Open Journal of Medical Imaging, 2012, 2, 19-22
http://dx.doi.org/10.4236/ojmi.2012.21003 Published Online March 2012 (http://www.SciRP.org/journal/ojmi)
Factors Affecting Brain Metabolism Measured with 18FDG
Hongyun Zhu1, Michael L. Goris2
1Department of Radiology, University of Arkansas for Medical Sciences (UAMS), Little Rock, USA
2Division of Nuclear Medicine, School of Medicine, Stanford University, Stanford, USA
Received December 2, 2011; revised January 26, 2012; accepted February 8, 2012
ABSTRACT
An observational finding found a large variation in the brain SUV in patients with multiple myeloma undergoing
PET/CT. The first hypothesis considered a toxic effect of chemotherapeutic agents, but no correlation was found with
hematological signs of toxicity. Low brain FDG uptake has been described with anesthesia, but this was not relevant in
this case. An alternative is the presence of a large FDG avid mass, but that was excluded. Since there was a question of
chemotherapy toxicity, the metrics used for comparison were Hemoglobin levels (Hgb, g/dl), Erythrocyte count (RBC,
M/μL), Lymphocytes absolute counts (Lymph#, K/μL) and % (lymph, %), Granulocytes Neutrophils, K/μL), age and
C-reactive protein levels (CRP, g/L). The liver SUV (standardized uptake value) was included to eliminate unexpected
global effects on the SUV values, since FDG uptake is a competitive system with a single source (plasma FDG levels).
There was in fact no correlation between brain SUV and hepatic SUV, eliminating the so-called super scan effect. Fur-
ther analysis, however, revealed a strong positive correlation with hemoglobin or RBC levels, but an inverse effect with
Neutrophils, C-reactive proteins and age (in years). The results suggest that brain metabolism strongly depends on oxy-
gen supply and may be depressed by general inflammatory diseases and independently with age. If the variation of glu-
cose metabolism correlates with cognitive deficits (CD), considering general measures of good health may be a first
step for relief of age related CD.
Keywords: Brain 18-FDG Uptake; Anemia; Inflammation
1. Introduction
An observational finding found a large variation in the
brain SUV (standardized uptake value) in patients with
multiple myeloma undergoing PET/CT. The first hypo-
thesis considered a toxic effect of chemotherapeutic agents,
but no correlation was found with hematological signs of
toxicity. Low brain FDG uptake has been described with
anesthesia, but this was not relevant in this case. An al-
ternative is the presence of a large FDG avid mass, but
that was excluded.
Since there was a question of chemotherapy toxicity,
the metrics used for comparison were Hemoglobin levels
(Hgb, g/dL), Erythrocyte count (RBC M/μL), Lympho-
cytes absolute counts (Lymph#, K/μL) and % (lymph, %),
Granulocytes (Neutrophils K/μL), age (in years) and C-
reactive protein levels (CRP, g/L). The liver SUV was
included to eliminate unexpected global effects on the
SUV values, since FDG uptake is a competitive system
with a single source (plasma FDG levels).
2. Materials and Methods
There were 40 patients, (70% male), average age 60.7
(range 44-83), with 173 FDG PET/CT scans. The reason
for the study was evaluation of Multiple Myeloma ther-
apy. Patients were routinely followed with the metrics
cited above. Brain SUV was measured in the central brain
cortex, but the value was not site sensitive, since the cor-
relation between central and visual was 0.95 and between
central and nuclei was 0.96. In addition, patients with
abnormal cerebral regional FDG distributions (e.g. as in
Alzheimer’s, Fronto-parietal dementia, multi-stroke de-
mentia) were not included. The central brain activity in
SUV units was on average 4.96 ± 1.84 (range 1.5 - 9.7).
The average glucose level was 134 ± 27 mg/dl (range 104-
198).
After the FDG injection, the patients were kept in a
quiet, dimly lighted room, for 1 hr, before imaging be-
gan.
The analysis consisted first in an evaluation of the inter-
correlation of the metric listed above prior to using multi-
ple regressions. Multiple regressions were performed first
with all metrics and later only with those not positively
inter-correlated or with a significant single regression
with brain SUV.
3. Results
Central brain SUV correlated significantly negative with
age (p < 0.001), CRP (p < 0.05) and Neutrophils (p <
C
opyright © 2012 SciRes. OJMI

H. Y. Zhu ET AL.
20
0.01) and significantly positive with RBC (p < 0.001),
Hgb (p < 0.001), lymph% (p < 0.005) and Lymph# (p <
0.05).
There was significant negative correlation between
WBC and liver SUV (p < 0.05) and between WBC and
lymph% (p < 0.001). However, the linear regression be-
tween WBC and brain SUV was not significant (p >
0.05). Lymph% correlated positively with RBC (p <
0.001) and Hgb (p < 0.001).
The results was that multiple regressions were per-
formed only for HGB (not RBC), Age, Neutrophils,
lymph# and CRP. The liver SUV was included to test for
a super-scan effect.
The results of single regression between the various
metric and brain SUV with their p values are shown in
Table 1.
Table 2 shows serial multiple regressions in which the
significant factors from the previous multiple regressions
are withdrawn.
4. Discussion
Recent and remote chemotherapy could influence brain
uptake. However, the fact that a potential effect of che-
motherapy would vary in the subject is not relevant in
this study. First, it is unlikely to be correlated with age.
Second, the effect on bone marrow, if there was one,
would be a decrease in both red and white blood ele-
ments. What we observe is an opposite effect on brain
uptake of RBC and Neutrophils. Age by itself did not
correlate with the next 2 important factors, the level of
hemoglobin or red cell counts, or the level of neutrophils.
Diffuse low glucose metabolism measured with FDG
has been described in Halothane anesthesia [1], in low
states of consciousness [2], and paradoxically with low
blood glucose [3]. In addition it has been observed in
diffuse brain injury [4] without large focal abnormalities.
An alternative explanation is due to a sink effect, if other
structures or pathologies are very FDG avid. They would
then act as a FDG sink, since all structures in the body
accumulate FDG from the same plasma pool. This phe-
nomenon has been described as a metabolic super scan
[5].
None of those factors was present in this population.
The lack of correlation between brain and liver SUV ex-
cludes the metabolic super scan. Glucose levels in the
plasma were within limit (above 6 mM). There was no
(generalized) history of trauma or low consciousness.
5. Conclusion
Non-neurological factors (anemia, age, inflammation)
seem to adversely affect FDG uptake in the brain. Those
effects are independent of age, and, if shown to be causa-
tive, would suggest methods of therapeutic interventions.
Table 1. Negative effects are age, CRP, Neutrophils and
WBC (not significant). Except for age they are all associ-
ated with inflammation. Positive effects are Hgb and RBC,
lymph#, and lymph%. There is no significant correlation
between br ain and liver SUV.
Slope p<
Age –0.048 0.001
HgB 0.376 0.001
Lymph# 0.368 0.001
CRP –0.011 0.033
Neutro –0.077 0.018
Liver 0.244 0.600
WBC –0.061 0.059
RBC 1.093 0.001
Lymph% 0.026 0.007
Table 2. With the uncorrelated metrics if all are used in
multiple regressions, Age and Hgb are significant. If those
are eliminated, Lymph# and Neutrophils are significant.
The next significant factor is the CRP. The liver does not
reach significance. The F-ratio is the ratio of the variance in
brain SUV explained by the regression over the residual
variance in brain SUV.
Metric p value
Hemoglobin1.1E–05
Age 0.0001
Neutro 0.0760 0.0295
Lymph 0.3103 0.0362
CRP 0.950 0.0560 0.0346
Liver 0.8122 0.7997 0.5388
F-Ratio 1.72E–08 0.007189 0.089
If the low FDG uptake is associated with cognitive defi-
cit, it may be that improving those general dysfunctions
(except for age), may be the first step and easiest thera-
peutic approach.
REFERENCES
[1] M. T. Alkire, C. J. Pomfrett, R. J. Haier , M. V. Gianzero,
C. M. Chan, B. P. Jacobsen and J. H. Fallon “Functional
Brain Imaging during Anesthesia in Humans: Effects of
Halothane on Global and Regional Cerebral Glucose Me-
tabolism,” Anesthesiology, Vol. 90, No. 3, 1999, pp. 701-
709. doi:10.1097/00000542-199903000-00011
[2] M. Boly, C. Phillips, L. Tshibanda, A. Vanhaudenhuyse,

H. Y. Zhu ET AL.
21
M. Schabus, T. T. Dang-Vu, G. Moonen, R. Hustinx, P.
Maquet and S. Laureys, “Intrinsic Brain Activity in Al-
tered States of Consciousness,” Annals of the New York
Academy of Sciences, Vol. 1129, 2008, pp. 119-129.
doi:10.1196/annals.1417.015
[3] R. Meierhans, M. Béchir, S. Ludwig, J Sommerfeld, G.
Brandi, C. Haberthür, R. Stocker and J. F. Stover, “Brain
Metabolism Is Significantly Impaired at Blood Glucose
below 6 mM and Brain Glucose below 1 mM in Patients
with Severe Traumatic Brain Injury,” Critical Care, Vol.
14, No. 1, 2010, p. R13. doi:10.1186/cc8869
[4] N. Nakayama, A. Okumura, J. Shinoda, T. Nakashima
and T. Iwama, “Relationship between Regional Cerebral
Metabolism and Consciousness Disturbance in Traumatic
Diffuse Brain Injury without Large Focal Lesions: An
FDG-PET Study with Statistical Parametric Mapping
Analysis,” Journal of Neurology, Neurosurgery & Psychi-
atry, Vol. 77, No. 7, 2006, pp. 856-862.
doi:10.1136/jnnp.2005.080523
[5] D.-W. Kim, C. G. Kim, S.-A. Park, S.-A. Jung and S-H.
Yang, “Metabolic Super Scan in 18F-FDG PET/CT Im-
aging,” Journal of Korean Medical Science, Vol. 25, No.
8, 2010, pp. 1256-1257.
doi:10.3346/jkms.2010.25.8.1256
Supporting Data
Table S1. Inter-corre l ation betw een metrics and brain SUV.
Brain frontal Age Liver WBC RBC HemoglobinLymph%Lymph# CRP Neutro
Brain frontal 1
Age –0.2747 1
Liver 0.047734 0.021864 1
WBC –0.14381 0.132306 –0.1327 1
RBC 0.33053 0.03899
–0.00581 0.0525841
Hemoglobin 0.37794 0.010832 0.05235 0.0142430.917851
Lymph% 0.205287 0.079511 0.119979 –0.296810.1217820.1985461
Lymph# 0.167963 0.068117 0.029356 0.173632 0.3040250.3208820.6621591
CRP –0.16137 0.006303 –0.00727 0.050013–0.23132–0.26407–0.08514–0.05294 1
Neutro –0.17894 0.120962 –0.14047 0.981173–0.00624–0.04847–0.43118–0.01984 0.0611561
Inter-correlation of brain SUV and the metrics used in this paper. The yellow background highlights the negative correlations.
Table S2. The significance of the correlation shown in Table S1 is not shown if p > 0.05.
Brain frontal Age liver WBC RBC HemoglobinLymph%Lymph# CRP Neutro
Brain frontal
Age 0.001
liver
WBC 0.05 0.05 0.05
RBC 0.001
Hemoglobin 0.001 0.001
Lymph% 0.005 0.0010.005
Lymph# 0.05 0.050.0010.0010.001
CRP 0.05 0.005 0.001
Neutro 0.01 0.05 0.0010.001
Brain central activity correlated negatively with Age (p < 0.001), WBC (p < 0.05), CRP (p < 0.05), and Neutrophils (Neutro) (p < 0.01), but positively with
RBC (p < 0.01), Hemoglobin (Hgb) (P, 0.01), lymphocytes % (lymph%) (p < 0.005) and lymphocytes absolute numbers (Lymph#) (p < 0.05).

H. Y. Zhu ET AL.
22
Table S3. Multple regressions using all metrics: significance.
Metric p-values
Age 0.0002
Hemoglobin 0.0735 0.0418
RBC 0.6784 0.9175 0.0000
Lymph # 0.2992 0.1389 0.09960.0234
Lymph % 0.1201 0.4007 0.24760.48410.0251
Neutro 0.1611 0.1000 0.05990.05030.91760.0312
CRP 0.6016 0.5178 0.36770.08310.21700.32620.0415
WBC 0.1678 0.0931 0.05700.0492 0.8632 0.0317 0.0834 0.0834 0.0693
liver 0.3891 0.7606 0.5700 0.1949 0.1277 0.0420 0.7020 0.7020 0.7037
F-ratio 3.1E–09 5.1E–07 1.2E–060.002260.0090.0360.048 0.048 0.15
With all the metric included, the significant factor is age. All other metrics do not significantly contribute. If age is eliminated, the significant contribution is
from hemoglobin. Next come RBC, then lymph#, then lymph% etc. Some factors contribute negatively (gray background). WBC and liver never contribute
significantly. The F-ratio is the ratio of the contribution of the regression to the variance of the data over the residual variance. When only liver and WBC re-
main the F-ratio is not significant.
Table S4. Range of significant metrics.
Metric Means Range
Liver SUV 2.19 ± 0.36 1.5 - 3.6
Hgb 10.62 ± 1.86 6.8 - 17.2
Lymph# 1.00 ± 0.84 0.0024 - 5.44
CRP 19.05 ± 25.51 0 - 148
Neutrophils 4.11 ± 4.30 0.04 - 42.46
The table includes the metrics selected for multiple regressions.
Table S5. Individual correlations w e re : Table S5.
Slope p<
Age –0.048 0.001
HgB 0.376 0.001
Lymph# 0.368 0.001
CRP –0.011 0.033
Neutro –0.077 0.018
Liver 0.244 0.600
WBC –0.061 0.059
RBC 1.093 0.001
Lymph% 0.06 20.007
Table S6. Multiple regressions between all metrics and brain
SUV.
Metric Slopes Standard Error p-value
Hgb 0.3332 0.0735 0.00001
Age –0.0472 0.0120 0.00012
Intercept 4.2009 1.2801 0.00126
Neutro –0.0528 0.0296 0.07603
Lymph# 0.1594 0.1566 0.31030
CRP –0.0042 0.0049 0.39498
liver 0.0832 0.3496 0.81221
When those (Table S6) metrics are used with their slope and the intercept,
the predicted value BeSUV = 4.2009 – 0.0472 × age – 0.0528 × Neutrophiles
+ 0.1594 × lymphocytes – 0.0042 × CRP + 0.0832 × LiverSUV the prediction
(expected) is highly significant and not significantly different from unity
(Table S7).
Table S7. Prediction of brain SUV using all significant not
intercorrelated metrics.
Coefficients Standard
Error p-value
Intercept 0.0000 0.6781 1
Expected 1.0000 0.1343 4.4754E–12
F-ratio 4.4754E–12