Radiotherapy and Oncology 55 (2000) 121±129
www.elsevier.com/locate/radonline
Cervical lymph node metastases from unknown primary tumours
Results from a national survey by the Danish Society for Head and Neck
Oncology
Cai Grau a,b,*, Lars Vendelbo Johansen a, John Jakobsen c, Poul Geertsen d,
Elo Andersen e, Brita Bjerregaard Jensen f
a
Department of Oncology and Department of Head and Neck Surgery, Aarhus University Hospital, 8000 Aarhus C, Denmark
b
Department of Experimental Clinical Oncology, Aarhus University Hospital, 8000 Aarhus C, Denmark
c
Department of Oncology and Department of Head and Neck Surgery, Odense University Hospital, Odense, Denmark
d
Department of Oncology and Department of Head and Neck Surgery, National University Hospital Copenhagen, Denmark
e
Department of Oncology and Department of Head and Neck Surgery, Copenhagen County Hospital, Herlev, Denmark
f
Department of Oncology and Department of Head and Neck Surgery, Aalborg Sygehus, Aalborg, Denmark
Received 17 May 1999; received in revised form 21 January 2000; accepted 16 February 2000
Abstract
Background and purpose: The management of patients with cervical lymph node metastases from unknown primary tumours is a major
challenge in oncology. This study presents data collected from all ®ve oncology centres in Denmark.
Material and methods: Of the 352 consecutive patients with squamous cell or undifferentiated tumours seen from 1975 to 1995, a total of
277 (79%) were treated with radical intent. The general treatment policy at all centres during the entire study period has been to treat all
suitable candidates with radiotherapy to both sides of the neck and include elective irradiation of the mucosal sites in nasopharynx, and
larynx, hypopharynx and larynx (81%). Irradiation of the ipsilateral neck only was done in 26 patients (10%). Radical surgery was the only
treatment in 23 N1±N2 patients (9%).
Results: The 5-year estimates of neck control, disease-speci®c survival and overall survival for radically treated patients were 51, 48 and
36%, respectively. The emergence of the occult primary was observed in 66 patients (19%). About half of the emerging primaries were
within the head and neck region with oropharynx, hypopharynx and oral cavity being the most common sites. Emerging primaries outside the
head and neck region were primarily located in the lung (19 patients) and oesophagus (®ve patients). The frequency of emerging primary in
the head and neck was signi®cantly higher in patients treated with surgery alone, the actuarial risks at 5-year being 54 ^ 1% (no RT) vs. 15 ^
3% (with RT), P , 0:0001. The most important factor for neck control was nodal stage (5-year estimates 69% (N1), 58% (N2) and 30%
(N3)). Other important parameters for neck control and disease-speci®c survival included haemoglobin, gender and overall treatment time.
Patients treated with ipsilateral radiotherapy had a relative risk of recurrence in the head and neck region of 1.9 compared with patients
treated to both neck and mucosa. At 5 years, the estimated control rates were 27% (ipsilateral) and 51% (bilateral; P ˆ 0:05). The 5-year
disease-speci®c survival estimates were 28 and 45%, respectively (P ˆ 0:10).
Conclusions: This study has con®rmed that patients with neck node metastases from occult head and neck cancer have clinical features and
prognosis similar to other head and neck malignancies. Extensive irradiation to both sides of the neck and the mucosa in the entire pharyngeal
axis and larynx resulted in signi®cantly less loco-regional failures compared with patients treated with ipsilateral techniques, but only a trend
towards better survival. A prospective randomized trial is required to determine the optimal strategy in terms of locoregional control, survival
and morbidity. q 2000 Elsevier Science Ireland Ltd. All rights reserved.
Keywords: Unknown primary tumor; Head and neck neoplasm; Radiotherapy; Head and neck surgery; Squamous cell carcinoma
1. Introduction
Cervical lymph node metastases from unknown primary
tumours are rare, constituting only about 2% of all new head
and neck cancers. However, the management of these
patients remains a major challenge in oncology. Recent
developments in imaging [2,13,16] and pathology [5,6,19]
have increased our diagnostic spectrum considerably, but
the impact of these techniques on decision-making has not
been well documented. The selection and timing of the
diagnostic measures in the work-up process is still under
* Corresponding author.
0167-8140/99/$ - see front matter q 2000 Elsevier Science Ireland Ltd. All rights reserved.
PII: S 0167-814 0(00)00172-9
122
C. Grau et al. / Radiotherapy and Oncology 55 (2000) 121±129
debate, and varies from centre to centre even within a small
country like Denmark. The choice of treatment is controversial. Recommendations vary from surgery alone in
selected cases [4], limited ®eld radiotherapy, where only
the ipsilateral neck is treated [8,18,21], or extensive prophylactic irradiation of all potential mucosal sites as well as
both sides of the neck [3,7,9,14,17,20]. No randomized or
prospective studies are available to support either of these
approaches in particular, and such a study would also be
dif®cult to undertake since the disease is so rare. The Danish
Society for Head and Neck Oncology therefore decided to
collect retrospective data from all ®ve oncology centres in
Denmark on the management and treatment outcome for
patients seen in the period 1975±1995. These data will
form the basis for a set of national management guidelines
for this disease. The present study summarizes the main
®ndings of the national survey.
gen (131 patients), Aarhus (92 patients), Odense (65
patients), Herlev (41 patients) and Aalborg (23 patients).
This distribution is consistent with the relative size of the
catchment areas for these centres. There were 248 males
(71%) and 104 females (29%). The primary symptoms
were enlarged lymph node (94%), pain (9%), and weight
loss (7%). The distribution of involved lymph nodes is
shown in Fig. 1. The sub-digastric nodes were by far the
most commonly involved, and with disease in this region
most patients were offered radical treatment. In contrast,
less than half of the patients with supraclavicular involvement, either isolated or in combination with more cranial
involvement were offered curative treatment. Fifteen
patients with isolated supraclavicular metastases were
offered curative treatment, and to make the series complete
these patients have been included in the analysis, although
their high risk of incurable disease below the clavicles is
acknowledged.
2. Patients and methods
The Danish Society for Head and Neck Oncology
initiated the study in July 1996. A common agreement
was made between all ®ve head and neck oncology centres
in Denmark to record all consecutive patients with metastatic cervical lymph nodes from occult primary tumour in a
common database. The ®ve co-authors of the present paper
were responsible for data collection and recording at each of
the ®ve institutions, respectively. The data recording was
®nished in 1997. A total of 491 patients with tumours of
various histological types were recorded in the database.
The present report included only patients with squamous
cell or undifferentiated tumours. No histology review was
performed for this report. The ®nal study material included
352 patients seen in the period January 1, 1975 to May 30,
1995. The recruitment from the ®ve centres was Copenha-
2.1. Diagnostic work-up
Classi®cation of a patient as having an unknown primary
tumour was done if adequate investigations failed to detect a
possible primary tumour site at the time when a ®nal treatment decision was made. The extent of investigations is
listed in Table 1. All patients had a biopsy of the enlarged
lymph node: 45% excision biopsy, 40% incision biopsy, 2%
neck dissection, 1% core biopsy, 12% ®ne needle aspiration.
Evaluation under anaesthesia (EUA) with laryngoscopy,
bronchoscopy and oesophagoscopy was part of the general
work-up, and 94% of all patients actually had this done.
Biopsies of potential sub clinical tumour sites in pharynx
(nasopharynx, tonsil, base of tongue) and larynx were done
in 55% of cases. All patients had a chest X-ray, and 14% had
chest CT scan. Imaging of the neck region was done by one
Fig. 1. Involved lymph nodes in all patients (left) and patients treated with radical intent (right). Up to three locations were recorded for each patient.
C. Grau et al. / Radiotherapy and Oncology 55 (2000) 121±129
Table 1
Applied work-up procedures in the 352 patients presenting with neck node
metastases from an unknown primary tumour
Investigations
Percentage
Node biopsy
(FNA only)
Evaluation under anaesthesia
Biopsy nasopharynx
Biopsy tonsil
Other biopsies
Chest X-ray
CT head and neck
CT chest
Ultrasound
MRI head and neck
PET
100
12
94
40
22
55
99
30
14
16
7
1
ore more of the following diagnostic approaches: CT scan
(30%), thyroid scintigraphy (26%), ultrasound (16%), MRI
(7%) and/or PET (1%). With ®ve different centres and a 20year period the work-up procedures were obviously heterogeneous. CT/MRI have not been applied systematically as it
has not been available throughout most of the period. The
work-up procedures were not signi®cantly different between
different treatment groups with the exception that relatively
more patients treated with extensive radiotherapy had CT of
the head and neck. All available clinical and diagnostic
information was used for clinical staging. For this report
all patients were N-staged based on chart information
according to the UICC TNM 1987 classi®cation.
2.2. Radical treatment
Decision about treatment intent and technique was made
at one of the ®ve oncology centres. The treatment intent
(curative or palliative) was declared in the patient chart
before the start of the treatment. Of the 352 patients in the
study, 277 (79%) were treated with radical intent. The distribution of nodal stages in the various treatment groups is
123
shown in Table 2. The general treatment policy at all centres
during the entire study period was to treat all suitable candidates with radiotherapy. This re¯ects the general policy in
Denmark to treat head and neck squamous cell carcinomas
with radiotherapy and reserve surgery for salvage. Most
patients had elective irradiation of both sides of the neck
and the mucosal sites in nasopharynx, oropharynx, hypopharynx and larynx. As a consequence, 224 patients (81%
of radically treated patients) received this treatment; 21 of
these also had some surgical procedure. Radiotherapy to the
ipsilateral neck only was used in 26 cases. The target
volume included the involved lymph node region and a
variable margin, most often encompassed in a direct electron ®eld or a wedge pair of photon ®elds. Four of the
patients receiving ipsilateral treatment also had surgery.
Radiation was delivered by linear accelerator (4±6 MV) or
cobalt in 2 Gy per fraction and 5 fractions per week. A few
patients received non-standard fractionation. For the current
study, the total radiation dose was calculated as the biological equivalent given in 2-Gy fractions using an a /b ratio
of 10 Gy for tumour response. Overall, 98% of the patients
treated to the neck and mucosa received a biologically effective dose of 661 Gy (median 66 Gy, range 20±79 Gy).
Patients treated to the neck only received a median radiation
dose of 59 Gy (range 28±93 Gy). Radical surgery alone,
either lymph node excision or modi®ed radical neck dissection, was done in 23 patients.
2.3. Follow-up
All patients were followed regularly at the oncology
centre for a period of 5 years after treatment. The absence
or presence of any recurrent disease in the neck, head and
neck primary sites or systemic was recorded in the patient
chart. For the present analysis, all patient charts were
reviewed and the disease and survival status, including
cross-check with the National Health Registry, was updated
per June 1, 1997.
Table 2
Distribution of nodal stage between treatment groups a
Nodal stage (UICC87)
Treatment
RT neck
RT neck 1 mucosa
Surgery only
Total
No.
%
No.
%
No.
%
No.
%
N1
5
19
32
14
9
39
46
17
N2
N2a
N2b
N2c
2
8
3
50
6
66
27
7
47
2
8
2
1
56
10
82
32
8
48
N3
Nx
Total
7
1
26
27
86
39
0
93
2
273
34
a
224
1
23
Only patients treated with curative intent are included (incomplete treatment information about four patients).
124
C. Grau et al. / Radiotherapy and Oncology 55 (2000) 121±129
2.4. Endpoints and statistics
Endpoints for the present study were neck control, mucosal control (de®ned as absence. of emerging primary tumour
in the head and neck region), loco-regional control
(combined neck and mucosal control), disease-speci®c
survival (death from or with the actual cancer) and overall
survival (death from any cause). All time estimates began at
the ®rst date of the initial treatment or, if no treatment was
given, the date of the initial consults at the oncology centre.
The data was analyzed using the SPSS for Windows version
8.0 statistical software. The tumour response and survival
data was analyzed actuarially using Kaplan±Meier method.
Patients were censored at time of their ®rst loco-regional
relapse in the evaluation of local and regional tumour
control. The statistical difference between various prognostic parameters was tested using log-rank test, and a test for
trend was used when applicable. A signi®cance level of 5%
(two-sided) was used for all tests. Data are represented as 5year actuarial value ^ standard error of the mean unless
otherwise mentioned. The independent signi®cance of parameters was tested in a Cox proportional hazard model. Parameters were included in the model using the enter method
and statistical analysis was performed using likelihood ratio.
The time for evaluation of loco-regional control and
disease-speci®c survival was 5 years after initial treatment,
since patients were only followed routinely at the oncology
centre for this time period. However, the date for evaluation
of overall survival was June 1, 1997, which gave a median
potential follow-up time of 137 months (range 24±226
months) for that endpoint.
3. Results
3.1. Clinical course for all patients (n ˆ 352)
At the time of analysis, 65 of the 352 patients were alive.
A total of 215 patients (61%) had died from their primary or
emerging head and neck cancer, 61 patients (17%) from
other diseases or unknown causes, and six patients (2%)
from other cancers. Five patients (1%) died from treatment
complications, including one patient with myelopathy and
paraplegia 11 years after treatment and one patient with
sepsis after successful salvage laryngectomy. The clinical
course in terms of loco-regional treatment outcome for the
352 patients is outlined in Fig. 2.
Of the 277 patients treated with radical intent, 111
patients were free of recurrence or emerging primary at
the time of analysis. The remaining 166 radically treated
patients experienced recurrent nodal disease and/or emerging primary with the following distribution: 20 primary
tumour recurrence only, 92 neck recurrence only, 19 distant
metastases, 16 combined T 1 N-position, 16 N 1 M-position, 2 T 1 M-position, and 1 T 1 N 1 M-position. Salvage
treatment (most often neck dissection) was attempted in 64
Fig. 2. Clinical course (in terms of treatment outcome) for 352 patients
with cervical lymph node metastases from unknown primary tumour.
of these patients, and was successful in 20 patients. The ®nal
result was that 131 patients obtained tumour control and 221
patients failed. The actuarial 5-year disease-speci®c and
crude survival for all patients was 38 and 29%, respectively.
3.2. Prognostic factors for patients treated with curative
intent (n ˆ 277)
The 5-year disease-speci®c and crude survival for patients
treated with curative intent was 48 and 36%, respectively.
The Kaplan±Meier estimates of nodal control, mucosal
control, combined loco-regional control, cause-speci®c and
overall survival is summarized in Table 3. The most important factor for treatment outcome and survival was the nodal
stage. Fig. 3 shows the loco-regional tumour control in the
three N-stages. Patients with N1 and N2 disease had a significantly better prognosis compared with N3 patients. Other
important factors for treatment outcome were gender
(females did better), haemoglobin (high haemoglobin was
better) and differentiation. Only 14 patients had undifferentiated carcinoma, so this group could not be analyzed separately, but was grouped with poorly differentiated tumours.
Differentiation was especially important for the risk of emerging primary, as signi®cantly fewer patients with undifferentiated/poorly differentiated lymph node metastases
experienced an emerging primary in the head and neck
mucosa. This group also had a marginally better 5-year
disease-speci®c survival (51 vs. 43%; P ˆ 0:05) and overall
survival (40 vs. 30%; not signi®cant (NS)). Age, when
divided at the median of 62 years, did not in¯uence the prognosis. For patients receiving radiotherapy, there was a signif-
C. Grau et al. / Radiotherapy and Oncology 55 (2000) 121±129
125
Table 3
Five-year actuarial probability for neck control, absence of emerging occult primary site within head and neck region, loco-regional control (neck and/or
primary site above clavicles), cause-speci®c survival (death from any cancer related to neck node in question) and overall survival for the 277 patients treated
with radical intent a
Parameter
No. of patients
Neck control
Mucosal control (H&N region) Loco-regional tumour control Cause-speci®c survival
Overall survival
%
%
%
P-value
P-value
Overall
277
51
81
Age
18±62 years
63±92 years
144
132
54
47 NS
77
87
90
17
14
59
63 NS
24
80
63
NS
Performance status
0
1
2±4
%
P-value
44
0.02
42
42
52
38
25
%
P-value
48
NS
NS
53
42
58
62
36
36
NS
0.03
43
28
50
41
21
Gender
Male
Female
201
76
48
61
0.007
80
85
NS
40
55
0.002
42
61
0.009
33
42
Haemoglobin
Low b
High b
128
118
60
41
0.004
85
83
NS
53
36
0.004
59
35
0.003
45
25
152
51
86
48 NS
62
71
86
87
Differentiation
Undifferentiated or
poorly diff.
Moderately or well diff.
60
48
0.0009
43
0.05
30
NS
58
50
30
0.0001
58
55
32
0.00001
48
38
25
NS
45
44
39
NS
52
40
52
29
44
59
NS
76
45
49
69
58
30
N-level c
Level 1
Level 2
Level 3
158
94
25
54
46 NS
48
78
87
86
Treatment modality
Surgery alone
Radiotherapy alone
Combined RT 1 S
23
213
26
58
50 NS
62
46
84
95
RT dose
1±65 Gy
66 1 Gy
127
124
52
48 NS
85
86
NS
45
45
RT treatment time
1±50 days
51 1 days
139
112
55
45
88
83
NS
51
38
RT technique
Neck only
Neck 1 mucosa
26
224
43
52 NS
77
87
NS
31
48
0.03
0.00001
NS
0.02
NS
0.002
0.009
NS
0.01
40
0.02
46
136
93
0.00001
51
33
N-stage (UICC)
NI
N2
N3
P-value
NS
0.025
39
29
38
65
37
28
NS
0.0008
NS
0.04
43
47
NS
30
35
NS
50
39
NS
36
29
NS
32
47
NS
22
34
NS
a
The parameters were tested using Kaplan±Meier analysis and differences between groups were tested for signi®cance using two-sided log-rank test (5%
signi®cance level). NS, not signi®cant.
b
Low haemoglobin: , 8 mmol/l (12.8 g/dl) in females, ,9 mmol/l (14.4 g/dl) in males.
c
Most distally involved node. Level 1: sub-digastric and above; level 2: midjugular and mid deep cervical; level 3: supraclavicular, low posterior or
pretracheal.
icant negative in¯uence of prolonged overall treatment …501
days) compared with shorter schedules. This effect was
evident for neck control and loco-regional tumour control,
but it was not signi®cant for survival. Radiation dose (when
divided at the median of 65 Gy) was not found to be signi®-
cant for tumour control nor survival. Node level, de®ned as
the most distal localization of the involved nodes, was not
important for treatment outcome or survival. Not shown in
Table 3 is extracapsular extension, which was not reported
suf®ciently to be analyzable.
126
C. Grau et al. / Radiotherapy and Oncology 55 (2000) 121±129
ynx and larynx (n ˆ 224). The actuarial values at 5/10 years
were 46/46% (surgery), 77/77% (RT neck), and 87/75%
(RT neck1mucosa), respectively. Both radiotherapy groups
were signi®cantly better than surgery to control occult
mucosal primaries (P , 0:05), and there was no difference
between the two radiotherapy groups.
3.4. Comparison of patients treated with different
radiotherapy techniques (n ˆ 250)
The potential difference in treatment outcome between the
two radiotherapy groups was further tested in Cox multivariate analysis to allow for variations in known prognostic
factors between the two groups (Table 5). Differentiation
was not included in the model, as it was too infrequently
recorded. However, the percentage of poorly differentiated
or undifferentiated tumours was the same (70%) in the two
groups. The relative risk of neck failure in N3 patients was
signi®cantly increased (P ˆ 0:03). Male patients had a
signi®cantly poorer neck control. There was no in¯uence
of radiotherapy technique, total dose or overall treatment
time. Ipsilateral treatment resulted in only one recurrence
in the untreated contralateral neck (4%). Bilateral treatment
resulted in ®ve contralateral recurrences (2%). These
numbers are too small to give meaningful statistical information. None of the parameters tested had any signi®cant independent information about the risk of mucosal relapse.
Again, the numbers of events were low (17 emerging
primaries). When the two endpoints were combined in locoregional control, the factors important included N-stage,
gender and radiotherapy technique. The latter meant that
patients treated with ipsilateral technique had a relative risk
of recurrence in the head and neck region of 1.9 compared
with patients treated to both neck and mucosa (P ˆ 0:05).
This difference is illustrated in Fig. 5 (left), where the
adjusted locoregional control curves for the two groups are
plotted at means of covariates. At 5 years, the estimated
control rates were 27 and 51%, respectively. For the endpoint
Fig. 3. Loco-regional tumour control as a function of nodal stage for
patients treated with radical intent (n ˆ 277). The 5-year actuarial values
were 58% (N1), 50% (N2) and 30% (N3), respectively.
3.3. Emerging primary tumours
The emergence of the occult primary, de®ned as the
appearance of a primary tumour with relevant clinical
features and same histology, was observed in 20% of all
patients. About half of these emerging primaries occurred
within the head and neck region (Table 4). Oropharynx
(base of tongue) was by far the most common site of emerging primary (n ˆ 14), whereas other occult sites in the
head and neck region were extremely rare. Emerging
primaries outside the head and neck region were primarily
located in the lung (14 patients) and oesophagus (®ve
patients). Fig. 4 shows the actuarial probability of being
free of emerging primary in the head and neck region for
radically treated patients assigned to one of three groups:
The ®rst group had neck dissection only (n ˆ 23). The
second received unilateral radiotherapy ^ neck surgery
(n ˆ 26). The remaining group received radiation treatment
to both sides of the neck and the mucosal sites in the phar-
Table 4
Observed incidence of emerging primary tumours according to treatment groups; only patients treated with curative intent included (n ˆ 277)
Site
Treatment
RT neck (n ˆ 26)
No.
Oropharynx
Hypopbarynx
Nasopharynx
Oral cavity
Larynx
Other H&N
Lung
Esophagus
GI
Other
Total
%
1
4
2
8
3
12
1
7
RT neck 1 mucosa
(n ˆ 224)
Surgery only (n ˆ 23)
Total (n ˆ 277)
No.
No.
%
No.
%
7
1
30
4
1
4
2
1
9
4
1
13
4
57
14
3
2
4
4
4
14
5
4
1
55
5
1
1
1
1
1
5
2
1
1
20
%
4
6
2
2
3
2
2
10
5
3
3
1
1
1
1
1
4
2
1
27
35
16
C. Grau et al. / Radiotherapy and Oncology 55 (2000) 121±129
Fig. 4. Actuarial estimate of being free of emerging primary in the head and
neck region (all mucosal sites above clavicles) as a function of time after
either radiotherapy to the neck and mucosa (n ˆ 224), neck irradiation
(n ˆ 26) or surgery alone (n ˆ 23).
of disease-speci®c survival, N-stage, gender and overall
treatment time were independent factors. The irradiated
volume did not in¯uence survival: Fig. 5 (right) shows the
corresponding survival plots from the Cox analysis. When
adjusted for cofactors the 5-year disease-speci®c survival
estimates were 28 and 45%, respectively (P ˆ 0:10).
4. Discussion
The present report summarizes the Danish experience
with occult head and neck cancer. With all eligible patients
from a 20-year period registered and reported, the series
should allow for some conclusions to be drawn on potential
prognostic factors of importance for choice of treatment.
127
The annual incidence of unknown primary with squamous
cell neck node metastases remained stable with an average of
17 new cases per year ± 0.34 cases per 100 000/year ± in the
20-year study period. In the same period the annual Danish
incidence of head and neck carcinoma increased from 700 to
1000 new cases per year. The proportion of unknown primary
cancers relative to total head and neck cancers thus decreased
from 2.5 to 1.7% in the 20-year period. It may be due to better
initial diagnostic work-up, but this cannot be documented in a
retrospective study like this.
Combined neck dissection and postoperative radiotherapy was used infrequently in this series, as the general
approach has been primary radiotherapy with surgery
reserved for salvage. More aggressive multi-modality
approaches may be advantageous, but will be so at the
cost of increased morbidity. This approach cannot be evaluated from the results of the present series. However, the
observed survival rates are comparable with survival rates
observed in most reports from the last decade
[8,11,14,17,18,20,21]. Treatment outcome is similar to
what can be seen for patients with known primary sites
and similar extent of lymph node metastases. The study
also con®rms N-stage as being the single most important
factor for treatment outcome and survival.
One of the most controversial topics in the management of
occult head and neck primaries is the role of mucosal irradiation. In Denmark, mucosal irradiation to prevent emerging
primaries has been the rule over the last two decades. Irradiation of the entire pharyngeal axis and the larynx causes
signi®cant acute and late morbidity. Although not demonstrated in the present retrospective analysis, it is well known
from other head and neck series that radiation morbidity is
highly related to the irradiated volume. The arguments for
ipsilateral treatment have been that by sparing most of the
Table 5
Cox proportional hazards analysis of radiotherapy technique and signi®cant prognostic parameters from univariate analysis using death from any cause as
endpoint (n ˆ 232); patients treated with surgery only not included
Variable
N-stage
Haemoglobin
Gender
RT technique
RT dose
Overall treatment time
Events/censored
a
NE, not evaluated.
Coding
N1
N2
N3
High
Low
Female
Male
Neck 1 mucosa
Neck
661 Gy
1±65 Gy
1±49 days
501 days
Loco-regional failure
Death from cancer
Death
P-value
P-value
P-value
0.28
0.01
0.34
0.01
0.05
0.50
0.06
RR (95% CI)
1.0
1.4 (0.7±2.8)
2.6 (1.3±5.1)
1.0
NE a
1.0
2.1 (1.2±3.6)
1.0
1.9 (1.0±3.6)
1.0
NE
1.0
NE
114/115
0.98
0.04
0.08
0.04
0.10
0.75
0.04
RR (95% CI)
1.0
1.0 (0.6±1.8)
1.8 (1.0±3.3)
1.0
NE
1.0
1.6 (1.0±2.6)
1.0
NI.
1.0
NI.
1.0
1.5 (1.0±2.1)
122/108
0.20
0.02
0.28
0.28
0.31
0.53
0.07
RR (95% CI)
1.0
1.4 (0.8±2.2)
1.8 (1.1±3.0)
1.0
NE
1.0
NE
1.0
NE
1.0
NE
1.0
NE
177/53
128
C. Grau et al. / Radiotherapy and Oncology 55 (2000) 121±129
Fig. 5. Loco-regional control and disease-speci®c survival for patients treated with neck radiotherapy (n ˆ 26) versus neck 1 mucosa (n ˆ 224) after
adjusting for covariates using Cox proportional hazard model. Parameters in the model were N-stage, haemoglobin, gender, radiation dose and overall
treatment time.
mucosa and the entire contralateral neck, patients will tolerate treatment much better and have the same survival rate
[8,18,21]. On the other hand, the arguments for large mucosal
®elds have been to prevent potentially incurable loco-regional relapses [7,9,14,17,20]. This dilemma between safety and
morbidity is also evident from the data in the present study.
There was a signi®cant twofold reduction in loco-regional
relapses by using extensive radiotherapy ®elds, primarily due
to less neck recurrences. Patients treated with bilateral ®elds
had a higher disease-speci®c survival. There was only few
patients in the ipsilateral group, so the difference did not
reach statistical signi®cance (P ˆ 0:10). Radiotherapy in
either form signi®cantly reduced the risk of having an emerging mucosal primary when compared with patients treated
with surgery alone. In fact, the incidence of emerging
primary in either radiotherapy group was similar to the incidence of metachronous cancers in other head and neck cancer
series, where a constant rate of 3% per year have been
reported [1,10,12,15]. One explanation for the `mucosal
effectiveness' of ipsilateral treatment may be that it unintentionally involves some mucosal irradiation. This effect may
be especially important in sterilising potential lateral
tumours in the oropharynx, because this region lies just medially to the commonly involved sub-digastric nodes. From
such indirect evidence it seems logic to include the ipsilateral
tonsillar fossa and base of tongue if decision is made to use
unilateral ®elds.
In conclusion, this study has con®rmed that patients with
neck node metastases from occult head and neck cancer
have clinical features and prognosis similar to other head
and neck malignancies. Prognostic factors include nodal
stage, gender, haemoglobin, tumour differentiation and
overall treatment time. Most patients were treated with
extensive irradiation to both sides of the neck and the
mucosa in the entire pharyngeal axis and larynx This
resulted in signi®cantly less loco-regional failures compared
with patients treated with ipsilateral techniques, but only a
trend towards better survival. A prospective randomized
trial is required to determine the optimal strategy in terms
of loco-regional control, survival and morbidity.
Acknowledgements
The authors would like to thank the staff and residents of
Department of Radiation Oncology, Vancouver Cancer
Centre for providing excellent research opportunities for
C.G. His fellowship at the Vancouver Cancer Centre was
also supported by grants from the Danish Cancer Society,
Ingeniùr Paul Lundbeck og Hustru's Fond til Fremme af
Radiologien i Danmark, and Radiumstationens Forskningsfond, Aarhus, Denmark.
References
[1] Boysen M, Loven JO. Second malignant neoplasms in patients with
head and neck squamous cell carcinomas. Acta Oncol. 1993;32:283±
288.
[2] Braams JW, Pruim J, Kole AC, Nikkels PG, Vaalburg W, Vermey A,
Roodenburg JL. Detection of unknown primary head and neck tumors
by positron emission tomography. Int. J. Oral Maxillofac. Surg.
1997;26:112±115.
[3] Colletier PJ, Garden AS, Morrison WH, Goepfert H, Geara F, Ang
KK. Postoperative radiation for squamous cell carcinoma metastatic
to cervical lymph nodes from an unknown primary site: outcomes and
patterns of failure. Head Neck 1998;20:674±681.
[4] Coster JR, Foote RL, Olsen KD, Jack SM, Schaid DJ, DeSanto LW.
Cervical nodal metastasis of squamous cell carcinoma of unknown
origin: indications for withholding radiation therapy. Int. J. Radiat.
Oncol. Biol. Phys. 1992;23:743±749.
[5] Feinmesser R, Feinmesser M, Freeman JL, Noyek AM, Livni N.
Detection of occult nasopharyngeal primary tumours by means of
in situ hybridization. J. Laryngol. Otol. 1992;106:345±348.
[6] Feinmesser R, Miyazaki I, Cheung R, Freeman JL, Noyek AM, Dosch
HM. Diagnosis of nasopharyngeal carcinoma by DNA ampli®cation
C. Grau et al. / Radiotherapy and Oncology 55 (2000) 121±129
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
of tissue obtained by ®ne-needle aspiration. N. Engl. J. Med.
1992;326:17±21.
Freeman D, Mendenhall WM, Parsons JT, Million RR. Unknown
primary squamous cell carcinoma of the head and neck: is mucosal
irradiation necessary? Int. J. Radiat. Oncol. Biol. Phys. 1992;23:889±
890.
Glynne-Jones RG, Anand AK, Young TE, Berry RJ. Metastatic carcinoma in the cervical lymph nodes from an occult primary: a conservative approach to the role of radiotherapy. Int. J. Radiat. Oncol. Biol.
Phys. 1990;18:289±294.
Harper CS, Mendenhall WM, Parsons JT, Stringer SP, Cassisi NJ,
Million RR. Cancer in neck nodes with unknown primary site: role of
mucosal radiotherapy. Head Neck 1990;12:463±469.
Johansen LV, Overgaard J, Hjelm-Hansen M, Gadeberg CC. Primary
radiotherapy of TI squamous cell carcinoma of the larynx: analysis of
478 patients treated from 1963 to 1985. Int. J. Radiat. Oncol. Biol.
Phys. 1990;18:1307±1313.
Jones AS, Cook JA, Phillips DE, Roland NR. Squamous carcinoma
presenting as an enlarged cervical lymph node. Cancer 1993;72:1756±
1761.
Jovanovic A, van der Tol IGH, Kostense PJ, Schulten EAJM, de Vries
N, Snow GB, van der Waal I. Second respiratory and upper digestive
tract cancer following oral squamous cell carcinoma. Oral Oncol. Eur.
J. Cancer 1994;30B:225±229.
Keyes JWJ, Watson NEJ, Williams DW, Greven KM, McGuirt WF.
FDG PET in head and neck cancer. Am. J. Roentgenol. 1997;169:
1663±1669.
Maulard C, Housset M, Brunel P, et al. Postoperative radiation ther-
[15]
[16]
[17]
[18]
[19]
[20]
[21]
129
apy for cervical lymph node metastases from an occult squamous cell
carcinoma. Laryngoscope 1992;102:884±890.
McGarry GW, Mackenzie K, Periasamy P, McGurk F, Gatehouse S.
Multiple primary malignant tumours in patients with head and neck
cancer: the implications for follow-up. Clin. Otolaryngol. 1992;17:
558±562.
Mukherji SK, Drane WE, Mancuso AA, Parsons JT, Mendenhall
WM, Stringer S. Occult primary tumors of the head and neck: detection with 2-[F-18] ¯uoro2-deoxy-d-glucose SPECT. Radiology
1996;199:761±766.
Reddy SP, Marks JR. Metastatic carcinoma in the cervical lymph
nodes from an unknown primary site: results of bilateral neck plus
mucosal irradiation vs. ipsilateral neck irradiation. Int. J. Radiat.
Oncol. Biol. Phys. 1997;37:797±802.
Sinnathamby K, Peters LJ, Laidlaw C, Hughes PG. The occult head
and neck primary: to treat or not to treat? Clin. Oncol. (R. Coll.
Radiol.) 1997;9:322±329.
Walter MA, Menarguez-Palanca J, Peiper SC. Epstein-Barr virus
detection in neck metastases by polymerase chain reaction. Laryngoscope 1992;102:481±485.
Wang RC, Goepfert H, Barber AE, Wolf P. Unknown primary squamous cell carcinoma metastatic to the neck. Arch. Otolaryngol. Head
Neck Surg. 1990;116:1388±1393.
Weir L, Keane T, Cummings B, Goodman P, O'Sullivan B, Payne
D, Warde P. Radiation treatment of cervical lymph node metastases
from an unknown primary: an analysis of outcome by treatment
volume and other prognostic factors. Radiother. Oncol.
1995;35:206±211.