- Email: [email protected]
Cerebral HMPAO spect in patients with major depression and healthy volunteers
Pmg. NetmH%ychopharmacoL
& BioL Psychic& Copyright
1996. Vol. 2.0. pp. 443-458 Q 1996 Elsevler All
ELSEVIER
PII
Science
0278 - 5846/96
SO278-5846(90)00008-5
Inc.
rightsreserved $32.00
CEREBRAL HMPAO SPECT IN PATIENTS WITH MAJOR DEPRESSION AND BEALTHY VOLUNTEERS
P. DAVID MOZLEY’ , MADYHORNIG-ROHAN*,ANNA MARIEWODA’ , HEE-JOUNG KIM’, ABASS ALAVI’, FRANZPAYER’,and JAYD. AMSTERDAM’ ‘Division of Nuclear Medicine, and *The Depression Research Unit, University of Pennsylvania Medical Center, Philadelphia, PA, IJ.S.A. (Final form, January 1396) Abstract Mozley David P., Mady Hornig-Rohan, Anna Marie Woda, Hee-Joung Kim, Abass Alavi, Franz Payer and Jay D. Amsterdam: Cerebral HMPAO SPECT in Patients with Major Depression and Healthy Volunteers. Prog. Neuro-Psychopharmacol. & Biol Psychiat. 1996,a 1. The authors examined the regional cerebral distribution of [Tc-99m] HMPACI using single photon emission computed tomography (SPECT) in patients with major depression and in healthy controls. 2. 19 patients and 16 healthy controls had SPECT images of the brain acquired with 740 MBq (20 mCi) of [Tc-99m] HMPAOon a triple-headed camera equipped with fan beam collimators. 3. Mean counts per pixel were measured in 13 regions of each hemisphere and compared to the mean activity in the whole brain, the ipsilateral hemisphere, and cerebellum. A “laterality score” was calculated for each structure by subtracting the mean counts per pixel in a region of the right hemisphere from the mean counts in the homotopic region of the left hemisphere and normalizing the difference by the average in both regions. The degree of hemispheric asymmetry was calculated from the absolute values of the laterality scores. 4. The distribution of HMPAOwas more variable in patients than in controls; while the mean activity ratios were not significantly different in any region. Asymmetries between homotopic regions of the limbic system were more pronounced in patients than in controls. However, there were no consistent left-to-right asymmetries in either group. 5. The present data indicate that regional cerebral distribution of HMPAOmay not be discretely abnormal in depression, but demonstrates heightened variability in depressives (vs. control subjects). Keywords:
brain SPECTimaging; depression: Tc-39 labeled HMPAO
Abbreviations:
dorsolateral
prefrontal
443
cortex
(DLPFC);
LgF-labeled
444
P.D. Mozley et al
fluorodeoxyglucose (FDG); hexamethylpropyleneamine oxime (HMPAO); I- 123 labeled iofetamine (IMP); laterality score (IS); major depressive disorder (MDD); positron emission tomography (PET); regional cerebral blood flow (rCBF); regions of interest (ROIs); single photon emission computed tomography (SPECT); Structured Clinical Interview for DSM-IIIR (SCID) Introduction Several radioactive blood flow tracers have been used to study regional cerebral blood flow (rCBF) in depression with single photon emission computed tomography (SPECT) (Alavi & Hirsch 1991; Holman & Devous 1992). Some investigators have reported a decrease in rCBF perfusion in the whole brain or some of its regions (Silfverskiold & Risberg 1989; Sackeim et al. 1990; Austin et al. 1992; Mayberg et al. 1994), while others have reported regional or global increases in rCBF in depression (Amsterdam & Mozley 1992; George et al. 1993; Ebert et al. 1991). In contrast, others have not replicated these observations (Mathew et al 1980; Gur et al 1984; Maes et al 1993). Several factors might contribute to the apparent discrepancy in neuroimaging observations in depression. For example, diagnostic heterogeneity within patient samples, and differences in illness severity have been cited (Baxter et al. 1992; Holman & Devous 1992; Bench et al. 1993). Others have suggested differences in instrumentation and the pharmacologic characteristics of various radioligands might contribute to variations in results (El1 et al. 1985; Anderson et al. 1988; Anderson 1989; Sokoloff 1992). Furthermore, these factors may be compounded by the effects of the disease process itself on cerebral perfusion, and could conceivably produce an increase in regional brain activity with one tracer and a decreased activity with another tracer. In the present study, the authors examined rCBF using [Tc-99m] labeled hexamethylpropyleneamine oxime (HMPAO) in carefully characterized patients with unipolar, major depression and compared these results to those from healthy non-psychiatric controls. Methods Patients Overall, 19 patients from the Depression Research Unit at the University of Pennsylvania Medical Center were studied: 13 men and 6 women with a mean (+SD)age of 38210 years (range 24-59 years). Eighteen were right handed and 1 was left handed. All had moderate to severe MDD with a mean Hamilton Depression Rating Scale (Hamilton 1960) score of 22.923.8 on the 17-item scale.
Cerebral HMPAO spect in patients with major depression
445
The minimum duration of depressive illness was six weeks with a mean duration of 63+87 weeks. Diagnostic assessments were made using the Structured Clinical Interview for DSM-III-R (SCID) (Spitzer et al. 1989) and each patient met criteria for unipolar, MDD (American Psychiatric Association 1987), single or recurrent episode. Patients with other DSM III-R Axis I diagnoses, including bipolar (manic depressive) disorder, were excluded. Similarly, patients demonstrating evidence of severe Axis II character pathology were also excluded. None of the patients had a history of drug or alcohol dependence. None of the patients had a meaningful, co-existing medical or surgical problem that could have affected the biodistribution of the HMPAO tracer. Moreover, no subject had any clinical evidence of a peripheral vascular, cardiovascular, or cerebrovascular disease. Physical and neurological examinations, along with KG and chest radiographs were within normal limits. No meaningful laboratory abnormalities were present in the complete blood count plus differential count, serum electrolytes, or in liver, kidney and thyroid function tests. All subjects were drug free for at least two weeks prior to the imaging procedure, and none had received fluoxetine within twelve weeks or electroconvulsive therapy within five years of the scan. Controls The control group consisted of 16 healthy, non-psychiatric subjects. All were recruited from the University and greater Philadelphia metropolitan area. Nine were men and 7 women with a mean age of 39511 years (range 24-62 years). All were fully employed and without any history of DSM III-R Axis I or I I diagnosis. The results of their semi-structured psychiatric histories, physical examinations, and laboratory screening studies were normal. None were on any medications other than oral contraceptives. Thirteen were right-handed, and 3 were left handed (with signs of mixed dominance). Image
Acquisition
The procedure began by intravenously administering 740 MBq (20 mCi) of Tc99m labeled HMPAO (Amersham Inc.) while the subjects sat quietly with their eyes open and ears unoccluded. Images of the brain were acquired 60 minutes later on a triple headed gamma camera equipped with fan beam collimators (Picker 3000, Cleveland, OH). The intrinsic resolution of the camera was rated as 8-9 mm FWHM (full width half maximum). A stop and shoot mode was used to obtain projection data for up to 1 hour. Acquisition parameters included a 13.5
446
P.D. Mozley et al
cm center of rotation, and a 128x128 matrix with a pixel width of 2.11mm in the projection domain. All images were reconstructed with the same procedure to produce cubic voxels with a side length of 3.56 mm. The projection images were reconstructed with a count-rate dependent Wiener filter (King et al. 1984) that was applied to partially correct for septal penetration, scatter, and finite detector resolution. The modulation transfer function applied by the filter was determined experimentally from the line spread function (Gilland et al. 1988; Press et al. 1988; Kim et al, 1392) for the camera that was used to scan the patients (Kim et al. 1993, 1995). A uniform ellipse was placed around each slice in order to apply Chang’s first order correction method for photon attenuation (Chang 1978). The images were re-aligned in all three planes as necessary in order to be resliced in transaxial sections that were approximately parallel to the plane containing the anterior and posterior commissure line. The orientation was actually chosen by fitting a line on the mid-sagittal section that ran through the apogees of the frontal and occipital poles. The coronal and sagittal slices were then cut in orthogonal planes. The images were then imported into an image analysis package (Resnick et al. 1993), where a “central punch biopsy” approach was used to draw regions of interest (ROIs) around 13 structures in each hemisphere. The technique required each ROI to be smaller than the structure it was placed on, and omitted the ROI from the first or last axial slice on which a structure could be visualized. This tended to minimize the effects of volume averaging around the edges and enhance the precision of measurement, particularly in structures without clearly visualizable boundaries in all directions. The whole supratentorial brain activity was measured on 3 contiguous mid sagittal sections. The coronal sections were used to measure the activity in the frontal and temporal poles. The right and left frontal poles were defined as the region anterior to a coronal plane containing the anterior most aspects of the temporal tips. The inferior aspect of the anterior cingulate gyrus was defined as a region beginning 7 mm below the inferior border of its body on the transaxial plane and continuing inferiorly for 3 slices. The dorsolateral prefrontal cortex (DLPFC) was defined as a 10.5 mm thick section of the lateral most aspect of the frontal cortex midway between the hemispheric fissure and the anterior most temporal lobe beginning 7 mm above the superior boundary of the body of the caudate in the transaxial plane, The ROIs for the temporal poles were placed on the 4 most anterior coronal sections beginning with the second slice on which they could be visualized. The basal ganglia were separated into the head of the caudate beginning below its body, and a lenticular region consisting of a strip between the internal and external capsules. The thalamic ROI was placed on 3 transaxial slices in the center of the structure. The activity
Cerebral HMPAO spect in patients with major depression
447
in the hippocampal-amygdaloid formation was measured on both the transaxial Only the anterior aspect of the and the coronal images for comparison. hippocampus was included. Data Analysis The mean counts per pixel in each region were measured and compared to the mean activity per pixel in the whole brain, the remaining hemisphere, and the cerebellum. Possible hemispheric differences between patients and controls in specific ROIs were examined using loge transformed and untransformed data with parametric unpaired Student t tests and the nonparametric Mann-Whitney U test, respectively. Pearson and Spearman correlation coefficients were then used to characterize the presence of any correlations between the demographic variables and specific measurements of HMPAOregional brain activity. A “laterality score” (Gur et al. 1991) was calculated for each ROI by taking the mean counts per pixel in each selected ROI of the right hemisphere and subtracting this from the mean counts per pixel in the homotopic region of the left hemisphere. The difference between the two hemispheres was then divided by the average activity in both regions. Positive “laterality score” (LS) values indicated that activity in the left-sided ROI relative to the corresponding rightsided ROI. LS values were compared using one-tailed, unpaired t-tests. The proportional differences in positive LS values between patients and controls were then compared by using chi-square analyses with Yates’ correction. The absolute value of the LS score for each region was also used in the analyses without regard to which side of the brain showed more activity. This was used as a measure of “degree of asymmetry” or variability between the two hemispheres. The differences between patients and controls were compared with one-tailed, unpaired t-tests. Results Visual
Inspection
of the Imapes
A review of the images did not reveal any occult brain disease. We obserVed mild asymmetry with decreased radioactivity in several regions of the cortex and diencephalon. However, no specific, consistent abnormalities in HMPAO activity were observed on any of the sagittal, coronal or transaxial images (Fig 1). Relative
HMPAO Activity
in Specific
Computer based, semi-quantitative
Brain Regions
image analysis of the SPECT data revealed
448
P.D. Mozley et al.
HMPAO activity to be highest in the diencephalic regions of patients and controls. The caudate/cortex ratio of mean counts per pixel was 1.81+0.33 in the patients (range 1.19 to 2.45) and 1.76fl.41 in the controls (range 1.27 to 2.33) (p=ns). Similarly, no differences were found between patients and controls in HMPAO activity in the putamen and thalamus, nor in putamen/cortex or thalamus/cortex ratios (Table 1A and 1B). Table 1A Diencephalic Cdudate Patients Controls Difference
to Whole Brain
1.81 F 0.39 1.76 f. 0.41 p = 0.74
Region to Whole Brain Ratios Lenticular
to Whole Brain
1.77 + 0.40 1.71 + 0.42 p = 0.66
Thalamus
to Whole Brain
1.78 + 0.45 1.75 ‘t 0.46 p = 0.84
Table 1B Relative Activity of HMPAO in the Caudate to
Caudate to Cerebellum
Caudate Frontal
1.81 + 0.39 1.76+0.41 p = 0.74
1.30 + 0.28 1.3 1 + 0.30 p = 0.93
1.68 + 0.42 1.77 + 0.45 p = 0.52
Caudate Cortex Patients Controls Indep. t-test
to Pole
Caudate to Temporal Pole 1.75 + 0.32 1.80 LO.59 p = 0.73
No significant differences were observed between subject groups in the ratio of HMPAOactivity in any other ROI to the activity in the cerebellum, the ipsilateral hemisphere, the frontal cortex, or the whole brain (Table 2). The supratentorial cortex to cerebellar ratios defined on 3 contiguous mid-sagittal slices were similar in patients and controls: 0.72fl.05 and 0.75fl.06, respectively (p=ns). Examining regions of the right and left hemisphere independently did not produce any specific findings. Correlations
with Ape and Gender
No significant interactions were observed between the mean HMPAO activity ratios in any brain region and age or gender in this sample using Pearson and Spear-man correlation coefficients.
Cerebral
HMPAO spect in patients
with major depression
449
Fig 1: images of a 37-year old right handed man with depression who committed suicide less than a weeh after study. The regionally specific findings that have been reported in other studies of depression are not apparent. The findings are typical in that many of the most severely affected patients in this sample had completely unremarkable examinations that in clinical practice would have been read as “definitely not abnormal.”
450
P.D. Mozleyet al Table 2 Relative
Region Frontal Frontal DLPFC DLPFC Anterior
HMPAO Activity Patients
Pole to Whole Brain Pole L5 score to Whole Brain LS Score Cingulate
Laterali
1.81 0.01 3.28 -0.01 1.75
in Cortical
Regions
Controls
+ 7.9 LO.03 + 0.82 + 0.12 i 0.51
1.76 F 0.00 + 3.34 * -0.05 + 1 .G7 +
Probability 8.6 0.03 0.90 0.25 0.41
p p p p p
= = = = =
0.82 0.30 0.83 0.55 0.59
tv
Left-to-right hemispheric asymmetry in all ROIs was just as common as rightto-left asymmetry in both subject groups (Table 3). The range of LS values in various diencephalic regions was highly variable in both patients and controls. 68% (13/13) of the patients showed asymmetry in at least one of the three diencephalic regions that was >7% of the mean activity. similarly, 56% (9/16) of controls also demonstrated an asymmetry. However, while 26% (5/19) of the patients showed more than one area of asymmetric rCBF in the dfencephalon, only one control (6%) had two regional asymmetries. Asymmetries between the left and right caudate ranged from 0.81% to 14.4% in patients, and from 0.0% to 15.8% in controls. There were no consistent diencephalic rCBF asymmetries within either subject group, or between groups. Table 3 Diencephalic
Lenticular
Caudate Patients Controls Difference
Laterality
0.89 i 7.90 1.38 + 6.66 p = 0.86
0.53 + 5.42 -0.04 + 5.49 p = 0.76
Scores Thalamus 2.54 + 10.11 -0.52 L 5.65 p = 0.29
Similarly , there were no significant differences between patients and controls in the presence of hemispheric rCBF asymmetries for any of the following ROIs:
frontal cortex, temporal cortex (as measured by coronal and transaxial images), hippocampus/amygdala, or dorsolateral prefrontal cortex. Degree
of Temooral
Asymmetry
in Patients
vs. Controls
The absolute value of the LS index (degree of asymmetry) was specifically examined in the anterior temporal lobes because of prior observations of
Cerebral
HMPAO spect
in patients
with major
451
depression
temporal lobe asymmetry with [l-123] IMP SPECT (Amsterdam and Mozley 1992). Interestingly, with HMPAO there was no apparent hemispheric laterality in temporal lobes when patients were compared to controls. While the patients did demonstrate a greater overall degree of asymmetry compared to controls (6.83-&30 vs. 3.4722.21; p=O.OS, unpaired t-test), the direction of right-left asymmetry was not consistent in either group, and there was a substantial overlap between groups (Table 4). As noted above, there were no statistically significant differences between the patient and control groups in the ratio
1.02+0.11 1.01 + 0.08
Difference
p = 0.91
Activity Temporal LS Score
of HMPAO in Temporal Pole
HippocampalIAmygdala to Whole Brain
1.10 + 9.97 -0.61 + 4.17 p=O.Sl
1.38 + 0.36 1.39 + 0.39 p = 0.93
Regions Hippocampal/Amygdala LS Score -0.31 i 6.91 0.19 + 8.09 p = 0.85
Discussion Observations from the present HMPAO study do not support prior reports of specific rCBF abnormalities in patients with MDD. However, the authors did observe greater variability in regional HMPAO activity in depressed patients compared to control subjects. However, there were no other findings which differentiated the patients from the controls. While some investigators have reported specific abnormalities in regional brain activity during depression (Sackeim et al. 1990; Amsterdam & Mozley 1992; Austin et al. 1992; Ebert et al. 1991; George et al. 1993; Mayberg et al. 1994), Maes et al. (1993) failed to detect any abnormalities in the distribution of HMPAO in depressed patients compal’ed to controls. Clinical Correlates
of SPECT FindinPs
Several clinical and technical reasons might explain the divergent observations in various SPECT studies in depression. For example, some studies reporting discrete rCBF abnormalities with HMPAO have included older natients who tiav
452
P.D. Mozley et al.
normally show more frequent rCBF disturbances (Philpot et al. 1993). Investigations with Xenon-133 SPECT suggest that there may be an interaction of age with depressive subtype for both global and rCBF (Devous et al. 1991). One of the authors previously found a significant effect of normal aging on the distribution of HMPAOin healthy subjects (Payer et al. 1994). In contrast, in the present study the patients were relatively young and without any significant effects of age on HMPAOactivity. Additionally, with Kumar et al. (1991), one of us previously reported a decreased cortical to cerebellar ratio in elderly depressed patients using IMP SPECT. In contrast, in the present study, only a slightly higher cortical to cerebellar ratio in the patients (vs. controls) was observed. The presence of depressive psychotic and/or melancholic features has been associated with several SPECT abnormalities (Austin et al. 1992; Ebert et al. 1991). In contrast, while some of our subjects did have melancholic symptoms, none had overt psychotic features. This clinical difference may have partly accounted for our failure to observe discrete SPECTabnormalities in our patients. Severity of depression has also been found to positively correlate with changes in rCBF (Schlegel et al. 1989). In contrast, we were unable to demonstrate any correlation between HDRS scores and regional HMPAO activity. Interestingly, in a prior study using IMP SPECT, we observed right temporal lobe asymmetry in depressed patients compared to medically ill controls (Amsterdam & Mozley 1992). Our failure to observe specific regional blood flow asymmetries in the present study using HMPAO suggests that the physical and pharmacologic properties of these tracers may result in different patterns of rCBF in depression. For example, HMPAOmay not be as sensitive to relatively small changes in brain function as other tracers. A high degree of concordance has been reported between HMPAO and other blood flow tracers in pathological processes that However, several independent groups of produce neuronal necrosis. investigators have now reported that HMPAO cannot detect small reductions in local blood flow that might be demonstrable with other agents (Heiss et al. 1990; Toyama et al. 1994; Pointon et al. 1994). Others have observed that the regional activity of HMPAO can actually be increased in areas of ischemia (Sperling & Lassen 1993). It is, therefore, possible that the present observations may reflect a relative insensitivity of HMPAO as a tracer for identifying discrete, functional abnormalities in MDD. Comparison
with Other
Imaging
Techniques
Positron emission tomography (PET) has been able to detect abnormalities in similar groups of patients (Buchsbaum et al. 1984; Baxter et al. 1985; Post et al.
Cerebral HMPAO spect in patients with major depression
453
1987;
Cohen et al. 1989). Cerebral blood flow and metabolism are frequently coupled (Sokoloff 1992). However, several unique biological features of HMPAO may limit its comparison with PET studies of depression. Metabolic tracers such as 18F-labeled fluorodeoxyglucose (FDG) tend to localize within the synaptic bouton, where the work of neural transmission is actually performed (Kadekaro et al. 1985). In contrast, HMPAO tends to localize in the cytoplasm of the neuronal cell body (Anderson et al. 1988) where only vegetative maintenance functions are being performed. The cell bodies are frequently located at some distance from their synaptic terminals, and have significantly less surface area to take up a radiotracer than their axonal processes. This may make HMPAO less biologically sensitive in from some functional disorders that produce reversible abnormalities of perisynaptic neurotransmission than FDG PET (Mielke et al. 1994). Thus, while HMPAO may a sensitive agent for detecting abnormalities characterized by neuronal necrosis, the present study suggests that it is less sensitive as a tracer for identifying the more subtle abnormalities of functional psychiatric disorders. The relatively enhanced resolution in our current study should have made it more likely for a true abnormality in HMPAO activity to be observed. While we have previously found that the restorative filter is more accurate than conventional low pass filters, the risks of using a count rate dependent filter usually include producing false positive results, not false negatives. The present image analysis technique has previously been validated and shown to be a reliable one. In the present study, image analysis benefited from the enhanced edge detection produced by the filter. However, it is important to note the caveat of a lack of co-registration with anatomical images (e.g., magnetic resonance imaging). However, this is not likely to have been a significant factor in identifying abnormalities in ROIs such as the head of the caudate and the temporal poles. The image thresholding technique in the present study enhanced precision and compensated for anatomical accuracy in these regions. Conclusions Visual inspection of HMPAO SPECT images of the brain in the depressed patients did not demonstrate any specific regional abnormalities. Moreover, computer-enhanced, semi-quantitative analysis failed to reveal any more subtle abnormalities in HMPAO activity that were not observed on visual inspection. The present results suggest that HMPAO may not identify discrete abnormalities in regional brain dysfunction in patients with major depression. Future investigations using diverse SPECTradioligands in the same individual may help to resolve the conflicting observations presently seen in brain imaging studies.
454
P.D. Mozley et aL
Acknowledgement This work was supported by an educational grant-in-aid from Eli Lilly and Company Central Research, NIMH Neuropsychopharmacology Fellowship PHS Grant MH 14654, and the Jack Warsaw Fund for Research in Biological Psychiatry. The technical support staffs in the Depression Research Unit and the Division of Nuclear Medicine made substantial contributions to the performance of this study. References ALAVI,A. and HIRSCH,L.J. (1991) Studies of central nervous system disorders with single photon emission computed tomography and positron emission tomography: evolution over the past 2 decades. Sem. Nucl. Med. 2: 58-81. AMERICANPSYCHIATRICASSOCIATION( 1987) Diagnostic and Statistical Manual of Mental Disorders, 3rd ed., revised, American Psychiatric Association, Washington. AMSTERDAM,J.D. and MOZLEY,P.D. (1992) Temporal lobe asymmetry with tofetamine (IMP) SPECT imaging in patients with major depression. J. Affective Disord. 24: 43-53. ANDERSON,A.R., FRIBERG,H.H., SCHMIDT,J.F. and HASSELBALCH, S.G. ( 1988) Quantitative measurements of cerebral blood flow using SPECTand [99mTc]d,l-HMPAO compared to Xenon-133. J. Cerebral Blood Flow Metab. 8: S69-S81. ANDERSON,AR. (1989) 99mTc-d,l-Hexamethylene-propyleneamine oxime (99mTc-HMPAO): Basic kinetic studies of a tracer of cerebral blood flow. Cerebrovasc. Brain Metab. Rev. 1: 288-3 18. AUSTIN,M.-P., DOUGALL,N., ROSS,M., MURRAY,C., O’CARROLL,R.E.,MOFFOOT,A., EBMEIER,K.P. and GOODWIN,G.M. (1992) Single photon emission computed tomography with 99mTc-exametazime in major depression and the pattern of brain activity underlying the psychotic/neurotic continuum. J. Affective Disord. 26: 31-44. BAXTER,L.R., PHELPS,M.E., MAZZIO’ITAJ.C., SCHWARTZ,J.M., GERNER,R.H., SELIN, C.E. and SUMIDA,R.M. (1985) Cerebral metabolic rates for glucose in mood disorders. Arch. Gen. Psychiatry 112:441-447. BAXTER,L.R.,GUZE,B.H.,SCHWARTZ,J.M., PHELPS,M.E.,MAZZIOTTAJ.C. and SZUBA,M.P. (1992) PET studies of cerebral function in major depression and related disorders. In: Brain Work and Mental Activity: Quantitative Studies
Cerebral HMPAO spect in patients with major depression
with Radioactive Tracers, N.A. Lassen, D.H. Ingvar, M.E. Raichle, H.H. Friberg (Eds.), pp 52-65, Munksgaard, Copenhagen. BENCH,C.J., FRISTON,K.J., BROWN,R.G.,FRACKOWLAK, R.S.J. and DOLAN,R.J. ( 1993) Regional cerebral blood flow in depression measured by positron emission tomography: the relationship with clinical dimensions. Psychol. Med. 23: 579-590. BUCHBAUM,M.S., DELISI,L.E., HOLCOMB,H.H., CAPPELLETTI,J., KING,A.C., JOHNSON,J., HAZLE’M’, E., DOWLING-ZIMMERMAN, S., POST, R.M., MORIHISA,J., CARPENTER,W., COHEN,R., PICKAR,D., WEINBERGER, D.R., MARGOLIN,R. and KESSLER,R.M. (1984) Anteroposterior gradients in cerebral glucose use in schizophrenia and affective disorders. Arch. Gen. Psychiatry 41: 1153-l 166. CHANG,L.T. (1978) A method for attenuation correction in radionuclide computed tomography. IEEETrans. Nucl. Sci. 25: 638- 643. COHEN,R.M.,SEMPLE,W.E., GROSS,M., NORDAHL,T.E., KING,AC., PUICKAR,D. and POST, R.M. (1989) Evidence for common alterations in cerebral glucose metabolism in major affective disorders and schizophrenia. Neuropsychopharmacol. 2: 241-254. DEVOUS,M.D. Sr., GULLION,C., GRANNERMAN,B. and RUSH,A.J. ( 1991) Regional cerebral blood flow alterations in unipolar depression. J. Nucl. Med. 32: 95 l952. EBERT,D., FEISTEL,H. and BAROCKA,A. (1991) Effects of sleep deprivation on the limbic system and the frontal lobes in affective disorders: a study with Tc99m-HMPAO SPECT. Psychiatry Res. Neuroimaging 40: 247-251. ELL,P.J., HOCKNELL,J.M.L., JARRI’I-F,P.H., CULLUM,I., LUI, D., CAMPOS-COSTA,D., NOWOTNIK,D.P., PICKETT,R.D., CANNING,L.R.and NEIRINCKX,R.D. ( 1985) A Tc99m-labelled radiotracer for the investigation of cerebral vascular disease. Nucl. Med. Commun. 4: 437-441. GEORGE,M.S., KEITER, T.A. and POST, R.M. (1993) SPECTand PET imaging in mood disorders. J. Clin. Psychiatry 54(suppl): 6-13. GILLAND,D.V., TSUI, B.M.W.,MCCARTNEY,W.H., PERRY,J.R. and BERG,J. (1988) Determination of the optimal filter function for SPECTimaging. J. Nucl. Med. 25): 643650. GUR,R.E.,SKOLNICK,B.E.,GUR,R.C.,CAROFF,S., RIEGER,W., OBRIST,W.D., YOUNKIN,D. and REMCH, M. (1984) Brain function in psychiatric disorders. II: Regional cerebral blood flow in medicated unipolar depression. Arch. Gen.
455
456
P.D. Mozley et al.
Psychiatry 41: 695-699. CUR,R.C., MOZLEY,P.D., GUR,R.E.,RESNICK,S.M.,GOTTLIEB,G.E.,KOHN,M., ZIMMERMAN,R., HERMAN,G., ATLAS,S., GROSSMAN,R., BARRE’ITA,D., ERWIN, R. and GUR, R.E. (1991) Gender differences in age effect on brain atrophy measured by magnetic resonance imaging. Proc. Natl. Acad. Sci. 88: 2845-2849. HAMILTON,M. ( 1960) A rating scale for depression. Psychiatry 23: 56-62.
J Neurol Neurosurg
HEISS,W.D., HERHOLZ,K., PODREKA,I., NEUAUER,I. and PIETRZYK,U. ( 1990) Comparison of [99mTc]HMPAO SPETwith [ 18Flfluoromethane PET in cerebrovascular disease. J. Cereb. Blood Flow Metab. 10: 687-697. HOLMES,B.L.and DEVOUS,M.D. (1992) Functional brain SPECT:the emergence of a powerful clinical method. J. Nucl. Med. 33: 1888-1904. KADEKARO,M., CRANE,A.M. and SOKOLOFFL (1985) Differential effects of electrical stimulation of sciatic nerve on metabolic activity in spinal cord and dorsal root ganglion in the rat. Proc. Natl. Acad. Sci. 73:4230-4234. KIM, H.J., ZEEBERG, B.R.,FAHN,
F.H., HOFFMAN,E.J. and REBA,R.C. (1992) 3-D SPECTsimulations of a complex 3-D mathematical brain model: Effects of 3-D geometric detector response, attenuation, scatter, and statistical noise. IEEE Trans. Med. Imag. 11: 176-184.
KIM,H.J., KARP,J.S., KUNG,J.S. and MOZLEY,P.D. (1993) Quantitative effects of a count rate dependent Wiener filter on image quality: a basal ganglia phantom study simulating [I-123] dynamic SPECTimaging. J. Nucl. Med. 34: 190 (abstract). KIM H.J., KARP,J.S., KUNG,J.S., MOZLN, P.D. and ALAVI,A. ( 1995) Collimator and filter selection: simulating technetium-99m cerebral perfusion studies with a three-dimensional Hoffman brain phantom. J. Nucl. Med. 34: 195 (abstract). KING, M.A.,
SCHWINGER,R.B.,DOHERTY,P.W. and PENNY,B.C. ( 1984) The twodimensional filtering of SPECTimages using the Metz and Wiener filters. J. Nucl. Med. 25: 1234-1240.
KUMAR, A., MOZLN,
P.D., VELCHIK,M.G., DUNHAM,C., REILLN, J., GOTTLIEB,G. and AIAVI, A. (1991) State dependent decrease in cortical to cerebellar IMP activity in patients with late onset depression. Internat J Geriatric Psychiatry 4: 775-777.
MAES,M., DIERCK,R., MELTZER,H.Y.,INGELS,M., SCHOTTE,C., VANDEWOUDE,M.,
Cerebral HMPAOspect in patients with major depression
CALABRESE, J. and COSYNS,P. (1993) Regional cerebral blood flow in unipolar depression measured with Tc-99m-HMPAO single photon emission computed tomography: negative findings. Psychiatry Res. Neuroimaging 30: 77-88. MATTHEW,R.J., MEYER,J.S., FRANCIS,D.J., SEMCHUK,K.M.,MORTEL,K. and CLAGHORN,J.I. (1980) Cerebral blood flow in depression. Am. J. Psychiatry 137: 1449-1450. MAYBERG,H.S., LEWIS,P.J., REGENOLD,W. and WAGNER,H.N. JR. ( 1994) Paralimbic hypoperfusion in unipolar depression. J. Nucl. Med. -35. 929-934. MIELKE,R., PIETRZYK,U., JACOBS,A., FINK, G.R., ICHIMIYA,A., KESSLER,J., HERHOLZ,K. and HEISS,W.D. (1994) HMPAOSPETand FDG PET in Alzheimer’s disease and vascular dementia: comparison of perfusion and metabolic pattern. Eur. J. Nucl. Med. 21, 1052-1060. PAYER,F., MOZLEY,P.D. and ALAVI,A. (1994) Age related changes in the cerebral distribution of HMPAO. J. NucI. Med. 35: 1007P (abstract). PHILPOT,M.P., BANERJEE,S., NEEDHAM-BENNETT, H., COSTA,DC. and ELL,P.J. (1993) 99mTc-HMPAO single photon emission tomography in late life depression: a pilot study of regional cerebral blood flow at rest and during a verbal fluency task. J. Affective Disord. 28: 233-240. POINTON,J., SINGH,P., EGAN,G., CHAN,G., TOCHON-DANGU,H., NEWTON,M.R., BERLANGIERI, S.U., O’KEEFE,G., MCKAY,W.J. and DONNAN,G.A. (1994) PET and SPECTimaging in acute stroke: perfusion patterns and clinical outcome. Eur. J. NucI. Med a(supp1): S23. POST,R.M., DELISI,L.E.,HOLCOMB,H.H,, UHDE,T.W., COHEN,R and BUCHSBAUM, M.S. (1987) Glucose utilization in the temporal cortex of affectively ill patients: positron emission tomography. Biol. Psychiatry 22: 545553. PRESS,W.H., FLANNERY,B.P., TEUKOLSKY,S.A. and VETTERLING,W.T. (1988) Numerical Recipes in C: The Art of Scientific Computing, Cambridge University Press, Cambridge, pp 407-412. RESNICK,S.M., KARP,J.S., TRETSKY,B.I. and GUR,R.E. ( 1993) Comparison of anatomically defined versus physiologically based regional localization: Effects on PET-FDGquantitation. J. Nucl. Med. 34: 201-208. SACKEIM,H.A., PROHOVNIK,I., MUELLER,J.R., BROWN,R.P., APTER,S., PRUDIC,J., DEVANAND,D.P. and MUKHERJEE,S. (1990) Regional cerebral blood flow in mood disorders: I. Comparison of major depressives and normal controls at re$t. Arch. Gen. Psychiatry 47: 60-70.
457
458
P.D. Mozley et aL
SCHLEGEL,S., ALDENHOFF,J.B., EISSNER,D., LINDNER,P. and NICKEL,0. ( 1989) Regional cerebral blood flow in depression: association with psychopathology. Affective Disord. 17: 211-218.
J.
SILFVERSKIOLD, P. and RISBERG,J. (1389) Regional cerebral blood flow in depression and mania. Arch. Gen. Psychiatry 46: 253-259. SOKOLOFF,L. (1992) Relationship between functional activity and energy metabolism in the nervous system: Whether, where, and why? In: Brain Work and Mental Activity: Ouantitative Studies WithRadioactive Tracers, N.A. Lassen, D.H. Ingvar, M.E. Raichle, H.H. Friberg (Eds.), pp 52-65, Munksgaard, Copenhagen. SPERLING,B. and LASSEN,N.A. (1933) Hyperfixation of HMPAOin sub-acute ischemic stroke leading to spuriously high estimates of cerebral blood flow by SPECT. Stroke 24: 133-l 94. SPITZER,R.L., WILLIAMS,J.B.W., GIBBON,M. and FIRST, M.B. ( 1989) Instruction Manual for the Structured Clinical Interview for DSM-IIIR (SCID, 5/l/89 Revision), Biometrics Dept., NYState Psychiatric Institute, New York. TOYAMA,H., MAEDA,H., TAKESHITA,G., SHIBATA,K., NOMURA,M., KATO,M., SAKAKIBARA,E., TAKEUCHI,A. and KOGA,S. (1994) Tc-99m HM-PA0 super dynamic SPECT to detect the early cerebral blood flow distribution. Eur J. Nucl. Med. 2l(suppl): S4. Inquiries and reprint requests should be addressed to: Jay D. Amsterdam, M.D. Depression Research Unit University of Pennsylvania Medical Center IJniversity Science Center, 8th floor 3600 Market Street Philadelphia, PA 19 104 U.S.A.
& BioL Psychic& Copyright
1996. Vol. 2.0. pp. 443-458 Q 1996 Elsevler All
ELSEVIER
PII
Science
0278 - 5846/96
SO278-5846(90)00008-5
Inc.
rightsreserved $32.00
CEREBRAL HMPAO SPECT IN PATIENTS WITH MAJOR DEPRESSION AND BEALTHY VOLUNTEERS
P. DAVID MOZLEY’ , MADYHORNIG-ROHAN*,ANNA MARIEWODA’ , HEE-JOUNG KIM’, ABASS ALAVI’, FRANZPAYER’,and JAYD. AMSTERDAM’ ‘Division of Nuclear Medicine, and *The Depression Research Unit, University of Pennsylvania Medical Center, Philadelphia, PA, IJ.S.A. (Final form, January 1396) Abstract Mozley David P., Mady Hornig-Rohan, Anna Marie Woda, Hee-Joung Kim, Abass Alavi, Franz Payer and Jay D. Amsterdam: Cerebral HMPAO SPECT in Patients with Major Depression and Healthy Volunteers. Prog. Neuro-Psychopharmacol. & Biol Psychiat. 1996,a 1. The authors examined the regional cerebral distribution of [Tc-99m] HMPACI using single photon emission computed tomography (SPECT) in patients with major depression and in healthy controls. 2. 19 patients and 16 healthy controls had SPECT images of the brain acquired with 740 MBq (20 mCi) of [Tc-99m] HMPAOon a triple-headed camera equipped with fan beam collimators. 3. Mean counts per pixel were measured in 13 regions of each hemisphere and compared to the mean activity in the whole brain, the ipsilateral hemisphere, and cerebellum. A “laterality score” was calculated for each structure by subtracting the mean counts per pixel in a region of the right hemisphere from the mean counts in the homotopic region of the left hemisphere and normalizing the difference by the average in both regions. The degree of hemispheric asymmetry was calculated from the absolute values of the laterality scores. 4. The distribution of HMPAOwas more variable in patients than in controls; while the mean activity ratios were not significantly different in any region. Asymmetries between homotopic regions of the limbic system were more pronounced in patients than in controls. However, there were no consistent left-to-right asymmetries in either group. 5. The present data indicate that regional cerebral distribution of HMPAOmay not be discretely abnormal in depression, but demonstrates heightened variability in depressives (vs. control subjects). Keywords:
brain SPECTimaging; depression: Tc-39 labeled HMPAO
Abbreviations:
dorsolateral
prefrontal
443
cortex
(DLPFC);
LgF-labeled
444
P.D. Mozley et al
fluorodeoxyglucose (FDG); hexamethylpropyleneamine oxime (HMPAO); I- 123 labeled iofetamine (IMP); laterality score (IS); major depressive disorder (MDD); positron emission tomography (PET); regional cerebral blood flow (rCBF); regions of interest (ROIs); single photon emission computed tomography (SPECT); Structured Clinical Interview for DSM-IIIR (SCID) Introduction Several radioactive blood flow tracers have been used to study regional cerebral blood flow (rCBF) in depression with single photon emission computed tomography (SPECT) (Alavi & Hirsch 1991; Holman & Devous 1992). Some investigators have reported a decrease in rCBF perfusion in the whole brain or some of its regions (Silfverskiold & Risberg 1989; Sackeim et al. 1990; Austin et al. 1992; Mayberg et al. 1994), while others have reported regional or global increases in rCBF in depression (Amsterdam & Mozley 1992; George et al. 1993; Ebert et al. 1991). In contrast, others have not replicated these observations (Mathew et al 1980; Gur et al 1984; Maes et al 1993). Several factors might contribute to the apparent discrepancy in neuroimaging observations in depression. For example, diagnostic heterogeneity within patient samples, and differences in illness severity have been cited (Baxter et al. 1992; Holman & Devous 1992; Bench et al. 1993). Others have suggested differences in instrumentation and the pharmacologic characteristics of various radioligands might contribute to variations in results (El1 et al. 1985; Anderson et al. 1988; Anderson 1989; Sokoloff 1992). Furthermore, these factors may be compounded by the effects of the disease process itself on cerebral perfusion, and could conceivably produce an increase in regional brain activity with one tracer and a decreased activity with another tracer. In the present study, the authors examined rCBF using [Tc-99m] labeled hexamethylpropyleneamine oxime (HMPAO) in carefully characterized patients with unipolar, major depression and compared these results to those from healthy non-psychiatric controls. Methods Patients Overall, 19 patients from the Depression Research Unit at the University of Pennsylvania Medical Center were studied: 13 men and 6 women with a mean (+SD)age of 38210 years (range 24-59 years). Eighteen were right handed and 1 was left handed. All had moderate to severe MDD with a mean Hamilton Depression Rating Scale (Hamilton 1960) score of 22.923.8 on the 17-item scale.
Cerebral HMPAO spect in patients with major depression
445
The minimum duration of depressive illness was six weeks with a mean duration of 63+87 weeks. Diagnostic assessments were made using the Structured Clinical Interview for DSM-III-R (SCID) (Spitzer et al. 1989) and each patient met criteria for unipolar, MDD (American Psychiatric Association 1987), single or recurrent episode. Patients with other DSM III-R Axis I diagnoses, including bipolar (manic depressive) disorder, were excluded. Similarly, patients demonstrating evidence of severe Axis II character pathology were also excluded. None of the patients had a history of drug or alcohol dependence. None of the patients had a meaningful, co-existing medical or surgical problem that could have affected the biodistribution of the HMPAO tracer. Moreover, no subject had any clinical evidence of a peripheral vascular, cardiovascular, or cerebrovascular disease. Physical and neurological examinations, along with KG and chest radiographs were within normal limits. No meaningful laboratory abnormalities were present in the complete blood count plus differential count, serum electrolytes, or in liver, kidney and thyroid function tests. All subjects were drug free for at least two weeks prior to the imaging procedure, and none had received fluoxetine within twelve weeks or electroconvulsive therapy within five years of the scan. Controls The control group consisted of 16 healthy, non-psychiatric subjects. All were recruited from the University and greater Philadelphia metropolitan area. Nine were men and 7 women with a mean age of 39511 years (range 24-62 years). All were fully employed and without any history of DSM III-R Axis I or I I diagnosis. The results of their semi-structured psychiatric histories, physical examinations, and laboratory screening studies were normal. None were on any medications other than oral contraceptives. Thirteen were right-handed, and 3 were left handed (with signs of mixed dominance). Image
Acquisition
The procedure began by intravenously administering 740 MBq (20 mCi) of Tc99m labeled HMPAO (Amersham Inc.) while the subjects sat quietly with their eyes open and ears unoccluded. Images of the brain were acquired 60 minutes later on a triple headed gamma camera equipped with fan beam collimators (Picker 3000, Cleveland, OH). The intrinsic resolution of the camera was rated as 8-9 mm FWHM (full width half maximum). A stop and shoot mode was used to obtain projection data for up to 1 hour. Acquisition parameters included a 13.5
446
P.D. Mozley et al
cm center of rotation, and a 128x128 matrix with a pixel width of 2.11mm in the projection domain. All images were reconstructed with the same procedure to produce cubic voxels with a side length of 3.56 mm. The projection images were reconstructed with a count-rate dependent Wiener filter (King et al. 1984) that was applied to partially correct for septal penetration, scatter, and finite detector resolution. The modulation transfer function applied by the filter was determined experimentally from the line spread function (Gilland et al. 1988; Press et al. 1988; Kim et al, 1392) for the camera that was used to scan the patients (Kim et al. 1993, 1995). A uniform ellipse was placed around each slice in order to apply Chang’s first order correction method for photon attenuation (Chang 1978). The images were re-aligned in all three planes as necessary in order to be resliced in transaxial sections that were approximately parallel to the plane containing the anterior and posterior commissure line. The orientation was actually chosen by fitting a line on the mid-sagittal section that ran through the apogees of the frontal and occipital poles. The coronal and sagittal slices were then cut in orthogonal planes. The images were then imported into an image analysis package (Resnick et al. 1993), where a “central punch biopsy” approach was used to draw regions of interest (ROIs) around 13 structures in each hemisphere. The technique required each ROI to be smaller than the structure it was placed on, and omitted the ROI from the first or last axial slice on which a structure could be visualized. This tended to minimize the effects of volume averaging around the edges and enhance the precision of measurement, particularly in structures without clearly visualizable boundaries in all directions. The whole supratentorial brain activity was measured on 3 contiguous mid sagittal sections. The coronal sections were used to measure the activity in the frontal and temporal poles. The right and left frontal poles were defined as the region anterior to a coronal plane containing the anterior most aspects of the temporal tips. The inferior aspect of the anterior cingulate gyrus was defined as a region beginning 7 mm below the inferior border of its body on the transaxial plane and continuing inferiorly for 3 slices. The dorsolateral prefrontal cortex (DLPFC) was defined as a 10.5 mm thick section of the lateral most aspect of the frontal cortex midway between the hemispheric fissure and the anterior most temporal lobe beginning 7 mm above the superior boundary of the body of the caudate in the transaxial plane, The ROIs for the temporal poles were placed on the 4 most anterior coronal sections beginning with the second slice on which they could be visualized. The basal ganglia were separated into the head of the caudate beginning below its body, and a lenticular region consisting of a strip between the internal and external capsules. The thalamic ROI was placed on 3 transaxial slices in the center of the structure. The activity
Cerebral HMPAO spect in patients with major depression
447
in the hippocampal-amygdaloid formation was measured on both the transaxial Only the anterior aspect of the and the coronal images for comparison. hippocampus was included. Data Analysis The mean counts per pixel in each region were measured and compared to the mean activity per pixel in the whole brain, the remaining hemisphere, and the cerebellum. Possible hemispheric differences between patients and controls in specific ROIs were examined using loge transformed and untransformed data with parametric unpaired Student t tests and the nonparametric Mann-Whitney U test, respectively. Pearson and Spearman correlation coefficients were then used to characterize the presence of any correlations between the demographic variables and specific measurements of HMPAOregional brain activity. A “laterality score” (Gur et al. 1991) was calculated for each ROI by taking the mean counts per pixel in each selected ROI of the right hemisphere and subtracting this from the mean counts per pixel in the homotopic region of the left hemisphere. The difference between the two hemispheres was then divided by the average activity in both regions. Positive “laterality score” (LS) values indicated that activity in the left-sided ROI relative to the corresponding rightsided ROI. LS values were compared using one-tailed, unpaired t-tests. The proportional differences in positive LS values between patients and controls were then compared by using chi-square analyses with Yates’ correction. The absolute value of the LS score for each region was also used in the analyses without regard to which side of the brain showed more activity. This was used as a measure of “degree of asymmetry” or variability between the two hemispheres. The differences between patients and controls were compared with one-tailed, unpaired t-tests. Results Visual
Inspection
of the Imapes
A review of the images did not reveal any occult brain disease. We obserVed mild asymmetry with decreased radioactivity in several regions of the cortex and diencephalon. However, no specific, consistent abnormalities in HMPAO activity were observed on any of the sagittal, coronal or transaxial images (Fig 1). Relative
HMPAO Activity
in Specific
Computer based, semi-quantitative
Brain Regions
image analysis of the SPECT data revealed
448
P.D. Mozley et al.
HMPAO activity to be highest in the diencephalic regions of patients and controls. The caudate/cortex ratio of mean counts per pixel was 1.81+0.33 in the patients (range 1.19 to 2.45) and 1.76fl.41 in the controls (range 1.27 to 2.33) (p=ns). Similarly, no differences were found between patients and controls in HMPAO activity in the putamen and thalamus, nor in putamen/cortex or thalamus/cortex ratios (Table 1A and 1B). Table 1A Diencephalic Cdudate Patients Controls Difference
to Whole Brain
1.81 F 0.39 1.76 f. 0.41 p = 0.74
Region to Whole Brain Ratios Lenticular
to Whole Brain
1.77 + 0.40 1.71 + 0.42 p = 0.66
Thalamus
to Whole Brain
1.78 + 0.45 1.75 ‘t 0.46 p = 0.84
Table 1B Relative Activity of HMPAO in the Caudate to
Caudate to Cerebellum
Caudate Frontal
1.81 + 0.39 1.76+0.41 p = 0.74
1.30 + 0.28 1.3 1 + 0.30 p = 0.93
1.68 + 0.42 1.77 + 0.45 p = 0.52
Caudate Cortex Patients Controls Indep. t-test
to Pole
Caudate to Temporal Pole 1.75 + 0.32 1.80 LO.59 p = 0.73
No significant differences were observed between subject groups in the ratio of HMPAOactivity in any other ROI to the activity in the cerebellum, the ipsilateral hemisphere, the frontal cortex, or the whole brain (Table 2). The supratentorial cortex to cerebellar ratios defined on 3 contiguous mid-sagittal slices were similar in patients and controls: 0.72fl.05 and 0.75fl.06, respectively (p=ns). Examining regions of the right and left hemisphere independently did not produce any specific findings. Correlations
with Ape and Gender
No significant interactions were observed between the mean HMPAO activity ratios in any brain region and age or gender in this sample using Pearson and Spear-man correlation coefficients.
Cerebral
HMPAO spect in patients
with major depression
449
Fig 1: images of a 37-year old right handed man with depression who committed suicide less than a weeh after study. The regionally specific findings that have been reported in other studies of depression are not apparent. The findings are typical in that many of the most severely affected patients in this sample had completely unremarkable examinations that in clinical practice would have been read as “definitely not abnormal.”
450
P.D. Mozleyet al Table 2 Relative
Region Frontal Frontal DLPFC DLPFC Anterior
HMPAO Activity Patients
Pole to Whole Brain Pole L5 score to Whole Brain LS Score Cingulate
Laterali
1.81 0.01 3.28 -0.01 1.75
in Cortical
Regions
Controls
+ 7.9 LO.03 + 0.82 + 0.12 i 0.51
1.76 F 0.00 + 3.34 * -0.05 + 1 .G7 +
Probability 8.6 0.03 0.90 0.25 0.41
p p p p p
= = = = =
0.82 0.30 0.83 0.55 0.59
tv
Left-to-right hemispheric asymmetry in all ROIs was just as common as rightto-left asymmetry in both subject groups (Table 3). The range of LS values in various diencephalic regions was highly variable in both patients and controls. 68% (13/13) of the patients showed asymmetry in at least one of the three diencephalic regions that was >7% of the mean activity. similarly, 56% (9/16) of controls also demonstrated an asymmetry. However, while 26% (5/19) of the patients showed more than one area of asymmetric rCBF in the dfencephalon, only one control (6%) had two regional asymmetries. Asymmetries between the left and right caudate ranged from 0.81% to 14.4% in patients, and from 0.0% to 15.8% in controls. There were no consistent diencephalic rCBF asymmetries within either subject group, or between groups. Table 3 Diencephalic
Lenticular
Caudate Patients Controls Difference
Laterality
0.89 i 7.90 1.38 + 6.66 p = 0.86
0.53 + 5.42 -0.04 + 5.49 p = 0.76
Scores Thalamus 2.54 + 10.11 -0.52 L 5.65 p = 0.29
Similarly , there were no significant differences between patients and controls in the presence of hemispheric rCBF asymmetries for any of the following ROIs:
frontal cortex, temporal cortex (as measured by coronal and transaxial images), hippocampus/amygdala, or dorsolateral prefrontal cortex. Degree
of Temooral
Asymmetry
in Patients
vs. Controls
The absolute value of the LS index (degree of asymmetry) was specifically examined in the anterior temporal lobes because of prior observations of
Cerebral
HMPAO spect
in patients
with major
451
depression
temporal lobe asymmetry with [l-123] IMP SPECT (Amsterdam and Mozley 1992). Interestingly, with HMPAO there was no apparent hemispheric laterality in temporal lobes when patients were compared to controls. While the patients did demonstrate a greater overall degree of asymmetry compared to controls (6.83-&30 vs. 3.4722.21; p=O.OS, unpaired t-test), the direction of right-left asymmetry was not consistent in either group, and there was a substantial overlap between groups (Table 4). As noted above, there were no statistically significant differences between the patient and control groups in the ratio
1.02+0.11 1.01 + 0.08
Difference
p = 0.91
Activity Temporal LS Score
of HMPAO in Temporal Pole
HippocampalIAmygdala to Whole Brain
1.10 + 9.97 -0.61 + 4.17 p=O.Sl
1.38 + 0.36 1.39 + 0.39 p = 0.93
Regions Hippocampal/Amygdala LS Score -0.31 i 6.91 0.19 + 8.09 p = 0.85
Discussion Observations from the present HMPAO study do not support prior reports of specific rCBF abnormalities in patients with MDD. However, the authors did observe greater variability in regional HMPAO activity in depressed patients compared to control subjects. However, there were no other findings which differentiated the patients from the controls. While some investigators have reported specific abnormalities in regional brain activity during depression (Sackeim et al. 1990; Amsterdam & Mozley 1992; Austin et al. 1992; Ebert et al. 1991; George et al. 1993; Mayberg et al. 1994), Maes et al. (1993) failed to detect any abnormalities in the distribution of HMPAO in depressed patients compal’ed to controls. Clinical Correlates
of SPECT FindinPs
Several clinical and technical reasons might explain the divergent observations in various SPECT studies in depression. For example, some studies reporting discrete rCBF abnormalities with HMPAO have included older natients who tiav
452
P.D. Mozley et al.
normally show more frequent rCBF disturbances (Philpot et al. 1993). Investigations with Xenon-133 SPECT suggest that there may be an interaction of age with depressive subtype for both global and rCBF (Devous et al. 1991). One of the authors previously found a significant effect of normal aging on the distribution of HMPAOin healthy subjects (Payer et al. 1994). In contrast, in the present study the patients were relatively young and without any significant effects of age on HMPAOactivity. Additionally, with Kumar et al. (1991), one of us previously reported a decreased cortical to cerebellar ratio in elderly depressed patients using IMP SPECT. In contrast, in the present study, only a slightly higher cortical to cerebellar ratio in the patients (vs. controls) was observed. The presence of depressive psychotic and/or melancholic features has been associated with several SPECT abnormalities (Austin et al. 1992; Ebert et al. 1991). In contrast, while some of our subjects did have melancholic symptoms, none had overt psychotic features. This clinical difference may have partly accounted for our failure to observe discrete SPECTabnormalities in our patients. Severity of depression has also been found to positively correlate with changes in rCBF (Schlegel et al. 1989). In contrast, we were unable to demonstrate any correlation between HDRS scores and regional HMPAO activity. Interestingly, in a prior study using IMP SPECT, we observed right temporal lobe asymmetry in depressed patients compared to medically ill controls (Amsterdam & Mozley 1992). Our failure to observe specific regional blood flow asymmetries in the present study using HMPAO suggests that the physical and pharmacologic properties of these tracers may result in different patterns of rCBF in depression. For example, HMPAOmay not be as sensitive to relatively small changes in brain function as other tracers. A high degree of concordance has been reported between HMPAO and other blood flow tracers in pathological processes that However, several independent groups of produce neuronal necrosis. investigators have now reported that HMPAO cannot detect small reductions in local blood flow that might be demonstrable with other agents (Heiss et al. 1990; Toyama et al. 1994; Pointon et al. 1994). Others have observed that the regional activity of HMPAO can actually be increased in areas of ischemia (Sperling & Lassen 1993). It is, therefore, possible that the present observations may reflect a relative insensitivity of HMPAO as a tracer for identifying discrete, functional abnormalities in MDD. Comparison
with Other
Imaging
Techniques
Positron emission tomography (PET) has been able to detect abnormalities in similar groups of patients (Buchsbaum et al. 1984; Baxter et al. 1985; Post et al.
Cerebral HMPAO spect in patients with major depression
453
1987;
Cohen et al. 1989). Cerebral blood flow and metabolism are frequently coupled (Sokoloff 1992). However, several unique biological features of HMPAO may limit its comparison with PET studies of depression. Metabolic tracers such as 18F-labeled fluorodeoxyglucose (FDG) tend to localize within the synaptic bouton, where the work of neural transmission is actually performed (Kadekaro et al. 1985). In contrast, HMPAO tends to localize in the cytoplasm of the neuronal cell body (Anderson et al. 1988) where only vegetative maintenance functions are being performed. The cell bodies are frequently located at some distance from their synaptic terminals, and have significantly less surface area to take up a radiotracer than their axonal processes. This may make HMPAO less biologically sensitive in from some functional disorders that produce reversible abnormalities of perisynaptic neurotransmission than FDG PET (Mielke et al. 1994). Thus, while HMPAO may a sensitive agent for detecting abnormalities characterized by neuronal necrosis, the present study suggests that it is less sensitive as a tracer for identifying the more subtle abnormalities of functional psychiatric disorders. The relatively enhanced resolution in our current study should have made it more likely for a true abnormality in HMPAO activity to be observed. While we have previously found that the restorative filter is more accurate than conventional low pass filters, the risks of using a count rate dependent filter usually include producing false positive results, not false negatives. The present image analysis technique has previously been validated and shown to be a reliable one. In the present study, image analysis benefited from the enhanced edge detection produced by the filter. However, it is important to note the caveat of a lack of co-registration with anatomical images (e.g., magnetic resonance imaging). However, this is not likely to have been a significant factor in identifying abnormalities in ROIs such as the head of the caudate and the temporal poles. The image thresholding technique in the present study enhanced precision and compensated for anatomical accuracy in these regions. Conclusions Visual inspection of HMPAO SPECT images of the brain in the depressed patients did not demonstrate any specific regional abnormalities. Moreover, computer-enhanced, semi-quantitative analysis failed to reveal any more subtle abnormalities in HMPAO activity that were not observed on visual inspection. The present results suggest that HMPAO may not identify discrete abnormalities in regional brain dysfunction in patients with major depression. Future investigations using diverse SPECTradioligands in the same individual may help to resolve the conflicting observations presently seen in brain imaging studies.
454
P.D. Mozley et aL
Acknowledgement This work was supported by an educational grant-in-aid from Eli Lilly and Company Central Research, NIMH Neuropsychopharmacology Fellowship PHS Grant MH 14654, and the Jack Warsaw Fund for Research in Biological Psychiatry. The technical support staffs in the Depression Research Unit and the Division of Nuclear Medicine made substantial contributions to the performance of this study. References ALAVI,A. and HIRSCH,L.J. (1991) Studies of central nervous system disorders with single photon emission computed tomography and positron emission tomography: evolution over the past 2 decades. Sem. Nucl. Med. 2: 58-81. AMERICANPSYCHIATRICASSOCIATION( 1987) Diagnostic and Statistical Manual of Mental Disorders, 3rd ed., revised, American Psychiatric Association, Washington. AMSTERDAM,J.D. and MOZLEY,P.D. (1992) Temporal lobe asymmetry with tofetamine (IMP) SPECT imaging in patients with major depression. J. Affective Disord. 24: 43-53. ANDERSON,A.R., FRIBERG,H.H., SCHMIDT,J.F. and HASSELBALCH, S.G. ( 1988) Quantitative measurements of cerebral blood flow using SPECTand [99mTc]d,l-HMPAO compared to Xenon-133. J. Cerebral Blood Flow Metab. 8: S69-S81. ANDERSON,AR. (1989) 99mTc-d,l-Hexamethylene-propyleneamine oxime (99mTc-HMPAO): Basic kinetic studies of a tracer of cerebral blood flow. Cerebrovasc. Brain Metab. Rev. 1: 288-3 18. AUSTIN,M.-P., DOUGALL,N., ROSS,M., MURRAY,C., O’CARROLL,R.E.,MOFFOOT,A., EBMEIER,K.P. and GOODWIN,G.M. (1992) Single photon emission computed tomography with 99mTc-exametazime in major depression and the pattern of brain activity underlying the psychotic/neurotic continuum. J. Affective Disord. 26: 31-44. BAXTER,L.R., PHELPS,M.E., MAZZIO’ITAJ.C., SCHWARTZ,J.M., GERNER,R.H., SELIN, C.E. and SUMIDA,R.M. (1985) Cerebral metabolic rates for glucose in mood disorders. Arch. Gen. Psychiatry 112:441-447. BAXTER,L.R.,GUZE,B.H.,SCHWARTZ,J.M., PHELPS,M.E.,MAZZIOTTAJ.C. and SZUBA,M.P. (1992) PET studies of cerebral function in major depression and related disorders. In: Brain Work and Mental Activity: Quantitative Studies
Cerebral HMPAO spect in patients with major depression
with Radioactive Tracers, N.A. Lassen, D.H. Ingvar, M.E. Raichle, H.H. Friberg (Eds.), pp 52-65, Munksgaard, Copenhagen. BENCH,C.J., FRISTON,K.J., BROWN,R.G.,FRACKOWLAK, R.S.J. and DOLAN,R.J. ( 1993) Regional cerebral blood flow in depression measured by positron emission tomography: the relationship with clinical dimensions. Psychol. Med. 23: 579-590. BUCHBAUM,M.S., DELISI,L.E., HOLCOMB,H.H., CAPPELLETTI,J., KING,A.C., JOHNSON,J., HAZLE’M’, E., DOWLING-ZIMMERMAN, S., POST, R.M., MORIHISA,J., CARPENTER,W., COHEN,R., PICKAR,D., WEINBERGER, D.R., MARGOLIN,R. and KESSLER,R.M. (1984) Anteroposterior gradients in cerebral glucose use in schizophrenia and affective disorders. Arch. Gen. Psychiatry 41: 1153-l 166. CHANG,L.T. (1978) A method for attenuation correction in radionuclide computed tomography. IEEETrans. Nucl. Sci. 25: 638- 643. COHEN,R.M.,SEMPLE,W.E., GROSS,M., NORDAHL,T.E., KING,AC., PUICKAR,D. and POST, R.M. (1989) Evidence for common alterations in cerebral glucose metabolism in major affective disorders and schizophrenia. Neuropsychopharmacol. 2: 241-254. DEVOUS,M.D. Sr., GULLION,C., GRANNERMAN,B. and RUSH,A.J. ( 1991) Regional cerebral blood flow alterations in unipolar depression. J. Nucl. Med. 32: 95 l952. EBERT,D., FEISTEL,H. and BAROCKA,A. (1991) Effects of sleep deprivation on the limbic system and the frontal lobes in affective disorders: a study with Tc99m-HMPAO SPECT. Psychiatry Res. Neuroimaging 40: 247-251. ELL,P.J., HOCKNELL,J.M.L., JARRI’I-F,P.H., CULLUM,I., LUI, D., CAMPOS-COSTA,D., NOWOTNIK,D.P., PICKETT,R.D., CANNING,L.R.and NEIRINCKX,R.D. ( 1985) A Tc99m-labelled radiotracer for the investigation of cerebral vascular disease. Nucl. Med. Commun. 4: 437-441. GEORGE,M.S., KEITER, T.A. and POST, R.M. (1993) SPECTand PET imaging in mood disorders. J. Clin. Psychiatry 54(suppl): 6-13. GILLAND,D.V., TSUI, B.M.W.,MCCARTNEY,W.H., PERRY,J.R. and BERG,J. (1988) Determination of the optimal filter function for SPECTimaging. J. Nucl. Med. 25): 643650. GUR,R.E.,SKOLNICK,B.E.,GUR,R.C.,CAROFF,S., RIEGER,W., OBRIST,W.D., YOUNKIN,D. and REMCH, M. (1984) Brain function in psychiatric disorders. II: Regional cerebral blood flow in medicated unipolar depression. Arch. Gen.
455
456
P.D. Mozley et al.
Psychiatry 41: 695-699. CUR,R.C., MOZLEY,P.D., GUR,R.E.,RESNICK,S.M.,GOTTLIEB,G.E.,KOHN,M., ZIMMERMAN,R., HERMAN,G., ATLAS,S., GROSSMAN,R., BARRE’ITA,D., ERWIN, R. and GUR, R.E. (1991) Gender differences in age effect on brain atrophy measured by magnetic resonance imaging. Proc. Natl. Acad. Sci. 88: 2845-2849. HAMILTON,M. ( 1960) A rating scale for depression. Psychiatry 23: 56-62.
J Neurol Neurosurg
HEISS,W.D., HERHOLZ,K., PODREKA,I., NEUAUER,I. and PIETRZYK,U. ( 1990) Comparison of [99mTc]HMPAO SPETwith [ 18Flfluoromethane PET in cerebrovascular disease. J. Cereb. Blood Flow Metab. 10: 687-697. HOLMES,B.L.and DEVOUS,M.D. (1992) Functional brain SPECT:the emergence of a powerful clinical method. J. Nucl. Med. 33: 1888-1904. KADEKARO,M., CRANE,A.M. and SOKOLOFFL (1985) Differential effects of electrical stimulation of sciatic nerve on metabolic activity in spinal cord and dorsal root ganglion in the rat. Proc. Natl. Acad. Sci. 73:4230-4234. KIM, H.J., ZEEBERG, B.R.,FAHN,
F.H., HOFFMAN,E.J. and REBA,R.C. (1992) 3-D SPECTsimulations of a complex 3-D mathematical brain model: Effects of 3-D geometric detector response, attenuation, scatter, and statistical noise. IEEE Trans. Med. Imag. 11: 176-184.
KIM,H.J., KARP,J.S., KUNG,J.S. and MOZLEY,P.D. (1993) Quantitative effects of a count rate dependent Wiener filter on image quality: a basal ganglia phantom study simulating [I-123] dynamic SPECTimaging. J. Nucl. Med. 34: 190 (abstract). KIM H.J., KARP,J.S., KUNG,J.S., MOZLN, P.D. and ALAVI,A. ( 1995) Collimator and filter selection: simulating technetium-99m cerebral perfusion studies with a three-dimensional Hoffman brain phantom. J. Nucl. Med. 34: 195 (abstract). KING, M.A.,
SCHWINGER,R.B.,DOHERTY,P.W. and PENNY,B.C. ( 1984) The twodimensional filtering of SPECTimages using the Metz and Wiener filters. J. Nucl. Med. 25: 1234-1240.
KUMAR, A., MOZLN,
P.D., VELCHIK,M.G., DUNHAM,C., REILLN, J., GOTTLIEB,G. and AIAVI, A. (1991) State dependent decrease in cortical to cerebellar IMP activity in patients with late onset depression. Internat J Geriatric Psychiatry 4: 775-777.
MAES,M., DIERCK,R., MELTZER,H.Y.,INGELS,M., SCHOTTE,C., VANDEWOUDE,M.,
Cerebral HMPAOspect in patients with major depression
CALABRESE, J. and COSYNS,P. (1993) Regional cerebral blood flow in unipolar depression measured with Tc-99m-HMPAO single photon emission computed tomography: negative findings. Psychiatry Res. Neuroimaging 30: 77-88. MATTHEW,R.J., MEYER,J.S., FRANCIS,D.J., SEMCHUK,K.M.,MORTEL,K. and CLAGHORN,J.I. (1980) Cerebral blood flow in depression. Am. J. Psychiatry 137: 1449-1450. MAYBERG,H.S., LEWIS,P.J., REGENOLD,W. and WAGNER,H.N. JR. ( 1994) Paralimbic hypoperfusion in unipolar depression. J. Nucl. Med. -35. 929-934. MIELKE,R., PIETRZYK,U., JACOBS,A., FINK, G.R., ICHIMIYA,A., KESSLER,J., HERHOLZ,K. and HEISS,W.D. (1994) HMPAOSPETand FDG PET in Alzheimer’s disease and vascular dementia: comparison of perfusion and metabolic pattern. Eur. J. Nucl. Med. 21, 1052-1060. PAYER,F., MOZLEY,P.D. and ALAVI,A. (1994) Age related changes in the cerebral distribution of HMPAO. J. NucI. Med. 35: 1007P (abstract). PHILPOT,M.P., BANERJEE,S., NEEDHAM-BENNETT, H., COSTA,DC. and ELL,P.J. (1993) 99mTc-HMPAO single photon emission tomography in late life depression: a pilot study of regional cerebral blood flow at rest and during a verbal fluency task. J. Affective Disord. 28: 233-240. POINTON,J., SINGH,P., EGAN,G., CHAN,G., TOCHON-DANGU,H., NEWTON,M.R., BERLANGIERI, S.U., O’KEEFE,G., MCKAY,W.J. and DONNAN,G.A. (1994) PET and SPECTimaging in acute stroke: perfusion patterns and clinical outcome. Eur. J. NucI. Med a(supp1): S23. POST,R.M., DELISI,L.E.,HOLCOMB,H.H,, UHDE,T.W., COHEN,R and BUCHSBAUM, M.S. (1987) Glucose utilization in the temporal cortex of affectively ill patients: positron emission tomography. Biol. Psychiatry 22: 545553. PRESS,W.H., FLANNERY,B.P., TEUKOLSKY,S.A. and VETTERLING,W.T. (1988) Numerical Recipes in C: The Art of Scientific Computing, Cambridge University Press, Cambridge, pp 407-412. RESNICK,S.M., KARP,J.S., TRETSKY,B.I. and GUR,R.E. ( 1993) Comparison of anatomically defined versus physiologically based regional localization: Effects on PET-FDGquantitation. J. Nucl. Med. 34: 201-208. SACKEIM,H.A., PROHOVNIK,I., MUELLER,J.R., BROWN,R.P., APTER,S., PRUDIC,J., DEVANAND,D.P. and MUKHERJEE,S. (1990) Regional cerebral blood flow in mood disorders: I. Comparison of major depressives and normal controls at re$t. Arch. Gen. Psychiatry 47: 60-70.
457
458
P.D. Mozley et aL
SCHLEGEL,S., ALDENHOFF,J.B., EISSNER,D., LINDNER,P. and NICKEL,0. ( 1989) Regional cerebral blood flow in depression: association with psychopathology. Affective Disord. 17: 211-218.
J.
SILFVERSKIOLD, P. and RISBERG,J. (1389) Regional cerebral blood flow in depression and mania. Arch. Gen. Psychiatry 46: 253-259. SOKOLOFF,L. (1992) Relationship between functional activity and energy metabolism in the nervous system: Whether, where, and why? In: Brain Work and Mental Activity: Ouantitative Studies WithRadioactive Tracers, N.A. Lassen, D.H. Ingvar, M.E. Raichle, H.H. Friberg (Eds.), pp 52-65, Munksgaard, Copenhagen. SPERLING,B. and LASSEN,N.A. (1933) Hyperfixation of HMPAOin sub-acute ischemic stroke leading to spuriously high estimates of cerebral blood flow by SPECT. Stroke 24: 133-l 94. SPITZER,R.L., WILLIAMS,J.B.W., GIBBON,M. and FIRST, M.B. ( 1989) Instruction Manual for the Structured Clinical Interview for DSM-IIIR (SCID, 5/l/89 Revision), Biometrics Dept., NYState Psychiatric Institute, New York. TOYAMA,H., MAEDA,H., TAKESHITA,G., SHIBATA,K., NOMURA,M., KATO,M., SAKAKIBARA,E., TAKEUCHI,A. and KOGA,S. (1994) Tc-99m HM-PA0 super dynamic SPECT to detect the early cerebral blood flow distribution. Eur J. Nucl. Med. 2l(suppl): S4. Inquiries and reprint requests should be addressed to: Jay D. Amsterdam, M.D. Depression Research Unit University of Pennsylvania Medical Center IJniversity Science Center, 8th floor 3600 Market Street Philadelphia, PA 19 104 U.S.A.