UC San Diego
UC San Diego Previously Published Works
Title
18Fluorodeoxyglucose PET in schizophrenia.
Permalink
https://escholarship.org/uc/item/0mf0n7ps
Journal
Psychiatry research, 16(4)
ISSN
0165-1781
Authors
Jernigan, TL
Sargent, T
Pfefferbaum, A
et al.
Publication Date
1985-12-01
DOI
10.1016/0165-1781(85)90123-4
Peer reviewed
eScholarship.org Powered by the California Digital Library
University of California
Psychiatry Research, 16,317-329
Elsevier
IgFluorodeoxyglucose PET in Schizophrenia
Terry L
. Jernigan, Thornton Sargent III, Adolf Pfefferbaum,
Natasha Kusubov, and Stephen M
. Stahl
Received May 6, 1985
; revised version received August 9, 1985
; accepted September 5, 1985
.
Abstract
. Six chronic schizophrenic patients and six age-matched controls were
studied with 'sfluorodeoxyglucose ('sFDG) positron emission tomography (PET)
.
All patients were scanned when they had been free of medication for at least 2
weeks
. Comparisons were made between the groups on regional ratios of cortical
15FDG, with manual and automated measures
. Only one of eight regions, the right
temporal cortical region, showed a significant group difference, and this effect was
not significant when adjustment was made for multiple comparisons
. Secondary
analyses suggest that ventricular enlargement and age may be associated with a
relatively "hypofrontal" pattern of '8FDG
.
Key Words
. Positron emission tomography, schizophrenia, ventricular
enlargement
.
As early as 1974, Ingvar and Franzen presented results suggesting that cerebral blood
flow was regionally reduced in the frontal lobes of schizophrenic patients (Ingvar and
Franzen, 1974
; Franzen and Ingvar, 1975)
. A number of subsequent studies appear to
confirm these findings
. The method used in cerebral blood flow studies requires
intracarotid injection or breathing of xenon-133, and external detection of emitted
gamma rays with an array of collimated crystals . The resulting data provide mapping
of a two-dimensional projection of the activity distribution with spatial resolution not
better than 5 cm
. The recent development of computerized tomographic recon-
struction techniques has resulted in devices with the capability of accurate quantitative
imaging of radiation from planar sections of the human body
. In particular, positron
emission tomography (PET) uses a circular array of crystal detectors to measure the
coincident annihilation gamma rays from positron-emitting isotopes and to
reconstruct quantitatively accurate images of planar sections through the body
. The
presently most widely used PET technique uses deoxyglucose labeled with fluorine's
('sFDG), which has a half-life of 110 minutes
. This labeled molecule is taken up and
metabolized in tissue in a manner analogous to that for glucose, except that after
phosphorylation the next step is metabolically blocked, and the fixed regional
Terry L
. Jernigan, Ph
.D
., is Assistant Professor of Psychiatry and Radiology, University of California, San
Diego School of Medicine
. Dr
. Jernigan
is
also in the Psychology Service, Veterans Administration
Medical Center, San Diego
. Thornton Sargent
III,
Ph.D
., and Natasha Kusubov,
B
.S
.,
M
.T
.
(A
.S
.C .P
.),
are at Donner Laboratory, University of California, Berkeley
. Adolf Pfefferbaum,
M
.D
.,
is Associate
Professor of Psychiatry, and Stephen M
. Stahl, M
. D
., Ph
.D
., is Assistant Professor of Psychiatry, Stanford
University School of Medicine
. Drs
. Pfefferbaum and Stahl are also in the Schizophrenia Biological
Research Center, Veterans Administration Medical Center, Palo Alto, CA
. (Reprint requests to Dr
. T
. L
.
Jernigan, Psychology (I 16B), VAMC San Diego,
3350
La Jolla Village Dr
., San Diego, CA
92161,
USA
.)
0165-1781/85/S03
.30
0
1985 Elsevier Science Publishers B
.V
.
317
3
1
8
concentration of the isotope provides a relative measure
of
local glucose metabolism
.
A method by which the local cerebral' metabolic rate for glucose (I CMRgIc) may be
calculated has been presented by Phelps et al
. (1979) and Huang et al
. (1980)
. It
requires a measure of the arterial input function and
a
number of assumptions with
regard to rate constants
. The concentration of 18FDG in specific brain regions may
also be compared to the total brain orcortical uptake without such assumptions, and
this provides a valid measure of relative glucose metabolic rate in any chosen brain
region for that individual, which can then be compared to other individuals
.
Since the development of this technique, two groups have publishedd findings with
lluoro-deoxyglucose (FDG) in schizophrenic patients
. Farkas and his associates at
New York University and the Brookhaven Laboratory (Farkas et al
., 1984)
. and
Buchsbaum and associates at the National Institute of Mental Health (Buchsbaum ct
al
., 1982, 1984
; De Lisi et al
., 1985), have reported relatively lower frontal glucose
utilization in schizophrenics compared to controls
. In both cases, what is being
measured is frontal glucose metabolic rate relative to that inn the posterior' part of the
brain, or relative to the whole slice
. Overall glucose metabolic rate does not appear to
be different between schizophrenics and controls
.
These findings raise a number of questions about the nature
of this difference in
schizophrenics
. First, does the change represent a chronic alteration of
brain activity,
or does the hypofrontal pattern occur only when psychotic symptoms are present? In
previous studies with schizophrenics (except for Buchsbaum et al
., 1984), uptake
occurred during a resting state
. It is possible that the pattern observed
is
linked to
differences in the cognition of the two groups during the rest period, rather than to a
non-state-dependent pathophysiological difference
. Another question that
arises
concerns the relationship of changes in cortical glucose utilization to structural brain
changes, e
.g
.,
do changes in the volume of brain tissue affect the metabolic pattern?
The present study was conducted as an attempt to address some
of
these questions
.
The procedure was performed while the patients and controls responded to an
auditory vigilance task
. The inclusion of the vigilance task in the procedure represents
an attempt to control and monitor the cognitive behavior of the subjects
. We reasoned
that if group differences were observed in cortical glucose utilization in spite of similar
performance on the task, such differences might be interpreted as resultiog from
chronic brain alterations
.
Previous studies at the Palo Alto Veterans Administration Medical Center
(PAVAMC) suggest that structural abnormalities occur only very rarely in the
population of schizophrenics treated at this center (
.Jernigan et al
., 1982)
. Another
question we hoped to address was whether the previously reported metabolic
abnormalities were linked to structural abnormalities, and thus might also be
relatively rare in our patient population
.
Methods
Subjects
. Six male chronic schizophrenics, meeting Research Diagnostic Criteria (Spitzer et
al
., 1978), were recruited from the inpatient research unit of the PAVAMC
. (For aa clinical
description of these patients, see Table
I)
.
Controls were
six
male volunteers matched for age
who were recruited
from
among hospital employees of two university-affiliated hospitals and
from posted advertisements in the medical center
. Patients who were being treated with
neuroleptics were withdrawn from medication at least 2 weeks before imaging
. Three of the
patients were also imaged while on neuroleptic medication
: because of the small sample size,
however, those data are not presented
. During the last week before the imaging date, two
warm-up sessions were conducted in which the subjects were reminded of the details of the
imaging procedure, and then given several sets of the vigilance task
. An attempt was made to
simulate the imaging conditions
: the subjects reclined with eyes blindfolded as in the subsequent
imaging session
. After the task, subjects were asked questions about their experience of the task
:
whether they found it easy or difficult
;
if
their minds wandered
;
if
they were feeling calm,
anxious, or bored
; and if they were having intrusive thoughts or hallucinations
.
Table 1
. Clinical data on schizophrenic subjects
BPRS = Brief Psychiatric Rating Scale
;Overall and Gorham,
1962)
.
Auditory Vigilance Task
. The stimuli were tones of 800-ms duration with an interstimulus
interval of approximately 2 seconds
. The tones were either 400 Hz or 430 Hz, two easily
discriminated frequencies
. The higher-pitch tones occurred at random intervals and were 25%
.
of the total number of tones
. The subjects were asked to press a button whenever the higher-
pitch tones occurred
. The stimuli were presented via earphones that clipped onto the subjects'
ears
. They were generated by a microcomputer, and the subjects' responses were collected and
the scores computed in real time
. A printer produced a continuous log of the presentations and
responses for the operator's information, so that the task could easily be monitored during
imaging
. When communication with the subject necessitated diversion of the subjects' attention
from the task, this printed log was marked by the operator and the performance scores were
later corrected to exclude the effects of errors made during such interruptions
.
Imaging Procedure
. On the day the images were taken, subjects were asked not to drink coffee
and no medications were given
. No meals were taken within 3 hours before the injection
.
Subjects were transported by car from Palo Alto to the Donner Laboratory in Berkeley
. Two
i
.v
. catheters were inserted for blood drawing and injection
. A set of marks, originally made to
guide positioning of the earlier CT scans, was reconstructed on the subject's face to guide
selection of angle and level of the PET images
. In this way, we attempted to match the PET
images to selected CT sections
. The CT sections were taken at angles ranging from
approximately 0° to 20° relative to the cantho meatal line
. Variability in the angle was due to
attempts to avoid irradiation of the cornea
. The CT angle was repeated on PET to ensure that
structures visualized on CT were always obtained in the small number of
planar PET sections
available
.
The subjects were blindfolded and positioned in the imaging ring
. The angle was matched to
that of the face marks by adjusting the tilt angle of the ring and superimposing the scanner's light
line indicator onto the face marks
. The earphones were then attached
. The head position was
secured by application of a warmed moldable plastic face mask that was attached to pins on the
3
1
9
Subject
No
.
~
(years)
Age
Subtype
Age at
1st
episode
(years)
Age at
current
episode
(years)
Duration
of
current
episode
(years)
BPRS
score at
testing
1
31
Chronic residual
23
23
8
44
2
34
Chronic paranoid
27 27 7 37
3 34
Chronic undifferentiated
24
24 10
45
4 36
Chronic undifferentiated
19 19
17
26
5
41
Chronic undifferentiated
32 32 9
26
6 54
Chronic residual
20 50 4 30
32
0
underside of the
-
head
.-holder
. When cool, this mask became rigid and thus form-fitted
to
the
subject's face, reducing head mobility
.
The Donner 280-crystal dynamic positron tomograph uses a fixed array of bismuth
germanate crystals with a 50 cm patient port
.(Derenzo et
al
.,
1983) Dataacquisition,
reconstruction, imagedisplay,
.and kinetic analysis are performed with a PDP t l i44 computer
:
The resolution is 8 mm full width-half maximum and I
cm
axially
. The level of the tomograpl
is
image plane is set by manual or computer-controlled bed movement to preestablished positions
monitored by a digital positron counter
.
The
18
FDG
:was produced at Crocker Nuclear Laboratory at the University of California at
Davis
. The target and synthesis cave were developed by Dr
. Chester Mathis at Donner
.
Laboratory
.
Thirty to forty-five minutes before injection, several' transmission images, lasting 10 minutes
each, were made at the selected levels
. These images are made with an external hoop source of
68-Gallium and provide information about the distribution of attenuating material
(e
.
g
., skull
bone) within the section
. They are usedfor attenuation correction of the emission dataa later
obtained at the same levels
. The transmission images also aided in locating thee best sections for
subsequent imaging because the shape
of
the skull could be compared to that in the C"l' sections
.
While the transmission images were beingg taken,
a
warming pad was applied to the hand and
forearm from which "arterialized" blood samples were drawn
.
Five minutes before injection, the vigilance task was
begun
.
At injection, a dose ranging from
3 to I I mCi of
t
5
FDG was administered
. Blood samples were drawn from the warmed
: hand,
initially at 5-second intervals and then at progressively longer intervals, until 40 minutes after
injection
. Emission data were collected at similar intervals during the uptake period
. For
thee
present study, three or four postuptake emission images were obtained, each representing
55
minutes of data and the first beginning at 40 minutes after injection
. They were contiguous, but
not overlapping, axial sections, approximately
. I
cm
thick
. These images were corrected for
attenuation using the earlier transmission image data
. The data were later transferred to another
image-processing
: system for automated regional analysis
.
After the imaging session subjects were again
: asked questions about their cognitive and
affective state during the session
.
Image Selection and
.
Analysis
.
Because of a within-study revision of the procedurefor
selection of image levels, only one level was available in all 12 subjects
. This is the level at which
thalamus and striatum are visualized
.
The "raw" values were converted to regional ratios as described below
. Analysis of the glucose
metabolic rates will be reported elsewhere
.
The measurement of relative regional riFDG distribution was attempted by three separatee
techniques
: two manual techniques, and one semiautomated technique
. The rationale for using
these techniques is as follows
: the automated technique was an attempt to maximize the
objectivity and reproducibility of the measures, and to provide specifically defined regions for
those cases in which the cortical rim might be hypoactive, and thus visually difficult to define
reliably
. These advantages were considered likely, however, to be gained at thee expense of
increased sensitivity possible with manual measures
.
The two
. manual techniques both involved using a cursor to draw an irregularly-shaped region
of interest around visually identified cortical areas at the display console
. The first
:
manual
technique involved taking the values of small cortical regions in superior frontal, middle frontal,
superior temporal, middle temporal, inferior temporal, and occipital lobes
. This technique will
be referred to as the "manual regions' technique
. In theseeond manualtechmque
;afine through
the structural midline was drawn manually on a plastic overlay, followed by a line perpendicular
to this line at its midpoint
. Two additional lines drawn at 45° through the midpoint divided each
hemisphere into quadrants
. The cursor wasthen used
-
d to draw contiguous regions visuallyy
chosen to include the cortex within each quadrant
. This technique is called "manual quadrants
."
The manual methods are illustrated in Figs I and 2
. All manual measurements were made
without knowledge of subject identity
.