of bismuth germanate (BGO), as currently used in PET detectors. Each BGO block is an 8 × 6 array of crystals which are 6.25 mm × 3.5 mm × 30 mm in size and separated by a gap of 0.5 mm. The intrinsic physical performance of the block in terms of its spatial resolution and efficiency, as well as the method of event positioning, has been reported previously (Digby et al., 1990). Data can be acquired in 1 of 4 modes: '1-, 3- or 8-nearest neighbours (NN)' or 'all coincidences'. The first 3 listed configurations produce a 2-dimensional (2D) 15 x 11 matrix of the activity distribution formed at the centre between the 2 detectors. Each picture element (pixel) is 3.3 mm × 2.0 mm in size. Accidental (or random) coincidences are corrected in real time by the delayed coincidence window technique (Hoffrnan et al., 1981). Data

TABLE I S P A T I A L R E S O L U T I O N S OF T H E P R O B E (full width at half maximum in mm) A L O N G T H E M I D L I N E OF B O T H AXES, F O R 1, 3 A N D 8-NN C O N F I G U R A T I O N S Detector axis (crystals)

Configuration (-NN)

8 8 8 6 6 6

1 3 8 1 3 8

Distance from centre of FOV (cm) 0

1

2

4.5 4.8 4.8 3.8 3.8 4.0

4.5 5.9 7.3 3.9 3.9 4.3

6.1 7.0 10.7 4.1 5.4 5.9

collection and matrix display are both under microcomputer control. A physical evaluation of the spatial resolution and sensitivity of each mode was made prior to the animal studies, at a block separation of 10 cm. From these results (shown in Tables I and II), the 3-NN mode, with a 350-850 keV energy window, was selected for the subsequent acquisition of experimental data, since it offered the best compromise between counting efficiency and spatial resolution and, in addition, provided a reasonable level of scattered 3,-ray rejection (these scattered events cause a degradation in image contrast).

Experimental studies Materials. The radioligands, 2-[t8F]fluoro-2 deoxy-D-glucose (FDG) (Hamacher et al., 1986) and [ttC]diprenorphine (sp.act. at injection: 5.416.3 TBq mmol -t) (Luthra et al., 1991) were prepared routinely at the Cyclotron Unit. 68Ge was purchased from Los Alamos National LaboT A B L E II RELATIVE EFFICIENCIES OF CONFIGURATIONS, S H O W I N G C O U N T S AS P E R C E N T A G E OF T H E 250 keV, 8-NN V A L U E Configuration(-NN)

Lower level energy discriminator (keV) * 250

350

450

1

16 62 100

15 55 88

10 38 59

3 8

* The upper level energy discriminator was set at 850 keV.

225

ratory, U.S.A. and a plane source (0.9 MBq in a solid gel volume of 30 ml) was subsequently prepared at the Cyclotron Unit. [15,1 6(N)-3H]di prenorphine (sp.act.: 1.2 TBq mmol - t ) was purchased from Amersham International plc, U.K. Naloxone chlorohydrate was purchased from Du Pont U.K. Ltd., and diprenorphine base was provided by Reckitt and Colman. Animals and methods. Male Sprague-Dawley rats (Harlan Olac Ltd.) weighing 250-260 g were anaesthetised by intraperitoneal injection of sodium pentobarbitone (Sagatal: May and Baker) at a dose of 60 mg/kg body weight and remained anaesthetised throughout all procedures. Following catheterisation of a tail vein, the rats were positioned on a perspex platform inside the bore of a lead collimator housing so that only the head projected into the centre of the field of view (FOV) of the detectors. The detectors were placed laterally (as shown in Fig. 1) such that the device produced a 2-D lateral image of radioactivity within the head. Prior to each dynamic scan, the position of the head in the image was defined by measuring the attenuation of ),-rays from a plane source of 68Ge, calculated from two 15-min, single-frame,

static scans performed without (blank scan) and with (transmission scan) the head in position. To delineate the brain within the head outline (in 2 rats only), the head was scanned (60 x 1-min frames) following an i.v. injection of 18.5 MBq [t8F]FDG and the counts for frames 30-60 min summated. Time-activity curves for the individual pixels within this outline were similar in shape to those previously reported by Redies et al. (1987) for a cylinder-shaped core of rat brain, in that they showed rapid accumulation of [18F]label over the initial 5-10 min after injection, followed by a plateau (data not shown). For each dynamic scan, 18.5 MBq [ltC]diprenorphine (in a volume of 0.5 ml saline) were injected via the tail vein over a 30-s period and the time course of radioactivity studied for 40 min using 1-min time frames. This injectate produced activity levels within the FOV which were within the linear count rate of the system, requiring no dead-time corrections. The scans were, however, normalised to correct for inefficiences in the edge detector crystals (see Data analysis). To compare directly the distribution of [tiC]radioactivity recorded using the positron-imaging device with the true regional localisation of

Fig. 1. Experimental arrangement for animal studies. The detectors are placed 10 cm apart on a sliding platform and the anaesthetised rat is positioned such that the head (from the back of the eyes to the back of the ears) projects into the FOV. A lead block shields the detectors from the remainder of the body radioactivity.

226 diprenorphine binding in the brain and surrounding head tissue, samples of thalamus, caudate putamen and cerebellum (together with the submaxillary and lachrymal glands) were taken for counting on a LKB Wallac y-counter. In addition, a single rat (200 g) was given an i.v. injection of [3H]diprenorphine (2.8 MBq in 0.30 ml sodium citrate:acetate buffer, pH 7.5). At 60 min after injection, the rat was killed with expiral and frozen in isopentane at - 4 0 °C. Whole-body, sagittal sections (20-/zm thick) were prepared for routine [3H]-autoradiography and conventional haematoxylin and eosin staining.

Data analysis Normalisation of the dynamic data was performed within the probe system software by multiplying the dynamic frames by pixel normalisation factors obtained from a 15-min 3-NN static scan of the 68Ge plane source, as the mean of the pixel counts divided by each pixel value. The dynamic scans were also corrected for attenuation, using factors obtained by dividing the blank scan by the transmission scan and each time frame was decay corrected back to time of injection, giving c o u n t s / s / p i x e l . Time-activity curves were then produced for each pixel within the delineated head region. The corrections and graphical display were exported to a SUN 3 / 6 0 workstation, using customised programmes, in order to convert the probe data into a format suitable for display using A N A L Y Z E image-processing software (Robb and Barillot, 1989). Summed images (collected over the 20-40-min time period after injection) were then constructed for the [llC]diprenorphine studies. During this latter half of the dynamic scan, any variations in radioactive content, due to regional blood flow rather than receptor concentration, were expected-to be minimal (Cunningham et al., 1991). Based on these summated images and also on the individual pixel time-activity curves, together with the ex vivo data, time-activity curves for pixel groups corresponding topographically to brain regions were produced. These groups were typically 4 - 6 pixels and are termed 'pixel regions of interest' or ROIs,

Results

Delineation of the brain in the matrix Fig. 2 diagramatically illustrates the reconstructed 2-D matrix obtained from the setup illustrated in Fig. 1. Within the 165 pixels, those which recorded counts primarily from the head were defined from the transmission scan as those having attenuation-correction factors greater than

® 0

@ Fig. 2. A 2-D image matrix showing the position of the head (light shading) and the 'brain' outline (dark shading). The pixels have been topographicallylabelled for future reference. The brain occupies rows 6-8 and columns A-K in the lateral image. The orientation is caudal/rostral = A/K, dorsal/ventral = 1/15.

227

unity (shown in light shading). The brain outline, as derived from the [18F]FDG distribution images, is superimposed as the darker area. As can be seen, the head occupied the centre of the FOV, i.e., pixel rows 5-11, and the brain was represented by the pixels in the upper half of the head outline. Fig. 3a illustrates a mid-sagittal section of rat head and neck, an orientation approximating to the 2-D projection of the volume scanned with the detectors in the lateral position. The outline of the scan FOV is indicated by the vertical line, confirming that the area occupied by

the brain approximates the upper half of the FOV.

Diprenorphine studies Fig. 3b is an ex vivo autoradiograph of the same section shown in Fig. 3a, obtained 60 min after an i.v. injection of [3H]diprenorphine. Within the brain, maximal binding was observed in thalamus, caudate putamen and brain stem, with some binding in the layers of cortex and hippocampus. No significant binding was seen in the cerebellum, giving a 'cold' region in the upper

Fig. 3. Mid-line, sagittal cryostat section (20 /~m) of a rat killed 60 min after i.v. injection of 2.8 MBq [3H]diprenorphine. The vertical lines in the haematoxylin- and eosin-stained section illustrated in (a) indicate the area included in the FOV of the detectors. Fig. 3b is an autofadiograph of the same section, exposed to 3H-hyperfilm (Amersham) for 13 weeks. The brain regions indicated are cerebellum (C) thalamus (T) caudate putamen (P) and spinal cord (S).

228 left portion of the FOV. Outside the brain, 'binding' can be seen in submaxillary and intraorbital lachrymal glands. Using the pixel labelling system shown in Fig. 2, brain regions defined anatomically both from the autoradiographs and from separately dissected brain slices, could be matched topographically to pixel groups or R O I s which recorded radioactivity from these regions. For example, counts from cerebellum would be recorded primarily in columns A - C , counts from thalamus in columns D - G and counts from striatum in columns H - J . Fig. 4a illustrates a summated image matrix, obtained over the period 20-40 min after i.v. injection of [llC]diprenorphine. High count re-

gions (hot spots) were seen in the 6-pixel group ( D - F , 6-7) and the 4-pixel group ( I - J , 7-8). Within the brain outline, a lower count region (cold spot) was observed in the 4-pixel group ( A - B , 6-7). The selected pixel groups were interpreted as reflecting, respectively, the brain regions thalamus, caudate putamen and cerebellum. Fig. 4b illustrates a similarly summed image matrix from a rat pre-dosed with naloxone 5 min prior to [llC]diprenorphine injection. Although the specific binding was displaced in 'brain', 2 remaining hot spots can be identified in regions corresponding anatomically to intraorbital and exorbital lachrymal glands. For comparison with the results obtained using the detector, the radioactive contents of selected

Fig. 4. Image matrices (detectors in lateral position, see Fig. 1) showing summed counts/pixel over the frames 21-40 min after i.v. injection of 18.5 MBq [llC]diprenorphine, given either (a) alone or (b) 5 rain after an i.v. injection of naloxone (1 mg/kg). As with Fig. 2, the nose is to the right-hand side of the images, the neck to the left; the brain signal is in rows 6-8, columns A-K. The scale covers the total range of counts recorded and has been arbitrarily divided into 12 equal portions.

229 TABLE III RADIOACTIVE CONTENT OF SELECTED REGIONS IN THE FOV OF THE DETECTORS, TAKEN POSTMORTEM 40 min AFTER i.v. INJECTION OF [IIC]DIPRENORPHINE Tissue samples were assayed using a y-counter. Uptake units are: (Bq/g tissue)/(injected Bq/g body weight). The values are from 5 animals/treatment. Dissected region

Uptake ( -+SD) Tracer

Thalamus Caudate putamen Cerebellum Submaxillary gland Exorbital lachrymal Intraorbital lachrymal Eyeball Masseter muscle

Naloxone pre-dosed

2.24:1:0.25 1.86 _+0.33 0.26 -+0.03 1.23-+0.17 1.61 + 0.29 2.63 _+0.26 0.26_+0.04 0.31 _+0.05

0.26 + 0.03 * 0.28 + 0.04 * 0.14 + 0.03 * 1.28_+0.11 2.23 _+0.29 * * 2.92 _+0.28 : 0.11 _+0.02 * 0.40_+0.14

* P < 0.001;

** P < 0.01 (Students's t-test).

brain and head regions, assayed by ex vivo counting 40 min after i.v. injection of [llC]diprenorphine, are given in Table III. Within the brain,

thalamus and caudate putamen showed high uptake relative to cerebellum. The latter had a radiOactive content similar to that in muscle adjacent to the skull. High counts were seen in the submaxillary gland and in the exorbital and intraorbital lachrymal glands. Whereas pre-treatment with naloxone resulted in a significant reduction in uptake in brain (the ratio of the mean uptake in thalamus compared with cerebellum was reduced from 8.6 to 1.9), it had either no significant effect on or caused an increase in radioactive content of the glands. Time-activity curves from selected pixel ROIs are shown in Fig. 5. The upper 3 curves are from the pixel groups representing thalamus, caudate putamen and cerebellum. Only the first 2 regions showed retention of radioactivity over the period of the experiment. Although the pixel group representing cerebellum showed a similar extraction of the radioligand, there was little retention of radioactivity with time. The lowest curve is from a pixel group in the centre of the lower half of the head outline (reflecting primarily counts from the soft palate and jaw).

15 14 13 12 )< E_

11 10 9 e 7 6 5

I

0

I

10

I

I

20

I

I

30

I

40

TIME(MIN)

Fig. 5. Attenuation-corrected counts as a function of time after i.v. injection of [llC]diprenorphine for 3 pixel groulSs (A-B, 6-8) = cerebellum (o) (E-F, 6-8) = thalamus ( + ) and (I-J, 7-8) = caudate-putamen ([]) in the upper half. of the head outline, defined from the corresponding transmission scan, and 1-pixel group (E-G, 10-11) (zx) in the lower half of the outline. The column and row annotation is as shown in Fig. 2.

230

The specificity of the signal was checked by comparing time-activity curves obtained for 3 experimental protocols, namely: (i) tracer alone, (ii) pulse-chase, in which non-radioactive diprenorphine was given 20 min after the injection of [ ~tC]diprenorphine, and (iii) pre-dosed, in which naloxone was given 5 min prior to the radioligand. The dose of cold diprenorphine in the tracer-alone protocol was of the order of 10 pmol/g body weight, having a negligible effect on [llC]diprenorphine uptake, whereas in the pulsechase protocol the dose of diprenorphine (470 pmgl/g body weight) was sufficient to saturate the binding sites (Cunningham et al., 1991). Fig. 6 illustrates the time-activity curves obtained for the pixel ROI (E-F, 6-8), representing thalamus. Data acquisition was sufficiently reproducible that, given the same injectate, the curves from different rats could be compared directly. The specific component of the [nC]binding obtained in the tracer-alone protocol (approximately half of the total counts recorded) was displaced by the subsequent administration of non-radioactive diprenorphine, in the pulse-chase experiment. Pre-treatment with naloxone, at a

dose sufficient to block in vivo binding of diprenorphine to opiate receptors (Seeger et al., 1984) resulted in time-activity curves which were similar in shape to that observed in the pixel group representing cerebellum (see Fig. 5) and characteristic of non-specific binding.

Discussion

From Table I, it can be seen that the measured spatial resolutions are close to the intrinsic resolution of the detectors (3.5 mm) (Digby et al., 1989, 1990) and optimal on the image plane, reducing with increasing distance from the centre of the FOV. The use of the 3-NN mode, balancing resolution with adequate sensitivity (Table II), allowed time-activity curves of good statistical quality to be obtained using comparatively small ROIs, while maintaining reasonable spatial resolution. (Anatomical structures which in size fall below the resolution of the system could still be imaged if the signal is of a sufficiently large magnitude).

15 1,4 13

12

J

x if_

b

11

N

lo

~ 8 u

9 8 7 6 5

I 0

I 10

I

1 20 TIME

I

i 30

i .40

(MIN)

Fig. 6. Attenuation-corrected counts in the selected pixel ROI (E-F, 6-8) (representing thalamus) as a function of time after injection of [nC]diprenorphine for: (i) tracer alone (+), (ii) pulse-chase (rn) and (iii) pre-dosed (c,) protocols. For (ii) diprenorphine (200 ng/kg) was given 20 min after the radioligand; for (iii) naloxone (1 mg/kg) was given 5 rain prior to the radioligand.

231 A comparison of the autoradiograph in Fig. 3 and the 2-D image matrix in Fig. 4 suggests that, with the detectors in the lateral position, the regional delineation of radioactivity within the 2-D image can b e interpreted as representative of the distribution of radiolabelled diprenorphine in the head. The differential pixel counts demonstrated hot spots in pixel groups corresponding topographically with thalamus and caudate putamen, with a cold spot in cerebellum. In addition, hot spots outside the brain, but within the FOV of the detectors, were identified. These corresponded to high concentrations of radioactivity measured in exorbital and intraorbital lachrymal glands, unrelated t o opiate receptor distribution. The effect of the radioactivity in the exorbital glands on the 'brain' pixel counts can be minimised by scanning with the detectors in the lateral position since these glands lie bilaterally in the lower half of the head outline. However, the radioactivity in the intraorbital glands contributes to the counts recorded in the 'fore-brain' pixels because of the juxtaposition of the glands to the skull and their relatively large mass compared with the individual brain regions. The influence of the intraorbital lachrymal glands is apparent in pixel columns J-K, in the summed image shown in Fig. 4b, and must be accounted for in the interpretation of data derived from fore-brain regions, including rostral neocortex. By using the dynamic data acquisition capabilities of the system, time-activity curves were obtained for selected regions in individual rats. These were similar in shape to those obtained by dissection of brain regions (1 rat/time point), as reported for [3H]diprenorphine by Cunningham et al. (1991). However, although the specificity of the signal could be verified by pre-dosing with naloxone and by pulse-chase with non-radioactive diprenorphine, the ratio of total compared with non-specific (cerebellum) counts was Very much reduced compared with that determined from ex vivo data. This was presumably due to an additional contribution to each image pixel from non-brain tissue; an unavoidable consequence of using only 2 BGO blocks, in a fixed position. Although the contamination of regional data negates the calculation of absolute values for

kinetic parameters, it is possible to obtain estimates of 'binding potential' using a reference tissue compartmental model (Inoue et al., 1991). In such a way, regional changes in signal following any pharmacological intervention can be quantified. The results from the probe, although limited by single-plane information, demonstrate that its spatial resolution is such that such devices can be used to facilitate the study of the kinetic properties of established or potential PET ligands in small animals, complementing more conventional ex vivo techniques. In addition, the results encourage further development of the current system, to one which has a tomographic capacity, by arranging a number of these high-resolution block detectors in a ring, smaller in diameter than previously used for animal scanners (Burnham et al., 1984; Tomitani et al., 1985; Yamashita et al., 1990). This would greatly increase the sensitivity and provide the ability to obtain information from multiple planes through the tissue or organ of interest. An additional feature of the proposed tomograph (Rajeswaran et al., 1991) is the absence of lead septa. This allows more coincidences to be collected and maximises the sensitivity for a given ring diameter, enabling 3-D volumetric images to be acquired. The resultant axial resolution should be sufficient to allow all regions of brain to be distinguished from extra-cerebral tissue. By recording the dynamic distributions of positron-emitting ligands in individual animals, these probes offer the possibility of performing more complex studies than are practical in multianimal experiments. In addition to facilitating PET ligand development, the effects of drugs on the in vivo pharmacokinetics of radiolabelled compounds could be studied with greater reproducibility and the more important issue of in vivo specificity might be feasibly addressed with simplified experimental design.

Acknowledgments The prototype probe system was supplied by CTI Inc., Knoxville, TN as part of an ongoing

232

collaborative program between CTI and the MRC Cyclotron Unit. The experimental set-up was constructed by the Medical Engineering Department of the MRC CU and the conversion of the probe data to a format suitable for a SUN workstation was done with the help of Jon Heather. We are grateful to Malcolm Seville and the Histology Department of the MRC Toxicology Unit, Carshalton, U.K., for their help with the ex vivo tritium autoradiography and for producing the whole-body cryostat section illustrated in Fig. 3. S. Rajeswaran is supported by a MRC studentship.

References Burnham, C.A., Bradshaw, J., Kaufman, D., Chesler, D. and Brownell, G.L. (1984) A stationary positron emission ring tomograph using BGO detector and analog readout, IEEE Trans. Nucl. Sci., 1984, 31: 632-636. Cunningham, V.J., Hume, S.P., Price, G.R., Ahier, R.G., Cremer, J.E., Jones, A.K.P. (1991) Compartmental analysis of diprenorphine binding to opiate receptors in vivo and its comparision with equilibrium data in vitro, J. Cereb. Blood Flow Metab., 11: 1-9. Digby, W.M., Hoffman, E.J., Phelps, M.E. et al. (1989) A multi-plane high-resolution PET system for animal studies, J. Cereb. Blood Flow Metab., 9 (Suppl. 1): $236.1. Digby, W.M., Dahlbom, M. and Hoffman, E.J. (1990) Detector shielding and geometric design factors for a high-resolution PET system, IEEE Trans. Nucl. Sci., 37: 664-670. Eriksson, L., Dahlbom, M. and Wid6n, L. (1990) Positron emission tomography - a new technique for the studies of the central nervous system, J. Microscopy, 157: 305-333. Frost, J.J., Mayberg~ H.S., Meltzer, C.C., et al. (1989) Measurement of [11Ckarfentanil and [11Cldiprenorphine binding to opiate receptors in temporal lobe epilepsy by PET, J. Cereb. Blood Flow Metab., 9 (Suppl. 1): $60. Hamacher, K., Coenen, H.H. and Stocklin, G. (1986) Efficient stereospecific synthesis of no-carrier-added 2-[lSF]-fluoro 2-deoxy-D-glucose using aminopolyether supported nucleophitic substitution, J. Nucl. Med., 27: 235-238. Hoffman, E.J., Huang S.-C., Phelps, M.E., Kuhl, D.E. (1981) Quantitation in positron emission computed tomography. 4. Effect of accidental coincidences, J. Comput. Assist. Tomogr., 3: 391-400. Inoue, O., Kobayashi, K., Tsukada, H., Itoh, T. and Langstrom,

B. (1991) Difference in in vivo receptor binding between [3H]N-methyl-spiperone and [3H]raclopride in reserpinetreated mouse brain, J. Neural Trans., 85: 1-10. Jones, A.K.P., Luthra, S.K., Maziere, B., et al. (1988) Regional cerebral opioid receptor studies with [11C]diprenorphine in normal volunteers, J. Neurosci. Methods, 23: 121-129. Lockwood, A,H. and Kenny, P.J. (1981) Serial measurements of positron-emitting isotope activity in rat brain, Stroke, 12: 173-176. Luthra, S.K., Turton, D.R., Dowsett, K., Bateman, D.M., Kensett, M.J., Waters, S.L., Pike, V.W. (1991) Improved and automated 'one-pot' radiosynthesis of [llCkliprenorphine and [IIC]buprenorphine, J. Label. Compd. Radiopharm., 30: 258-259. Nakai, H., Matsuda, H., Takara, E., et al. (1987) Simultaneous in vivo measurement of lumped constants and rate constants in experimental cerebral ischemia using F-18 FDG, Stroke, 18: 158-167. Perry, D.C., Mullis, K.B., Pie, S. and Sadee, W. (1980) Opiate antagonist receptor binding in vivo evidence for a new receptor binding model, Brain Res., 199: 49-61. Rajeswaran, S., Bailey, D.L, Hume, S.P., et al. (1991) 2D and 3D imaging of small animals and the human radial artery with a high resolution detector for, PET, IEEE Trans. Med. Imaging, in press. Redies, C., Matsuda, H., Diksic, M., Meyer, E. and Yamamoto, Y.L. (1987) In vivo measurement of [lSF]fluorodeoxyglucose rate constants in rat brain by external coincidence counting, Neuroscience, 22: 593-599. Robb, R.A. and Barillot, C. (1989) Interactive display and analysis of 3-D medical images, IEEE Trans. Med. Imaging, 8: 217-226. Seeger, T.F., Sforzo, G.A., Pert, C.B. and Pert, A. (1984) In vivo autoradiography: visualization of stress-induced changes in opiate receptor occupancy in the rat brain, Brain Res., 305: 303-311. Takagi, S., Ehara, K., Kenny, P.J., Finn, R.D., Kothari, P.J. and Gilson, A.J. (1984) Measurement of cerebral blood flow in the rat with intravenous injection of [ltC]butanol by external coincidence counting: a repeatable and noninvasive method in the brain, J. Cereb. Blood Flow Metab., 4: 275-283. Tomitani, T., Nohara, H., Murayama, H., Yamamoto, M. and Tanaka, E. (1985) Development of a high resolution positron CT for animal studies, IEEE Trans. Nucl. Sci., 32: 822-825. Yamashita, T., Watanabe, M., Shimizu, K. and Uehida, H. (1990) High resolution block detectors for PET, IEEE Trans. Nuct. Sci., 37: 589-593.