G Model

REMNIM-932; No. of Pages 7 ARTICLE IN PRESS

Rev Esp Med Nucl Imagen Mol. 2017;xxx(xx):xxx–xxx

Original Article

Correlation of 18 F-FDG uptake on PET/CT with Ki67

immunohistochemistry in pre-treatment epithelial ovarian cancer夽
M. Mayoral a,1 , P. Paredes a,b,∗,1 , A. Saco c , P. Fusté d , P. Perlaza a , A. Tapias a , A. Fernandez-Martinez e ,
L. Vidal e , J. Ordi c,f , J. Pavia a,b,g , S. Martinez-Roman b,d , F. Lomeña a,b
a
Nuclear Medicine Department, Hospital Clínic, Barcelona, Spain
b
Institut d’Investigacions Biomèdiques August Pi I Sunyer (IDIBAPS), Barcelona, Spain
c
Department of Pathology, Hospital Clínic, Barcelona, Spain
d
Gynaecology Department, Hospital Clínic, Barcelona, Spain
e
Medical Oncology Department, Hospital Clínic, Barcelona, Spain
f
Centre de Recerca en Salut Internacional de Barcelona (CRESIB), Barcelona, Spain
g
CIBER de Bioingeniería, Biomateriales y Nanomedicina (CIBER-BBN), Barcelona, Spain

a r t i c l e i n f o a b s t r a c t

Article history: Objective: Standardized uptake value (SUV) and volumetric parameters such as metabolic tumour volume
Received 12 March 2017 (MTV) and total lesion glycolysis (TLG) from 18 F-FDG PET/CT are useful criteria for disease prognosis in
Accepted 9 July 2017 pre-operative and post-treatment epithelial ovarian cancer (EOC). Ki67 is another prognostic biomarker
Available online xxx
in EOC, associated with tumour aggressiveness. The aim of this study is to evaluate the association
between 18 F-FDG PET/CT measurements and Ki67 in pre-treatment EOC to determine if PET/CT parame-
Keywords: ters could non-invasively predict tumour aggressiveness.
Epithelial ovarian cancer
18 Material and methods: A pre-treatment PET/CT was performed on 18 patients with suspected or newly
F-FDG PET/CT
SUV
diagnosed EOC. Maximum SUV (SUVmax), mean SUV (SUVmean), whole-body MTV (wbMTV), and whole-
Metabolic tumour volume body TLG (wbTLG) with a threshold of 30% and 40% of the SUVmax were obtained. Furthermore, Ki67
Total lesion glycolysis index (mean and hotspot) was estimated in tumour tissue specimens. Immunohistochemical findings
Ki67 were correlated with PET parameters.
Results: The mean age was 57.0 years old (standard deviation 13.6 years). A moderate correlation was
observed between mean Ki67 index and SUVmax (r = 0.392), SUVmean 30% (r = 0.437), and SUVmean 40%
(r = 0.443), and also between hotspot Ki67 index and SUVmax (r = 0.360), SUVmean 30% (r = 0.362) and
SUVmean 40% (r = 0.319). There was a weaker correlation, which was inversely negative, between mean
and hotspot Ki67 and volumetric PET parameters. However, no statistical significant differences were
found for any correlations.
Conclusions: SUVmax and SUVmean were moderately correlated with Ki67 index, whereas volumetric
PET parameters overall showed a weaker correlation. Thus, SUVmax and SUVmean could be used to assess
tumour aggressiveness in pre-treatment EOC.
© 2017 Elsevier España, S.L.U. y SEMNIM. All rights reserved.

Correlación de la captación de 18 F-FDG de la PET/TC con el Ki67 de la

inmunohistoquímica en el cáncer epitelial de ovario pretratamiento

r e s u m e n

Palabras clave: Objetivo: El standardized uptake value (SUV) y los parámetros volumétricos volumen metabólico tumoral
Cáncer epitelial de ovario (MTV) y glicólisis total de la lesión (TLG) de la 18 F-FDG PET/TC son útiles para determinar el pronóstico
18
F-FDG PET/TC preoperatorio y postratamiento del cáncer epitelial de ovario (CEO). El Ki67 es otro marcador pronóstico
SUV
en el CEO asociado con la agresividad tumoral. El objetivo fue estudiar la asociación entre los parámetros
Volumen metabólico tumoral
de la 18 F-FDG PET/TC y el Ki67 en el CEO pretratamiento para determinar si la PET/TC puede predecir la
Glicólisis total de la lesión
Ki67 agresividad tumoral de forma no invasiva.
Material y métodos: Se realizó una PET/TC a 18 pacientes con sospecha o diagnóstico reciente de CEO. Se
obtuvo el SUV máximo (SUVmáx), SUV medio (SUVmean) y el MTV y la TLG corporal (wbMTV y wbTLG,
respectivamente), con un dintel del 30-40% del SUVmáx. Se estimó el índice de Ki67 (medio y máximo)
en muestras del tejido tumoral, y se correlacionó con los parámetros de la PET.

夽 This work was supported by Premi Fi de Residència Emili Letang 2015 and AGAUR 2014 SGR 279 grants. This financial source did not have any participation in the study
design, data collection, analysis or interpretation of these in the writing of the manuscript or in the decision to submit it for publication.
∗ Corresponding author.
E-mail address: pparedes@clinic.cat (P. Paredes).
1
Both authors contributed equally to this work.

http://dx.doi.org/10.1016/j.remn.2017.07.005
2253-654X/© 2017 Elsevier España, S.L.U. y SEMNIM. All rights reserved.

Please cite this article in press as: Mayoral M, et al. Correlation of 18 F-FDG uptake on PET/CT with Ki67 immunohistochemistry in
pre-treatment epithelial ovarian cancer. Rev Esp Med Nucl Imagen Mol. 2017. http://dx.doi.org/10.1016/j.remn.2017.07.005
G Model
REMNIM-932; No. of Pages 7 ARTICLE IN PRESS
2 M. Mayoral et al. / Rev Esp Med Nucl Imagen Mol. 2017;xxx(xx):xxx–xxx

Resultados: La edad media fue 57,0 años (desviación estándar 13,6 años). Se observó una moderada
correlación entre el Ki67 medio y el SUVmáx (r = 0,392), SUVmean 30% (r = 0,437) y SUVmean 40%
(r = 0,443), así como entre el Ki67 máximo y el SUVmáx (r = 0,360), SUVmean 30% (r = 0,362) y SUVmean
40% (r = 0,319). La correlación fue más débil, e inversamente negativa, entre el Ki67 medio y máximo
y los parámetros volumétricos de la PET. No hubo diferencias estadísticamente significativas entre las
correlaciones.
Conclusiones: SUVmáx y SUVmean se correlacionaron moderadamente con el Ki67 mientras que los
parámetros volumétricos globalmente mostraron una correlación más débil. SUVmáx y SUVmean podrían
utilizarse para predecir la agresividad tumoral en el CEO pretratamiento.
© 2017 Elsevier España, S.L.U. y SEMNIM. Todos los derechos reservados.

Introduction
Ovarian cancer is the fourth most common gynaecological
malignancy and the fifth cause of death from cancer among women
in developed countries.1 Epithelial ovarian cancer (EOC) is the
most common histological subtype. Contrast-enhanced computed
tomography (CT) is the imaging procedure of choice for pretreat-
ment staging of EOC.2 However, functional techniques, comprising
diffusion-weighted magnetic resonance imaging (DW-MRI) and
18 F-fluorodeoxyglucose (18 F-FDG) positron emission tomography
(PET)/CT, are emerging imaging modalities that may overcome lim-
itations of CT. It has been reported that DW-MRI could be superior
to CT in the assessment of bowel serosal and mesenteric disease
as well as metastases outside the abdomen.3 On the other hand,
PET/CT has also shown to be more accurate than CT in the detection
of supradiaphragmatic metastases.4
In contrast with conventional morphological imaging tech-
niques, 18 F-FDG PET/CT combines both anatomical information
and tumour metabolic activity. As a glucose analogue, the radio-
pharmaceutical 18 F-FDG is taken up in metabolically active cells
and so it can be used clinically to detect tumour tissue. The
most commonly used parameter to estimate 18 F-FDG uptake is
the standardized uptake value (SUV). Given the high mortality
of EOC,1 some studies have evaluated if 18 F-FDG uptake could
be used as a non-invasive procedure to predict disease progno-
sis. High maximum SUV (SUVmax) values in the primary tumour
have been associated with poor outcome in patients with EOC.5 In
other malignancies, 18 F-FDG uptake has been correlated with anti-
gen Ki67 obtained by immunohistochemistry analysis, which is a
nuclear protein needed for cell proliferation and reflects tumour
aggressiveness.6 However, although Ki67 is a known prognostic
factor in EOC,7 its association with SUV has not yet been well
established in this malignancy. On the other hand, SUVmax reflects
the highest 18 F-FDG uptake within the tumour and may be not
representative of a heterogeneous hypermetabolic lesion. In the
recent years, volumetric parameters total lesion glycolysis (TLG)
and metabolic tumour volume (MTV) have proven to be superior
to SUVmax as prognostic factors in some neoplasms,8,9 as well as in
EOC.10–12 However, their association with biomarker Ki67 for any
malignancy is unknown to date. Considering the above conditions,
the aim of this study was to evaluate the association between Ki67
and 18 F-FDG PET/CT measurements SUVmax, SUVmean, MTV and
TLG in pretreatment EOC to determine whether PET/CT parameters
could non-invasively predict tumour aggressiveness.
Materials and methods
Patient population
This prospective study was approved by the hospital Ethics
Committee and written informed consent was obtained from
patients. Women with suspected or newly diagnosed EOC at our
centre during the time period comprised between December 2013
and April 2016 were consecutively enrolled and underwent an ini-
tial staging PET/CT study (n = 41). Exclusion criteria were 1) not
having EOC as a final diagnosis in patients in whom there was
only a suspicion of the disease and PET/CT was performed to evalu-
ate an annexal mass (n = 8); 2) having undergone any treatment
previous to staging PET/CT, that is, neoadjuvant chemotherapy
or cytoreductive surgery (n = 1); 3) pretreatment Ki67 was not
obtained, as patients directly proceed to neoadjuvant therapy after
PET/CT without obtaining a tumoural biopsy, only a cytology was
obtained to diagnose EOC (n = 13); and 4) pretreatment Ki67 could
not be obtained, as tumoural biopsy was performed and analyzed in
another centre (n = 1). Finally, 18 patients met the eligibility crite-
ria. The following clinical data was recorded for each patient: age at
the time of diagnosis of EOC, histological subtype and International
Federation of Gynaecology and Obstetrics (FIGO) staging.
PET/CT simulation study
MTV is an estimation of the volume of an uptake within a manu-
ally drawn region of interest (ROI), most usually a hypermetabolic
tumoural lesion. The segmentation of MTV is based on a fixed
SUVmax threshold, that is, all voxels with the same uptake or higher
than a selected SUVmax percentage are included in the volume
estimation. MTV value may be subject to operator-dependent con-
touring variations when drawing the ROI, PET/CT equipment, and it
also may vary depending on tumour to background uptake ratio (if
background uptake is similar to tumoural uptake more voxels are
wrongly included in the volume estimation). For this reason, we
aimed to determine which SUVmax percentage should be used as a
threshold in our study to estimate the volume of tumoural lesions
more precisely. We used NEMA IEC Body phantom (Fig. 1) that con-
tains six spheres of different volumes (0.5, 1.2, 2.6, 5.6, 11.5 and
26.5 mL). Four SUVmax thresholds were used (20%, 30%, 40% and
50% of the SUVmax). Additionally, three activity concentrations,
similar to those found in tumoural lesions in patients with EOC,
were placed in each sphere in three separate time interval acquisi-
tions, with the same background activity concentration in all three
cases to obtain three different tumour to background uptake ratios.
PET/CT acquisition parameters for the simulation study were the
same than the routinely used in clinical practice at our centre.
PET/CT technique and image interpretation
After a 6-h fasting period to achieve blood glucose concentra-
tions below 140 mg/dL, patients were required to rest quietly in a
dimly lit room during the 60 min following 18 F-FDG intravenous
administration of 4.07 MBq/kg (0.11 mCi/kg). PET/CT scans were
performed using a hybrid PET/TC Biograph mCT TrueV (Siemens,
USA) for attenuation correction and image fusion. PET/CT imaging
was acquired from the skull base to the proximal third thigh with
arms raised above the head. Iodine-based oral and intravenous

Please cite this article in press as: Mayoral M, et al. Correlation of 18 F-FDG uptake on PET/CT with Ki67 immunohistochemistry in
pre-treatment epithelial ovarian cancer. Rev Esp Med Nucl Imagen Mol. 2017. http://dx.doi.org/10.1016/j.remn.2017.07.005
G Model
REMNIM-932; No. of Pages 7 ARTICLE IN PRESS
M. Mayoral et al. / Rev Esp Med Nucl Imagen Mol. 2017;xxx(xx):xxx–xxx 3

A B

C D

Fig. 1. Phantom acquisition. Axial CT (A), fused PET/CT (B) and PET (C) images of the phantom (D) used in the simulation study. This phantom contains 6 spheres of different
volumes (diameters): 0.5 mL (10 mm), 1.2 mL (13 mm), 2.6 mL (17 mm), 5.6 mL (22 mm), 11.5 mL (28 mm) and 26.5 mL (37 mm).

contrasts were used. PET data reconstruction was performed with
a standard iterative algorithm from CT records.
PET/CT images were assessed using a workstation
(syngo.PET&CT Oncology, Siemens) by two experienced nuclear
medicine physicians. Any discrepancies were overcome by
consensus. Spherical or ellipsoidal ROI were placed over all
the pathological uptakes on attenuation-corrected PET images,
assuring the enclosure of the entire hypermetabolic lesion on
axial, sagittal, and coronal projections. SUVmax and mean SUV
(SUVmean) were then automatically calculated based on measured
activity concentration (Bq/mL) multiplied by patient weight (kg)
divided by injected activity (Bq). The highest SUVmax of all the
malignant lesions was taken into account for analysis. To obtain
MTV values, the contouring margins of each lesion were delimited
by the thresholds of SUVmax which best estimated the volume of
the spheres of the phantom. The sum of the MTV of every lesion
was defined as the whole-body MTV (wbMTV) and whole-body
TLG (wbTLG) was the summation of the MTV of each lesion
multiplied by its SUVmean.
Immunohistochemistry
The rate of cell proliferation was estimated in tumoural
tissue specimens obtained from diagnostic biopsy or cytore-
ductive surgery. Tumoural samples were fixed in 10% neutral
buffered formalin, dehydrated in alcohol, cleared in xylene
and embedded in paraffin. Blocks were sectioned at 3-␮m
thickness slices using a microtome, previously pretreated with
3-aminopropyltriethoxysilane. Slides were then deparaffinized in
three consecutive 5-min xylene baths and hydrated through suc-
cessive baths of ethanol at decreasing concentrations (99%, 96%
and 70%). After rehydration, slides were immersed in a distilled
water bath or a PBS Buffer, then pretreated in PT-Link pH9 at 95 ◦ C
for 20 min and cleared in Buffer for 5 min. Afterwards, peroxidase
inhibition was performed for 5 min and cleared in Buffer. Ready-
to-use antigen Ki67 and amplification polymer were consecutively
applied for 20 min each and slides were cleared in Buffer after both
steps. Then, slides were developed in diaminobenzidine for 5 min
with distilled water, contrasted in haematoxylin for 2 min, dehy-
drated and mounted in DPX. Finally, slides were analyzed at the
microscope by a pathologist. All coloured nucleus were considered
positive, independently of the intensity and distribution, and a per-
centage was given (mean and hotspot Ki67 index) using manual
count.
Statistical analysis
Absolute and percentage frequencies were used to illustrate the
distribution of histological subtype and FIGO staging and mean
value and standard deviation (SD) were calculated for PET param-
eters. MedCalc statistical software (version 17.2) was used to
perform the statistical analysis. PET parameters were compared
with histological subtypes using Kruskal–Wallis test (mixed sub-
types was excluded from the analysis). Correlations between Ki67
and SUVmax, SUVmean 30%, SUVmean 40%, wbMTV 30%, wbMTV
40%, wbTLG 30% and wbTLG 40% were assessed with Spearman’s
rho test. Correlation was considered low for r > 0 and r < 0.3, moder-
ate for r > 0.31 and r < 0.5, strong for r > 0.51 and r < 0.8 and excellent
for r > 0.81. Significant differences were considered to exist when
the p value was less than 0.05.
Results
PET/CT simulation study
The relative error in volume estimation for the three tumour
to background uptake ratios when using each SUVmax threshold
is shown in Fig. 2. The volume of the spheres was overestimated
when threshold 20% of the SUVmax was used, whereas thresholds
40% and 50% had a tendency to underestimate their volume. The
most precise estimation was obtained when threshold 30% was
used, especially for ratios 2 and 3, the ones with the least contrasted
tumour to background uptake. The relative error in volume estima-
tion was higher for the three spheres with the smallest diameters
(10, 13 and 17 mm) when using any of the four thresholds.

Please cite this article in press as: Mayoral M, et al. Correlation of 18 F-FDG uptake on PET/CT with Ki67 immunohistochemistry in
pre-treatment epithelial ovarian cancer. Rev Esp Med Nucl Imagen Mol. 2017. http://dx.doi.org/10.1016/j.remn.2017.07.005
G Model
REMNIM-932; No. of Pages 7 ARTICLE IN PRESS
4 M. Mayoral et al. / Rev Esp Med Nucl Imagen Mol. 2017;xxx(xx):xxx–xxx

A Threshold 20%
B Threshold 30%
450 350
Relative error in volume (%)

Relative error in volume (%)

400 300
350 250
300
Ratio 1 200 Ratio 1
250
Ratio 2 150 Ratio 2
200
Ratio 3 100 Ratio 3
150
100 50
50 0
0 10 20 30 40
0 -50
0 10 20 30 40
Sphere diameter (mm) Sphere diameter (mm)
C Threshold 40%
D Threshold 50%
100 40
Relative error in volume (%)

Relative error in volume (%)

80
20
60
0
40 Ratio 1 0 10 20 30 40 Ratio 1
20 Ratio 2 -20 Ratio 2
0 Ratio 3 Ratio 3
0 10 40 -40
20 30
-20
-60
-40
-60 -80
Sphere diameter (mm) Sphere diameter (mm)

Fig. 2. Volume estimation in the simulation study. Relative error in volume estimation obtained with a threshold of 20% (A), 30% (B), 40% (C) and 50% (D) of the SUVmax. From
ratio 1–3 the uptake contrast between tumour and background tissues decreased. Sphere diameters were 10, 13, 17, 22, 28 and 37 mm. The threshold which best estimated
the real volume of the spheres was 30% of SUVmax (B).

Table 1 were associated with lower SUVmean 30% (mean 2.73, SD 0.13)
Age, histological subtype and FIGO staging (overall values).
in contrast with endometrioid carcinoma (unique value of 6.01)
Variable Value and serous adenocarcinomas (mean 8.50, SD 3.25) that had higher
Age, years values of this parameter, and these differences were statistically
Mean 57.0 significant. However, no statistically significant differences were
Standard deviation 13.6 found between the distribution of SUVmax, SUVmean 40% and
Histological subtype, n (%) volumetric parameters in relation to the histological subtype.
Serous adenocarcinoma 14 (77.7) The correlation between Ki67 index and PET parameters is
Clear cell carcinoma 2 (11.1) shown in Table 4. A moderate correlation was observed between
Endometrioid 1 (5.6)
mean Ki67 index and SUVmax (r = 0.392), SUVmean 30% (r = 0.437)
Mixed 1 (5.6)
and SUVmean 40% (r = 0.443), and also between hotspot Ki67 index
FIGO staging, n (%) and SUVmax (r = 0.360), SUVmean 30% (r = 0.362) and SUVmean
I 3 (16.7)
40% (r = 0.319). Fig. 3 illustrates the correlation obtained between
II 0 (0.0)
IIIC 7 (38.9) mean Ki67 index and SUVmax. There was a weaker correlation,
IV 8 (44.4) which was inversely negative, between mean and hotspot Ki67 and
volumetric PET parameters. Only wbMTV 30% obtained a moder-
ate correlation between mean (r = −0.315) and hotspot Ki67 index
According to these findings, we chose to use threshold 30% of (r = −0.342). However, no statistical significant differences were
the SUVmax in PET/CT studies performed to evaluate our patients. found for any correlations.
However, although it was less precise, we also decided to use
threshold 40% as it is the most used percentage in previous articles
of the literature. Discussion

Tumour characteristics, PET parameters and Ki67 values Ovarian cancer is one of the most lethal neoplasies. Therefore, it
is crucial to obtain a precise staging of the disease and identify its
The mean age was 57.0 years old (SD 13.6 years). The distri- prognostic factors in order to provide the most appropriate treat-
bution of histological subtypes and FIGO staging and the values ment. In this study we aimed to evaluate the association between
obtained for PET parameters and Ki67 index are shown in Table 1 Ki67 index, a known prognostic factor in EOC,7 and 18 F-FDG uptake
and Table 2. Most patients belonged to stage IIIC and IV according measured by pretreatment PET/CT to determine whether PET
to the FIGO staging system (46.7%). The most common tumour type parameters could non-invasively predict tumour aggressiveness.
was serous adenocarcinoma (77.7%). The mean values of SUVmax, Different measurements of radioactivity concentration can be
SUVmean 30%, wbMTV 30%, wbTLG 30%, SUVmean 40%, wbMTV used to estimate metabolic activity in PET/CT studies although
40% and wbTLG 40% were 14.67 (SD 5.75), 7.57 (SD 3.44), 681.37 mL none of them is totally precise. SUVmax, despite being the most
(SD 372.18 mL), 3225.42 g (SD 2128.14 g), 8.75 (SD 3.86), 474.54 mL commonly used parameter, may not be representative of a het-
(SD 285.84 mL) and 2507.19 g (SD 1752.10 g), respectively. The erogeneous uptake as it depicts the maximum activity obtained
mean values of mean and hotspot Ki67 index were 35.05% (SD within a ROI. SUVmean, which represents the average radioactivity
19.53%) and 60.37% (SD 24.51%), respectively. concentration, may be subject to operator-dependent contouring
SUVmean 30% values were differently distributed depending on variations. Volumetric PET parameters MTV and TLG are emerging
the histological subtype, as shown in Table 3. Clear-cell carcinomas measurements that estimate the volume of metabolically active

Please cite this article in press as: Mayoral M, et al. Correlation of 18 F-FDG uptake on PET/CT with Ki67 immunohistochemistry in
pre-treatment epithelial ovarian cancer. Rev Esp Med Nucl Imagen Mol. 2017. http://dx.doi.org/10.1016/j.remn.2017.07.005
G Model
REMNIM-932; No. of Pages 7 ARTICLE IN PRESS
M. Mayoral et al. / Rev Esp Med Nucl Imagen Mol. 2017;xxx(xx):xxx–xxx 5

Table 2
Age, histological subtype, FIGO staging, PET parameters and Ki67 index (per patient and mean values).

No. Age Histology FIGO SUVmax SUVmean30% wbMTV30% (mL) wbTLG30% (g) SUVmean40% wbMTV40% (mL) wbTLG40% (g) Ki67 (%)a Ki67(%)b

1 54 SA IV 23.88 13.98 630.74 7418.11 14.80 448.48 6013.27 26.5 49.0

2 62 SA IV 14.93 7.18 599.00 4103.74 8.28 433.75 3338.96 25.4 29.2
3 47 SA IIIC 16.70 8.96 1098.49 6446.64 9.84 766.29 5129.78 38.5 73.0
4 66 SA IIIC 12.82 6.14 911.12 3014.82 7.22 604.33 2268.71 33.6 45.5
5 37 SA IIIC 15.86 8.28 541.88 2184.74 13.09 392.37 1805.81 48.0 84.2
6 50 SA IIIC 23.87 14.54 575.22 4636.72 16.36 376.52 3467.31 9.1 22.5
7 74 SA IV 14.42 7.60 498.73 1822.38 8.93 328.87 1334.28 41.3 87.0
8 68 SA IIIC 13.45 6.94 668.46 2582.85 7.96 487.38 2081.87 42.8 70.9
9 19 SA IIIC 7.85 3.66 733.76 1950.78 4.38 515.54 1525.00 14.9 33.3
10 58 SA IV 12.17 6.46 199.99 701.31 7.16 146.46 562.60 39.5 54.4
11 70 SA IV 18.56 8.99 345.19 2440.70 10.23 290.22 2177.72 79.0 95.9
12 64 Mixedc IC3 14.79 5.80 512.02 2882.44 7.01 186.71 1262.08 18.6 60.5
13 69 CCC IC2 4.33 2.82 1548.59 2468.29 2.82 1176.26 2027.61 5.6 24.8
14 45 CCC IC2 4.65 2.64 171.14 389.96 3.01 130.28 326.61 19.7 45.0
15 60 SA IIIC 24.42 12.94 534.27 2125.17 14.31 406.42 1738.16 74.9 95.6
16 55 SA IV 16.37 8.26 735.83 3249.33 8.72 448.75 2404.30 39.6 67.3
17 66 E IV 12.84 6.01 1461.12 7912.39 7.38 1089.77 6458.70 34.7 55.6
18 61 SA IV 12.17 5.02 499.08 1727.18 6.01 313.26 1206.64 39.2 93.0

Mean 14.67 7.57 681.37 3225.42 8.75 474.54 2507.19 35.05 60.37
Standard deviation 5.75 3.44 372.18 2128.14 3.86 285.84 1752.10 19.53 24.51

SA = serous adenocarcinoma. CCC = clear cell carcinoma. E = endometrioid.

a
Mean Ki67 index.
b
Hotspot Ki67 index.
c
80% Serous adenocarcinoma and 20% clear cell carcinoma.

Table 3 showed the highest correlation coefficient (r = 0.81).14,15 In our

Comparison between histological subtypes and PET parameters.
study we obtained a moderate positive correlation between mean
Histological subtype vs p Ki67 index and SUVmax (r = 0.392), SUVmean 30% (r = 0.437) and
SUVmax 0.068 SUVmean 40% (r = 0.443), and also between hotspot Ki67 index
SUVmean 30% *
0.049 and SUVmax (r = 0.360), SUVmean 30% (r = 0.362) and SUVmean
wbMTV 30% 0.357 40% (r = 0.319). According to Deng et al.’s meta-analysis,6 52
wbTLG 30% 0.152 and 12 studies used SUVmax and SUVmean, respectively, as
SUVmean 40% 0.068
wbMTV 40% 0.357
the PET parameter to be associated with Ki67 index (mean or
wbTLG 40% 0.129 hotspot). Both PET measurements obtained a moderate corre-
* lation with r = 0.41 for SUVmax and r = 0.51 for SUVmean. We
p < 0.05.
also found a similar correlation coefficient between mean and
hotspot Ki67 index and SUVmax (r = 0.392 and r = 0.360, respec-
Table 4 tively), SUVmean 30% (r = 0.437 and r = 0.362, respectively) and
Correlation between Ki67 index and PET parameters.
SUVmean 40% (r = 0.443 and r = 0.319, respectively). However, in
Ki67 r p our study, hotspot Ki67 index globally obtained a lower cor-
Mean Ki67 vs
relation coefficient with PET parameters than mean Ki67 index
SUVmax 0.392 0.11 , and these findings are in discordance with Deng et al.’s meta-
SUVmean 30% 0.437 0.07 analysis in which PET parameters showed a stronger correlation
wbMTV 30% −0.315 0.20 with hotspot Ki67 (r = 0.48) than mean Ki67 index (r = 0.40).
wbTLG 30% −0.228 0.36
On the other hand, we also studied the association between
SUVmean 40% 0.443 0.07
wbMTV 40% −0.199 0.43 Ki67 index and volumetric PET parameters which have shown
wbTLG 40% −0.071 0.78 to be superior to SUVmax for disease prognostication in various
Hotspot Ki67 vs
neoplasies.8–12 We found a low negative correlation between vol-
SUVmax 0.360 0.14 umetric PET parameters and both mean and hotspot Ki67 index.
SUVmean 30% 0.362 0.17 Only wbMTV 30% obtained a moderate correlation between mean
wbMTV 30% −0.342 0.17 (r = −0.315) and hotspot Ki67 index (r = −0.342). Thus, these mea-
wbTLG 30% −0.253 0.31
surements would not seem to be useful to predict disease prognosis
SUVmean 40% 0.319 0.20
wbMTV 40% −0.251 0.32 in patients with EOC. These findings may have a theoretical expla-
wbTLG 40% −0.183 0.47 nation after interpreting the results of the simulation study. The
relative error in the estimated volume was higher for the three
spheres with the smallest diameters (10, 13 and 17 mm) when
tumour. However, MTV is based on a fixed SUVmax threshold and any of the four thresholds was used (20%–50%). In PET/CT stud-
TLG is additionally contingent upon SUVmean. ies performed in patients these findings would represent a worse
A recent meta-analysis that included 81 studies involving a volume estimation of those tumoural lesions approximately below
total of 3242 patients with different tumours reported that 18 F- 1.7 cm (the diameter of the third sphere with the smallest diam-
FDG uptake measured by SUVmax and SUVmean had a moderate eter), a lesion size characteristically found in EOC (i.e. peritoneal
positive correlation with tumour cell proliferation with a com- carcinomatosis). Our research is the first study in which the asso-
bined r of 0.44.6 The degree of correlation varied depending on ciation between Ki67 index and volumetric PET parameters in
tumour lineage. The slightest correlation was observed in malig- patients with EOC was evaluated. To the best of our knowledge,
nant melanoma (r = −0.22),13 whereas thymic epithelial tumours this association has only been studied in breast16 and lung cancer.17

Please cite this article in press as: Mayoral M, et al. Correlation of 18 F-FDG uptake on PET/CT with Ki67 immunohistochemistry in
pre-treatment epithelial ovarian cancer. Rev Esp Med Nucl Imagen Mol. 2017. http://dx.doi.org/10.1016/j.remn.2017.07.005
G Model
REMNIM-932; No. of Pages 7 ARTICLE IN PRESS
6 M. Mayoral et al. / Rev Esp Med Nucl Imagen Mol. 2017;xxx(xx):xxx–xxx

A B

C D

E F

Fig. 3. Correlation between mean Ki67 index and SUVmax. Three examples which illustrate the correlation between mean Ki67 index and SUVmax. The lesion with the
highest SUVmax value of all the malignant lesions is depicted with the green arrow. The first case is a 19-year-old woman with stage IIIC serous adenocarcinoma (Table 1,
patient 9) whose mean Ki67 index (A) and SUVmax (B) values were 14.9% and 7.85 respectively, both below the mean value of each parameter. The second case is a 60-year-old
woman with stage IIIC serous adenocarcinoma (Table 1, patient 15) whose mean Ki67 index (C) and SUVmax (D) values were 74.9% and 24.42 respectively, both higher than
the mean value of each parameter. The third patient is an example of a discordant case. This patient is a 50-year-old woman with stage IIIC serous adenocarcinoma (Table 1,
patient 6) whose mean Ki67 index (E) and SUVmax (F) values were 9.1% and 23.87, respectively. The mean Ki67 index was below its mean value whereas SUVmax was higher
than its mean value.

Kajáry et al.’s study found similar results in their series than in our
research, in which Ki67 index had a stronger association with SUV
than volumetric parameters. In Kajáry et al.’s article a moderate cor-
relation was observed between Ki67 index and SUVmax (r = 0.556)
and SUVmean (r = 0.426) and a lower correlation between Ki67
index and MTV (r = 0.313) and TLG (r = 0.348).16 On the contrary,
Park et al. reported that Ki67 index had a stronger association with
volumetric measurements wbMTV (r = 0.254) and wbTLG (r = 0.239)
in comparison with SUVmax (r = 0.116) and SUVmean (r = 0.031).17
Our study is also one of the firsts to describe the association
between SUV and Ki67 in epithelial ovarian tumours. The arti-
cle by Kurokawa et al. is the only previous published research on
this topic.18 These authors found a moderate association between
SUVmean and hotspot Ki67 with r = 0.457, a stronger correla-
tion than in our study as we obtained r = 0.362 and r = 0.319
for SUVmean 30% and SUVmean 40% respectively versus hotspot
Ki67 index. However, 4 out of 17 patients included in Kurokawa
et al. study had borderline or benign tumours which commonly
have lower 18 F-FDG uptake in comparison with advanced stage
EOC. The heterogeneity of the included patients may condition a
higher contrast between 18 F-FDG uptakes and this may increase
the strength of the correlation coefficient. We found the maximum
correlation when associating SUVmean 40% and mean Ki67 index
(r = 0.443).
It is also noteworthy that we found a statistically significant
different distribution of SUVmean 30% depending on the histologi-
cal subtype (p = 0.049). Clear-cell carcinomas had tumoural lesions
with lower SUVmean 30% values than serous adenocarcinomas and
endometrioid carcinomas. These findings are in accordance with
Tanizaki et al.’s study.19 The authors also obtained lower 18 F-FDG
uptake in patients with clear cell carcinoma in relation to serous and
endometrioid carcinomas (p < 0.01). However, they used SUVmax
to evaluate 18 F-FDG uptake while we did not find a statistically
significant difference in the distribution of this parameter depend-
ing on the histological subtype (p = 0.068). Comparably with our
results, Karantanis et al. concluded that SUVmax is unrelated to his-
tological subtype.20 However, this study only included two patients
with clear cell carcinoma (n = 42), the same number than in our

Please cite this article in press as: Mayoral M, et al. Correlation of 18 F-FDG uptake on PET/CT with Ki67 immunohistochemistry in
pre-treatment epithelial ovarian cancer. Rev Esp Med Nucl Imagen Mol. 2017. http://dx.doi.org/10.1016/j.remn.2017.07.005
G Model
REMNIM-932; No. of Pages 7 ARTICLE IN PRESS
M. Mayoral et al. / Rev Esp Med Nucl Imagen Mol. 2017;xxx(xx):xxx–xxx
research (n = 18), while 11 patients with this histological subtype
were included in Tanizaki et al.’s study (n = 67).
The main limitation of this study is the small sample size,
which may explain the absence of statistically significant differ-
ences between correlations. However, as mentioned above, we
included a higher number of patients than Kurokawa et al.’s study
(n = 13) if only patients with cancer are considered.18 A strength
of our study is being the second article to evaluate the correlation
between cell proliferation index and SUV in EOC and the first to
describe the association between Ki67 index and volumetric PET
parameters in this neoplasm. Our findings should be further eval-
uated in future studies with bigger sample sizes.
A new prognostic stratification system incorporating volume-
based PET parameters could help to select patients who would
benefit from cytoreductive surgery from those who would do better
with adjuvant chemotherapy. Additionally, one potential clinical
use of volume-based assessment by 18 F-FDG PET/CT could be to
identify appropriate patients for risk-adapted treatment. Consid-
ering the high rate of recurrence for citorreductive surgery in EOC,
patients with high PET parameter values could be closely moni-
tored postoperatively and enrolled in prospective clinical trials of
novel treatment approaches.
Conclusions
SUVmax and SUVmean (30% and 40%) were moderately associ-
ated with mean and hotspot Ki67 index whereas volumetric PET
parameters wbMTV and wbTLG globally showed a weaker correla-
tion. Thus, SUVmax and SUVmean could be used to assess tumour
aggressiveness in pretreatment EOC while volumetric PET param-
eters would not seem to be useful in this situation. Given the
increasing role of molecular imaging for personalized cancer treat-
ment, our findings merit to be explored in prospective studies and
validated with a larger population of EOC patients.
Conflicts of interest
The authors declare no conflicts of interest.
References
1. Siegel RL, Miller KD, Jemal A. Cancer Statistics, 2016. CA Cancer J Clin.
2016;66:7–30.
2. Morgan RJ, Armstrong DK, Alvarez RD, Bakkum-Gamez JN, Behbakht K, Chen
L-M, et al. Ovarian cancer including fallopian tube cancer and primary peri-
toneal cancer. In: NCCN clinical practice guidelines in oncology 2015; version 2;
2015.
3. Michielsen K, Vergote I, Op de Beeck K, Amant F, Leunen K, Moerman P, et al.
Whole-body MRI with diffusion-weighted sequence for staging of patients with
Please cite this article in press as: Mayoral M, et al. Correlation of 18 F-FDG uptake on PET/CT with Ki67 immunohistochemistry in
pre-treatment epithelial ovarian cancer. Rev Esp Med Nucl Imagen Mol. 2017. http://dx.doi.org/10.1016/j.remn.2017.07.005
7
suspected ovarian cancer: a clinical feasibility study in comparison to CT and
FDG-PET/CT. Eur Radiol. 2014;24:889–901.
4. Schmidt S, Meuli RA, Achtari C, Prior JO. Peritoneal carcinomatosis in primary
ovarian cancer staging: comparison between MDCT, MRI, and 18F-FDG PET/CT.
Clin Nucl Med. 2015;40:371–7.
5. Nakamura K, Hongo A, Kodama J, Hiramatsu Y. The pretreatment of maximum
standardized uptake values (SUVmax) of the primary tumor is predictor for
poor prognosis for patients with epithelial ovarian cancer. Acta Med Okayama.
2012;66:53–60.
6. Deng S, Zhang W, Zhang B, Chen Y-Y, Li J-H, Wu Y-W. Correlation between
the uptake of 18F-fluorodeoxyglucose (18F-FDG) and the expression of
proliferation-associated antigen Ki-67 in cancer patients: a meta-analysis. PLOS
ONE. 2015;44:1–18.
7. Garzetti GG, Ciavattini A, Goteri G, De Nictolis M, Stramazzoti D, Lucarini G, et al.
Ki67 antigen immunostaining (MIB 1 monoclonal antibody) in serous ovarian
tumors: index proliferative activity with prognostic significance. Gynecol Oncol.
1995;56:169–74.
8. Im HJ, Pak K, Cheon GJ, Kang KW, Kim SJ, Kim IJ, et al. Prognostic value of volumet-
ric parameters of (18)F-FDG PET in non-small-cell lung cancer: a meta-analysis.
Eur J Nucl Med Mol Imaging. 2015;42:241–51.
9. Pak K, Cheon GJ, Nam HY, Kim S-J, Kang KW, Chung JK, et al. Prognostic value of
metabolic tumor volume and total lesion glycolysis in head and neck cancer: a
systematic review and meta-analysis. J Nucl Med. 2014;55:884–90.
10. Chung HH, Kwon HW, Kang KW, Park NH, Song Y-S, Chung JK, et al. Prognostic
value of preoperative metabolic tumor volume and total lesion glycolysis in
patients with epithelial ovarian cancer. Ann Surg Oncol. 2012;19:1966–72.
11. Liao S, Lan X, Cao G, Yuan H, Zhang Y. Prognostic predictive value of total lesion
glycolysis from 18F-FDG PET/CT in post-surgical patients with epithelial ovarian
cancer. Clin Nucl Med. 2013;38:715–20.
12. Lee JW, Cho A, Lee JH, Yun M, Lee JD, Kim YT, et al. The role of metabolic tumor vol-
ume and total lesion glycolysis on 18 F-FDG PET/CT in the prognosis of epithelial
ovarian cancer. Eur J Nucl Med Mol Imaging. 2014;41:1898–906.
13. Park S, Lee J, Lee W, Han KM. Biologic correlation between glucose transporters,
hexokinase-II, Ki-67 and FDG uptake in malignant melanoma. Nucl Med Biol.
2012;39:1167–72.
14. Matsumoto I, Oda M, Takizawa M, Waseda R, Nakajima K, Kawano M, et al.
Usefulness of fluorine-18 fluorodeoxyglucose-positron emission tomography
in management strategy for thymic epithelial tumors. Ann Thorac Surg.
2013;95:305–10.
15. Viti A, Bertolaccini L, Cavallo A, Fortunato M, Bianchi A, Terzi A. 18-Fluorine
fluorodeoxyglucose positron emission tomography in the pretreatment evalu-
ation of thymic epithelial neoplasms: a metabolic biopsy confirmed by Ki-67
expression. Eur J Cardiothorac Surg. 2014;46:369–74.
16. Kajáry K, Tőkés T, Dank M, Kulka J, Szakáll S, Lengyel Z. Correlation of the value of
18
F-FDG uptake, described by SUVmax, SUVavg, metabolic tumour volume and
total lesion glycolysis, to clinicopathological prognostic factors and biological
subtypes in breast cancer. Nucl Med Commun. 2015;36:28–37.
17. Park S, Lee E, Rhee S, Cho J, Choi S, Lee S, et al. Correlation between semi-
quantitative (18)F-FDG PET/CT parameters and Ki-67 expression in small cell
lung cancer. Nucl Med Mol Imaging. 2016;50:24–30.
18. Kurokawa T, Yoshida Y, Kawahara K, Tsuchida T, Okazawa H, Fujibayashi
Yasuhisa, et al. Expression of GLUT-1 glucose transfer, cellular proliferation
activity and grade of tumor correlate with [F-18]-fluorodeoxyglucose uptake
by positron emission tomography in epithelial tumors of the ovary. Int J Cancer.
2004;109:926–32.
19. Tanizaki Y, Kobayashi A, Shiro M, Ota N, Takano R, Mabuchi Y, et al. Diagnos-
tic value of preoperative SUVmax on FDG-PET/CT for the detection of ovarian
cancer. Int J Gynecol Cancer. 2014;24:454–60.
20. Karantanis D, Allen-Auerbach M, Czernin J. Relationship among glycolytic phe-
notype, grade, and histological subtype in ovarian carcinoma. Clin Nucl Med.
2012;37:49–53.