FDG-PET. A possible prognostic factor in head and neck cancer.

Abstract

Previous studies have shown that high uptake of (18)F-fluoro-2-deoxy-glucose in head and neck cancer, as determined by the standardized uptake value on positron emission tomography scan, was associated with poor survival. The aim of this study was to confirm the association and to establish whether a high standardized uptake value had prognostic significance. Seventy-three consecutive patients with newly diagnosed squamous cell carcinoma of the head and neck underwent a positron emission tomography study before treatment. Age, gender, performance status tumour grade, stage, maximal tumour diameter and standardized uptake value were analyzed for their possible association with survival. The median standardized uptake value for all primary tumours was 7.16 (90% range 2.30 to 18.60). In univariate survival analysis the cumulative survival was decreased as the stage, tumour diameter and standardized uptake value increased. An standardized uptake value of 10 was taken as a cut-off for high and low uptake tumours. When these two groups were compared, an standardized uptake value >10 predicted for significantly worse outcome (P=0.003). Multivariate analysis demonstrated that an standardized uptake value >10 provided prognostic information independent of the tumour stage and diameter (P=0.002). We conclude that high FDG uptake (standardized uptake value>10) on positron emission tomography is an important marker for poor outcome in primary squamous cell carcinoma of the head and neck. Standardized uptake value may be useful in distinguishing those tumours with a more aggressive biological nature and hence identifying patients that require intensive treatment protocols including hyperfractionated radiotherapy and/or chemotherapy.

Full text

Page 1

FDG–PET. A possible prognostic factor in head and neck cancer
W Halfpenny1, SF Hain2, L Biassoni2, MN Maisey2, JA Sherman1and M McGurk*,1
1Department of Oral and Maxillofacial Surgery, Guy’s and St Thomas’s Hospitals, St Thomas’s Street, London SE1 9RT, UK;2Clinical PET Centre, Guy’s and St
Thomas’s Hospitals, St Thomas’s Street, London SE1 9RT, UK
Previous studies have shown that high uptake of18F-fluoro-2-deoxy-glucose in head and neck cancer, as determined by the
standardized uptake value on positron emission tomography scan, was associated with poor survival. The aim of this study was
to confirm the association and to establish whether a high standardized uptake value had prognostic significance. Seventy-
three consecutive patients with newly diagnosed squamous cell carcinoma of the head and neck underwent a positron
emission tomography study before treatment. Age, gender, performance status tumour grade, stage, maximal tumour diameter
and standardized uptake value were analyzed for their possible association with survival. The median standardized uptake value
for all primary tumours was 7.16 (90% range 2.30 to 18.60). In univariate survival analysis the cumulative survival was
decreased as the stage, tumour diameter and standardized uptake value increased. An standardized uptake value of 10 was
taken as a cut-off for high and low uptake tumours. When these two groups were compared, an standardized uptake value
410 predicted for significantly worse outcome (P=0.003). Multivariate analysis demonstrated that an standardized uptake
value 410 provided prognostic information independent of the tumour stage and diameter (P=0.002). We conclude that high
FDG uptake (standardized uptake value410) on positron emission tomography is an important marker for poor outcome in
primary squamous cell carcinoma of the head and neck. Standardized uptake value may be useful in distinguishing those
tumours with a more aggressive biological nature and hence identifying patients that require intensive treatment protocols
including hyperfractionated radiotherapy and/or chemotherapy.
British Journal of Cancer (2002) 86, 512–516. DOI: 10.1038/sj/bjc/6600114
ª 2002 Cancer Research UK
www.bjcancer.com
Keywords: head and neck cancer; squamous cell carcinoma; positron emission tomography; survival
Squamous cell carcinomas (SCC) of the head and neck are a
distinct group of neoplasms with an unpredictable clinical beha-
viour. The stage of the disease is currently the most important
prognostic parameter and depends partly on local invasion but
more to cervical node status. The biology of each tumour is impor-
tant to outcome, for even in advanced disease (stages 3 and 4) a
third of patients can be cured. However, at present these patients
cannot be distinguished.
In positron emission tomography (PET) the most commonly
used radiopharmaceutical is
an analogue of glucose, which is metabolized in normal, and
neoplastic tissues in proportion to the rate of tissue metabolism.
This tracer is particularly useful in Oncology (Conti and Strauss,
1991; Minn et al, 1995), because it exploits the increased glucose
metabolism present in the majority of malignant tumours
(Warburg, 1956).
FDG–PET has already been applied to staging and follow-up of
patients with head and neck cancer (Rege et al, 1994; Braams et al,
1995), and Minn et al (1997) have shown that high uptake as
determined by the standardized uptake value (SUV) was associated
with poor survival. However it is not clear whether a high SUV
simply reflects the stage of disease or is an independent predictor
of survival. The aim of this study was to investigate the relationship
of SUV values to outcome in head and neck cancer.
18F-fluoro-2-deoxy-glucose (FDG),
MATERIALS AND METHODS
Patients and treatment
Between February 1993 and September 1998, 73 consecutive patients
with newly diagnosed Squamous cell carcinoma of the head and neck
underwent an FDG–PET evaluation before treatment. Diabetic
patients (eight patients) and those with incomplete data (seven
patients) were excluded from the study. Diabetics were excluded
because of their altered glucose metabolism and its effect on SUV
(Hallett et al, 2001). Incomplete data included stage and/or SUV leav-
ing a total of 58 patients suitable for analysis.
All patients had panendoscopy and CT scans of the head and
neck and were staged according to the UICC TNM classification
(Hermanek and Sobin, 1992). The average diameter of the primary
tumour was calculated from clinical inspection and the CT scan.
Patient characteristics are listed in Table 1.
All patients were treated with curative intent. Fifty-three patients
(91%) underwent surgery as the primary treatment modality, and of
these patients 75% received postoperative radiotherapy. Five patients
received radiotherapy alone. Of the patients undergoing surgery 38
hadlocalexcisionwithaneckdissectionandfreetissuetransferrecon-
struction, 10 local excision and a neck dissection and five local
excision alone. Patients were followed up at 2 monthly intervals with
a median follow-up of 39 months (range 4–75 months).
FDG–PET study
The patient fasted for 6 h prior to the PET study and a blood
glucose evaluation proceeded the test. FDG was synthesized with
Clinical
Received 3 October 2001; revised 8 November 2001; accepted 5 Decem-
ber 2001
*Correspondence: Professor Mark McGurk, Floor 23, Guys Hospital,
London SE1 9RT, UK
British Journal of Cancer (2002) 86, 512–516
ª 2002 Cancer Research UKAll rights reserved 0007–0920/02$25.00
www.bjcancer.com

Page 2

radiochemical purity 499% and 350 MbQ were injected intra-
venously.Aftera30–60 min
acquired in 2D mode on an ECAT 951R scanner (Siemens/
CTI, Knoxville, TN, USA). Emission data were acquired at
15 min per bed position (two positions) each of 31 slices over
a 10.6 cm axial field of view. A transmission scan using Germa-
nium 68 source was obtained for attenuation correction for
10 min following the emission. Images were reconstructed using
filtered back projection (Hann filter, cut-off frequency 0.4) and
viewed as coronal, transaxial and sagittal slices with a spatial
resolution of 12 mm.
Standard uptake values (SUV) of the main lesion were calculated
using the attenuation corrected image. One observer positioned a
4.5 mm region of interest over the maximum pixel in the lesion
and the average of two perpendicular diameters of the lesion was
noted. The SUV was calculated using the formula: tissue concentra-
tion of FDG measured by PET divided by the injected dose divided
by body weight. No correction was made for glucose or partial
volume.
uptake period imageswere
Statistical analysis
Statistical analysis was performed using a standard software pack-
age (Stata: Release 7.0, Stata Corp., College Station, TX, USA).
Standard diagnostics showed that age and tumour diameter had
a normal distribution and SUV a log-normal. For age and
diameter, mean and s.d. are given. For SUV, geometric mean,
median and 90% range (5th to 95th centile). For categories,
percentages are used.
The effect of tumour diameter and stage on SUV was examined
by linear regression on the log of SUV. Analysis of SUV initially
examined survival within SUV subgroups that were defined by
the quartiles of SUV distribution. A log rank test was used to
determine a statistically significant cut-off value of 10, which was
used for survival analysis.
Survival times were calculated from the time of the PET
study. The eight intercurrent deaths were all regarded as clearly
unrelated to head and neck cancer and were treated as censored
at time of death. A stepwise Cox regression analysis was
performed for recognized risk factors: first unadjusted (‘univari-
ate’) and then adjusting (‘multivariate’) for the main predictors
(stage tumour diameter and SUV). Due to the small number
of patients with stage 1 disease they were combined with the
stage 2 patients for analysis. For similar reasons those patients
classified as performance status 2 were combined with the
performance status 1 group. Survival curves were obtained by
the Kaplan–Meier method. Robust standard errors were used
to correct for possible unequal variances and any residual diver-
gence from normality. A P value of less than 0.05 was
considered significant.
RESULTS
All primary tumours were correctly identified on the PET scans.
The mean tumour diameter was 25.60 mm (s.d. 8.08). Outcome
evaluation showed 28 out of 58 patients (48%) were dead by the
end of the study period: 20 (72%) from disease and eight (28%)
from other causes (myocardial infarction (four), stroke (three),
prostate cancer (one)). The stage of disease at initial presentation
of those who died of head and neck cancer was: stage 2 (n=1),
stage 3 (n=3) and stage 4 (n=16).
The median SUV for all primary tumours was 7.16 (90% range
2.30 to 18.60). Stage was a more important predictor of SUV than
tumour diameter (Figures 1 and 2). An increase in tumour
diameter of 1 cm resulted in only a 1.06 times rise in SUV (Ratio
1.06, 95% CI 0.8–1.3 P=0.55), compared to a four-fold rise in
moving from stage 1 to stage 4 disease (Ratio 4.17, 95% CI
2.4–7.4 P=0.004).
In univariate survival analysis high stage (P=0.015), increasing
tumour diameter (P=0.041) and SUV 410 (P=0.003) were the
main negative predictors of survival (Table 2). The Kaplan–Meier
survival curves are shown in Figures 3–5.
In multivariate analysis the joint effect of stage, tumour
diameter and SUV was examined to determine whether SUV had
prognostic significance beyond that provided by stage and tumour
diameter (Table 3). Patients with an SUV 410 had a significantly
Clinical
Table 1
Patient and tumour characteristics
No. of patients(%)
Gender
Male
Female
42
16
72.0
28.0
Age (years)
460
460
25
33
Mean 64.0 s.d. 12.7
43.0
57.0
WHO performance status
0
1
2
42
12
72.0
21.0
7.04
Tumour site
Lip
Oral cavity
Oropharynx
Nasopharynx
Hypopharynx
Maxillary sinus
1 1.5
83.0
11.0
1.5
1.5
1.5
48
6
1
1
1
Tumour grade
Grade 1
Grade 2
Grade 3
8 15.0
50.0
35.0
26
18
Tumour stage
1
2
3
4
2 3.0
20.0
21.0
56.0
11
12
33
Tumour diameter (cm)Mean 25.6 s.d. 8.08
Standardized uptake valueGeometric mean 6.54
Median 7.16
90% range (2.3—18.6)
30
20
10
0
SUV
10 20 30 40 50
Diameter (mm)
Figure 1
Scattergram of SUV with tumour diameter.
FDG–PET and prognosis in head and neck cancer
W Halfpenny et al
513
ª 2002 Cancer Research UK British Journal of Cancer (2002) 86(4), 512–516

Page 3

poorer prognosis (P=0.002). The relative risk was increased for
patients with high stage disease (Figure 6).
DISCUSSION
The role of PET–FDG in the management of squamous cell carci-
noma (SCC) of the head and neck continues to evolve and its full
potential has yet to be established. It has been applied to staging
and surveillance of patients with head and neck SCC (Rege et al,
1994; Braams et al, 1995), but a more promising observation is that
high FDG uptake in head and neck SCC is associated with poor
survival (Minn et al, 1997). A similar association has also been
demonstrated for other malignancies such as brain, lung and
malignant lymphomas (Patronas et al, 1985; Okada et al, 1992;
Ahuja et al, 1998), where it appears that the prognostic value of
a high SUV may act independently of tumour stage.
Clinical
30
20
10
0
SUV
1 2 3 4
Stage
Figure 2
Scattergram of SUV with stage.
Table 2
Univariate analysis
eb
95% confidence
interval
b
(relative risk)P
Age (decades)
70.0530.0948 0.871—1.035 0.240
Female
Male0.757 2.131 0.620—7.3210.229
Grade 1
Grade 2
Grade 3
70.0107
70.273
0.899
0.761
0.270—2.988
0.285—2.0370.587
Perf status 0
Perf status 1+20.018 1.018 0.368—2.816 0.972
Stage 1+2
Stage 3
Stage 4
1.199
1.339
3.315
3.815
1.239—8.870
1.518—9.584 0.015
Diameter (cm) 0.0641.720 1.022—2.8950.041
SUV410
SUV4101.475 4.3691.652—11.558 0.003
1.00
0.75
0.50
0.25
0.00
Cumulative survival
0 10 20 30
Analysis time (months)
Stage 1+2
Stage 3
Stage 4
Figure 3
Kaplan–Meier survival estimates, by stage.
1.00
0.75
0.50
0.25
0.00
Cumulative survival
0 10 20 30
Analysis time (months)
Tumour diameter
Figure 4
Kaplan–Meier survival estimate, by tumour diameter.
1.00
0.75
0.50
0.25
0.00
Cumulative survival
0 10 20 30
Analysis time (months)
SUV ≤10
SUV >10
Figure 5
Kaplan–Meier survival estimates, by SUV10.
Table 3
Multivariate analysis
eb
95% confidence
interval
b
(relative risk)P
Diameter (cm)0.0291.0290.771—2.282 0.410
Stage 1+2
Stage 3
Stage 4
———
1.071
1.277
2.919
3.587
1.130—7.535
1.211—10.6290.021
SUV410
SUV410
———
1.1734.170 1.146—8.4960.002
FDG–PET and prognosis in head and neck cancer
W Halfpenny et al
514
British Journal of Cancer (2002) 86(4), 512–516
ª 2002 Cancer Research UK

Page 4

The role of SUV as a prognostic factor has been extensively
studied in non-small cell lung cancer (NSCLC) which consists
largely of patients with SCC. This tumour type has a very similar
biology to SCC of the head and neck, and smoking is a common
aetiological agent. It is not surprising therefore that the present
results mirror those for lung cancer, where SUV has proved to
be a significant prognostic indicator (Ahuja et al, 1998; Saunders
et al, 1999; Vansteenkiste et al, 1999).
Vansteenkiste et al (1999) suggested a SUV of seven, and Ahuja
et al (1998) one of 10 as a cut-off value that distinguishes good and
bad prognostic groups, although there is probably a range of 6–11
where prognosis deteriorates. Previous studies in head and neck
cancer have found SUV important in differentiating benign and
malignant disease but were unable to define a cut-off value (Nowak
et al, 1999). In the present study a SUV 410 proved to be an inde-
pendent marker of poor prognosis and this is consistent with
results obtained for lung cancer.
Uptake time is important in sarcomas (Lodge et al, 1999) and
possibly testicular cancer (Hain et al, 2000). The optimal time
allowed for FDG uptake in squamous cell carcinoma of the head
and neck is unknown. The uptake period was consistent through-
out this study (30–60 min). Where differences have been seen the
uptake has varied over a period of hours rather than 30–60 min.
The association of high FDG uptake with poor survival may
be relatedto several factors.
advanced disease present with large tumours, and FDG uptake
may simply reflect the larger tumour burden in these patients.
Indeed some of the lung studies have shown that size as well
as SUV has prognostic value (Saunders et al, 1999; Vansteen-
kiste et al, 1999). However in the current study increasing
tumour diameter had little effect on FDG uptake (SUV),
suggesting that tumour burden alone cannot account for these
findings.
Secondly, it has been proposed that the abnormal carbohydrate
metabolism seen in the cachectic state is responsible for the dispro-
portionate glucose uptake, in tumours of patients with advanced
malignancies (Tayek, 1992). An association has been demonstrated
between FDG uptake and cellular proliferation and/or tumour
growth rate in a number of cancers including head and neck, lung
tumours and non-Hodgkin’s lymphoma (Minn et al, 1988, 1997;
Haberkorn et al, 1991).
Thirdly, Kitagawa et al (1999) found that SUV value in SCC of
the head and neck pre-treatment predicted response to chemo/
radiotherapy, with tumours demonstrating higher SUV having
greater treatment resistance. Together these observations provide
tentative evidence that the rate of glucose metabolism in SCC
may reflect adverse tumour biology and survival. If this proves
correct, FDG uptake represents one of the few objective measures
of tumour malignancy.
In this study high FDG uptake (SUV 410) proved to be a
marker for poor outcome in SCC of the head and neck, particu-
larly in those patients with stages 3 and 4 disease. SUV increased
with stage, but multivariate analysis suggests SUV was an inde-
pendent predictor of survival, and not dependent on stage.
Therefore patients with tumours that are more active metaboli-
cally,as demonstratedbyFDG–PET
considered at high risk for relapse, regardless of clinical stage at
presentation. This is similar to results obtained in lung cancer
where both PET staging and SUV are currently the best predictors
of survival, apart from operative stage (Saunders et al, 1999). If
these findings are substantiated it will allow the clinician to distin-
guish those tumours with a more aggressive biological nature, and
hence identify those patients that require intensive treatment
protocols, including hyperfractionated
chemotherapy.
Firstly, mostpatientswith
imaging, shouldbe
radiotherapyand/or
REFERENCES
Ahuja V, Coleman RE, Herndon J, Patz Jr EF (1998) The prognostic signifi-
cance of fluorodeoxyglucose positron emission tomography imaging for
patients with nonsmall cell lung carcinoma. Cancer 83: 918–924
Braams JW, Pruim J, Freling NJM, Nikkels PG, Roodenburg JL, Boering G,
Vaalburg W, Vermey A (1995) Detection of lymph node metastases of
squamous cell cancer of the head and neck with FDG-PET and MRI. J Nucl
Med 36: 211–216
Conti PS, Strauss LG (1991) The applications of PET in clinical oncology. J
Nucl Med 32: 623–648
Haberkorn U, Strauss L, Reisser CH, Haag D, Dimitrikopoulou A, Ziegler S,
Oberdorfer F, Rudat V, van Kaick G (1991) Glucose uptake, perfusion and
cell proliferation in head and neck tumours: relation of positron emission
tomography to flow cytometry. J Nucl Med 32: 1548–1555
Hain SF, O’Doherty MJ, Timothy AR, Leslie MD, Harper PG, Huddart RA
(2000) Fluorodeoxyglucose positron emission tomography in the evalua-
tion of germ cell tumours at relapse. Br J Cancer 83: 863–869
Hallett WA, Marsden PK, Cronin BF, O’Doherty MJ (2001) Effect of correc-
tions for blood glucose and body size on [18F] FDG PET standardised
uptake values in lung cancer. Eur J Nucl Med 28: 919–922
Hermanek P, Sobin LH (1992) UICC TNM classification of malignant tumours
4th edn Berlin: Springer-Verlag
Kitagawa Y, Sadato N, Azuma H, Ogasawara T, Yoshida M, Ishii Y, Yonekura
Y (1999) FDG PET to evaluate combined intra-arterial chemotherapy and
radiotherapy of head and neck neoplasms. J Nucl Med 40: 1132–1137
Lodge MA, Lucas JD, Marsden PK, Cronin BF, O’Doherty MJ, Smith MA
(1999) A PET study of 18FDG uptake in soft tissue masses. Euro J Nuc
Med 26: 22–30
Minn H, Joensuu H, Ahonen A, Klemi P (1988) Fluorodeoxyglucose imaging:
a method to assess the proliferative activity of human cancer in vivo.
Cancer 61: 1776–1781
Minn H, Zasadny KR, Quint LE (1995) Lung cancer: reproducibility of quan-
titative measurements for evaluating 2-[F-18]-fluoro-2-deoxy-D-glucose
uptake at PET. Radiology 196: 167–173
Minn H, Lapela M, Klemi PJ, Grenman R, Leskinen S, Liindholm P, Bergman
J, Eronen E, Haaparanta M, Joensuu H (1997) Prediction of survival with
fluorine-18-fluorodeoxyglucose and PET in head and neck cancer. J Nucl
Med 38: 1907–1911
Nowak B, Di Martino E, Janicke S, Cremerius U, Adam G, Zimny M, Rein-
artz P, Bull U (1999) Diagnostic evaluation of malignant head and neck
cancer by F-18-FDG PET compared to CT/MRI. Nuklearmedizin 38:
312–318
Clinical
4
3
2
1
0
Relative risk, by stage and SUV level
SUV >10
SUV ≤10
4 3 1+2
Tumour stage
Figure 6
Relative risk of death by SUV and stage.
FDG–PET and prognosis in head and neck cancer
W Halfpenny et al
515
ª 2002 Cancer Research UK British Journal of Cancer (2002) 86(4), 512–516

Page 5

Okada J, Yoshikawa K, Itami M, Imaseki K, Uho K, Itami J, Kuyoma J, Mika-
ta A, Arimozu N (1992) Positron emission tomography using fluorine-18-
fluorodeoxyglucose in malignant lymphoma: a comparison with prolifera-
tive activity. J Nucl Med 33: 325–329
Patronas NJ, Di Chiro G, Kufta C, Bairamian D, Kornblith PL, Simon R,
Larson SM (1985) Prediction of survival in glioma patients by means of
positron emission tomography. J Neurosurg 62: 816–822
Rege S, Maas A, Chaiken L, Hoh CK, Choi Y, Lufkin R, Anzai Y, Julliard G,
Maddahi J, Phelps ME (1994) Use of positron emission tomography with
flurodeoxyglucose in patients with extracranial head and neck cancers.
Cancer 73: 3047–3058
Saunders CA, Dussek JE, O’Doherty MJ, Maisey MN (1999) Evaluation of
fluorine-18-Fluorodeoxyglucose whole Body Positron emission Tomogra-
phy Imaging in the Staging of Lung cancer. Ann Thorac Surg 67: 7790–
7797
Tayek JA (1992) A review of cancer cachexia and abnormal glucose metabo-
lism in humans with cancer. J Am Coll Nutr 11: 445–456
Vansteenkiste JF, Stroobants SG, Dupont PJ, De Leyn PR, Verbeken EK,
Deneffer GJ, Mortlemans LA, Demedts MG (1999) Prognostic importance
of the standardized uptake value on 18F-Fluoro-2-deoxy-glucose positron
emission tomography scan in non-small-cell lung cancer: an analysis of 125
cases. J Clin Oncol 17: 3201–3206
Warburg O (1956) On the origin of cancer cells. Science 123: 309–314
Clinical
FDG–PET and prognosis in head and neck cancer
W Halfpenny et al
516
British Journal of Cancer (2002) 86(4), 512–516
ª 2002 Cancer Research UK