Tonsillectomy in diagnosis of the unknown primary tumor of
the head and neck
DAVID A. RANDALL, MD, PETER A. S. JOHNSTONE, MD, ROBERT D. FOSS,DDS, and PETER J. MARTIN, MD, San Diego, California
OBJECTIVES: The purpose of this study was to dis-
cuss the experience of one tumor registry with per-
forming tonsillectomy in the diagnostic approach
to unknown head and neck primary tumors. It also
describes the importance of including tonsillecto-
my in this evaluation algorithm.
STUDY DESIGN: A retrospective chart review was
done of 68 patients with either tonsillar or unknown
primary squamous cell carcinoma culled from 829
patients seen from 1956 to 1996 at the head and
neck tumor registry at the Naval Medical Center
San Diego.
METHODS: Records from the head and neck tumor
registry, radiation oncology service, and pathology
department were reviewed with attention to pre-
senting symptom, initial examination, diagnostic
studies performed, and type and result of biopsies
performed.
RESULTS:Thirty-four patients sought treatment for a
neck lymph node metastasis of squamous cell car-
cinoma without an identifiable primary tumor site.
Six of these (18%) had the primary site diagnosed
by performing tonsillectomy ipsilateral to the pre-
senting neck mass. Six of 14 T1 tonsillar carcinomas
in this series had the primary site identified by ton-
sillectomy.
CONCLUSIONS: Despite a diligent search, a prima-
ry tumor site may not be found in the head and neck
cancer patient. The tonsil may harbor an occult
squamous cell carcinoma. The patient benefits from
identification of the initial tumor site because
postoperative irradiation ports may be reduced and
because surveillance for recurrence
From the Departments of Otorhinolaryngology (Drs Randall and
Martin), Radiology/Oncology (Dr Johnstone), and Oral Pathology
and Clinical Investigation (Dr Foss), Naval Medical Center San
Diego.
The Chief, Bureau of Medicine and Surgery, Navy Department,
Washington, DC, Clinical Investigation Program sponsored this
report (no. S-95-092), as required by HSETCINST 6000.41A.
The views expressed in this article are those of the authors and do not
reflect the official policy or position of the Department of the Navy,
Department of Defense, or the US government.
Reprint requests: LCDR David A. Randall, MC, USN, c/o Clinical
Investigation Department, Naval Medical Center San Diego, 34800
Bob Wilson Dr, San Diego, CA 92134-5000.
23/1/99073
52
may be improved. For these reasons, tonsillectomy
should be performed ipsilateral to the presenting
cervical metastasis if no other primary tumor site is
identified. (Otolaryngol Head Neck Surg 2000;122: 52-
5.)
Although most primary tumor sites of the head and
neck region can be identified by physical examination,
a small number cannot. On occasion, a cancer will pre-
sent initially as a cervical nodal metastasis without an
identifiable source. In 2% to 8% of such cases, no pri-
mary malignancy is ever identified despite a diligent
search.1-6 Identification of a primary site is important
for several reasons. More specific therapy may be pro-
vided to that area. The site may be more closely scruti-
nized during subsequent observation for recurrence.
Treatment morbidity to surrounding areas is decreased
by identification of a specific location because the area
treated with radiotherapy may be reduced, thereby
decreasing the associated side effects.
Physical examination, including endoscopic evalua-
tion, identifies 90% of head and neck tumors.1-3,7 In
1944, Dr Hayes Martin described the diagnostic evalu-
ation of the unknown primary lesion and popularized
the use of rigid laryngoscopy, bronchoscopy, and esopha-
goscopy (triple endoscopy, panendoscopy) with direct-
ed biopsies. This remains the mainstay for diagnosis.2,8
In cases of unknown primary sites, it is fairly standard
practice to biopsy areas known to harbor occult tumors,
such as the nasopharynx, tongue base, palatine tonsil,
and pyriform sinus. The tonsil deserves special attention
as a potential source of hidden malignancy because
small tumors can originate in the depths of tonsil crypts
without being detected at the surface.9 Also, the rich
lymphatic drainage of the tonsil facilitates early spread
to regional nodes despite a low tumor volume of the pri-
mary site.
Panendoscopic biopsy is generally performed super-
ficially with cup- or biting-type forceps, which is appro-
priate for most mucosal sites but inadequate for tonsil
biopsy. Obtaining only a small superficial amount of tis-
sue can result in a false-negative specimen because of
sampling error or a “geographical miss.” Tonsillectomy
ipsilateral to the cervical metastasis has therefore been
recommended.5,9-12 Unfortunately, this procedure is
Otolaryngology–
Head and Neck Surgery
Volume 122 Number 1 RANDALL et al 53

Table 1. Tumor stage of patients with unknown and

tonsil primaries
Tumor size No. of patients

Tx 28
T1 14*
T2 11
T3 11
T4 4
TOTAL 68
*Six were identified by tonsillectomy, and 8 were identified clinically.

often omitted.13-16 It has been our experience that ton-

sillectomy plays an important role in the management
of the occult head and neck squamous cell carcinoma
(SCC) primary. Review of this institution’s tumor board
files demonstrates that tonsillectomy has provided the
diagnosis in a significant number of cases.

METHODS AND MATERIAL

This study was a retrospective chart review of patients
from the Naval Medical Center San Diego head and neck
tumor registry. Sources of data included medical records from
the otolaryngology and radiation oncology departments, out-
patient/inpatient charts, and pathologic slides and records.
Charts were reviewed with attention to presenting symptoms,
initial examinations, diagnostic studies performed, and biop-
sies obtained. Histology was reviewed on available Tx and T1 Fig 1. Poorly differentiate d SCC arises from the depth of a
tonsil tumors to determine whether superficial tonsil biopsy or tonsillar crypt.Tu mor was associate d with the normal lym-
phoid tonsil tissue and m e asure d less than 1.0 cm in
tonsillectomy was performed and whether this yielded a diag-
gre atest dim ension. (He matoxylin-eosin stain; original
nosis. Because of the relatively small numbers of patients, no magnificatio n at medium power.)
specific statistical analysis was performed.

RESULTS Although only 5 superficial tonsil biopsies were per-

Between 1956 and 1996 a total of 829 patients were
seen in the head and neck tumor board at the Naval
Medical Center San Diego. Of these, 38 (4.6%) were
unknown primaries and 56 (6.7%) were tonsillar carci-
nomas. Sixty-eight of these 94 charts were available for
review. Table 1 illustrates the proportions of tumor
types.
Of the patients in Table 1, 34 had tonsil primaries
that were identified on examination in the clinic or vis-
ible by endoscopy. The remaining 34 patients (28 with
Tx and 6 with occult T1) demonstrated no visible or
palpable disease. Of these 34 patients, 6 underwent ipsi-
lateral tonsillectomies, 5 had superficial tonsil biopsies,
and 2 had neither. Documentation of the procedures
performed on the remainder of patients was unavailable.
In this series, 6 of 14 T1 tonsil carcinomas were occult
and required tonsillectomy for diagnosis. And of the
total patients without identifiable primary sites, 6 of 34
(18%) had the diagnosis made by tonsillectomy.
formed, none yielded tumor. In the remaining 23
patients, either tonsillectomy/tonsil fossa biopsy was
not performed or documentation was unavailable. No
tonsillectomies performed were negative for tumor. No
complications were noted as a result of tonsillectomy.
Among the tonsil primaries reviewed in this study,
14 (35%) were T1, 11 (27.5%) were T2, 11 (27.5%)
were T3, and 4 (10%) were T4 lesions. Tonsil carcino-
mas represented 6.7% of our head and neck tumor reg-
istry. Other reviewers described the percentages of their
tonsil lesions as 10% to 34% T1, 33% to 37% T2, 19%
to 41% T3, and 10% to 17% T4.17-19 This suggests that
although this series is smaller and has slightly more T1
and fewer T2 lesions, its distribution of sizes compares
with those in other experiences. The increased number
of T1 lesions might be explained by the diagnoses made
on occult tonsil lesions. Unknown primary tumors rep-
resented 4.6% of our head and neck patients, consistent
with the 2% to 8% reported in the literature.1-6

54 RANDALL et al
HISTOLOGY
Histology was reviewed on all occult tonsil tumors
where diagnosis was made by tonsillectomy. All were
SCCs, varying from poorly differentiated to well differ-
entiated. Most primaries and metastatic nodal disease
were poorly to moderately differentiated with limited
keratinization. One less differentiated SCC produced a
well-differentiated metastasis to a cervical node. One
tonsil tumor was 1.5 cm in diameter, and the rest were
1 cm or less. The tonsillectomy specimens were 1 to 3
cm in greatest dimension, and the occult primaries in all
of these were located in crypts beneath the surface.
Figure 1 illustrates a representative example. The
metastatic tumor deposits showed at least some degree
of comedonecrosis or early cystic degeneration of
tumor islands.
DISCUSSION
Approaches to the diagnostic evaluation of the
unknown head and neck primary tumor appear rather
inconsistent in the literature, and its management is
somewhat controversial. Cervical node metastases are
often the first symptom of cancer. Search for the prima-
ry usually involves triple endoscopy, and most series
addressing this topic include the previously discussed
directed biopsies.4-6,8,15,20,21 Although the yield of these
biopsies is unclear, it appears to be low. However, little
additional operative time, effort, or risk is involved to
obtain them during endoscopy. Review of 204 patients
culled from 12 reports in which a primary site was ulti-
mately diagnosed supports the fact that traditionally
described locations such as the nasopharynx, tonsil,
tongue base, and hypopharynx remain likely
sites.1,4,5,8,11,14,15,22-27
Several sources recommend tonsillectomy. 9-11,28
Righi and Sofferman12 described 6 occult primary ton-
sil tumors identified in their small series, whereas
Lapeyre et al10 found carcinoma in 23 of 87 (26%) ton-
sillectomies. This compares with 43% of our T1 tonsil
tumors found this way.
Several reasons support the need to locate a primary
tumor site. Identification of a specific area permits
focused surveillance for recurrence. With respect to
radiotherapy, there are potential benefits to the patient
from properly defining a tonsillar primary cancer rather
than using a standard treatment protocol for an
unknown primary. Whereas the neck should be irradiat-
ed bilaterally in the presence of palpable node disease,
the doses and portals to the primary may be adjusted to
minimize salivary gland toxicity if a typical postopera-
tive tonsil field is treated. In the case of unknown head
and neck primaries, therapy may include coverage of
the at-risk mucosa (nasopharynx, oropharynx, hypophar-
Otolaryngology–
Head and Neck Surgery
January 2000
ynx) to 60 Gy by use of opposed lateral ports. This pro-
duces significant xerostomia. For a tonsillar regimen,
fields need not include the nasopharynx and can have
more parotid blocking. In fact, treatment volume may
be adjusted to treat the ipsilateral tonsil and spare the
contralateral parotid.
It is important to note that there is some controversy
as to whether prophylactic irradiation of potential head
and neck mucosal tumor sites is effective or justifies the
cost and morbidity.13 Several authors found that the
subsequent development of a primary tumor was asso-
ciated with a decrease in the 3- and 5-year survival rates
by roughly half.3,5,9,24 Therefore, despite the associated
morbidity, some physicians may prefer to avoid under-
treatment and the subsequent development of a primary
site.1,3,14,16,23,25-28
SUMMARY
Despite thorough evaluation, 2% to 8% of patients
with cervical metastases have no primary tumor site
identified. Although most series discussing this subject
do not include ipsilateral tonsillectomy in their diag-
nostic protocol, our experience has been that removal of
the entire tonsil provides a worthwhile yield. This
allows focused therapy, reduced morbidity, and improved
follow-up. We therefore recommend that in the absence
of an obvious primary site, the standard endoscopy with
directed biopsies also include ipsilateral tonsillectomy.
In the case of previous tonsillectomy, any residual ton-
sil tissue should be removed (if present), or the tonsil
fossa should be biopsied.
REFERENCES
1. Bataini J, Rodriguez J, Jaulerry C, et al. Treatment of metastatic
neck nodes secondary to an occult epidermoid carcinoma of the
head and neck. Laryngoscope 1987;97:1080-4.
2. Martin H, Romieu C. The diagnostic significance of a “lump in
the neck.” Postgrad Med 1952;11:491-500.
3. Million R, Cassisi N, Mancuso A. The unknown primary. In:
Million R, Cassisi N, editors. Management of head and neck can-
cer: a multidisciplinary approach. 2nd ed. Philadelphia: JB
Lippincott; 1994. p. 311-20.
4. Acquarelli M, Matsunaga R, Cruze K. Metastatic carcinoma of
the neck of unknown primary origin. Laryngoscope 1961;71:
962-74.
5. Coker D, Casterline P, Chambers R, et al. Metastases to lymph
nodes of the head and neck from an unknown primary site. Am J
Surg 1977;134:517-22.
6. DeBraud F, Heilbrun L, Sakr W, et al. Metastatic squamous cell
carcinoma of an unknown primary localized to the neck. Cancer
1989;64:510-5.
7. Freeman D, Mendenhall W, Parsons J, et al. Unknown primary
squamous cell carcinoma of the head and neck: is mucosal irra-
diation necessary? Int J Radiat Oncol Biol Phys 1992;23:889-90.
8. Leipzig B, Winter M, Hokanson J. Cervical nodal metastases of
unknown origin. Laryngoscope 1981;91:593-8.
9. Marcial-Vega V, Cardenes H, Perez C, et al. Cervical metastases
from unknown primaries: radiotherapeutic management and
appearance of subsequent primaries. Int J Radiat Oncol Biol Phys
1990;19:919-28.
Otolaryngology–
Head and Neck Surgery
Volume 122 Number 1 RANDALL et al 55

10. Lapeyre M, Malissard L, Peiffert D, et al. Cervical lymph node
metastasis from an unknown primary: is a tonsillectomy neces-
sary? Int J Radiation Oncology Biol Phys 1997;39:291-6.
11. Jakobsen J, Aschenfeldt P, Johansen J, et al. Lymph node metas-
tases in the neck from unknown primary tumor. Acta Oncol
1992;31:653-5.
12. Righi P, Sofferman R. Screening unilateral tonsillectomy in the
unknown primary. Laryngoscope 1995;105:548-50.
13. Coster J, Foote R, Olsen K, et al. Cervical nodal metastasis of
squamous cell carcinoma of unknown origin: indications for
withholding radiation therapy. Int J Radiat Oncol Biol Phys
1992;23:743-9.
14. Fitzpatrick P, Kotalik J. Cervical metastases from an unknown
primary tumor. Radiology 1974;110:659-63.
15. Fried M, Diehl W, Brownson R, et al. Cervical metastasis from
an unknown primary. Ann Otol 1975;84:152-7.
16. McCunniff A, Raben M. Metastatic carcinoma of the neck from
an unknown primary. Int J Radiat Oncol Biol Phys 1986;12:
1849-52.
17. Kaplan R, Million R, Cassisi N. Carcinoma of the tonsil: results
of radical irradiation with surgery reserved for radiation failure.
Laryngoscope 1977;87:600-7.
18. Mizono G, Diaz R, Fu K, et al. Carcinoma of the tonsillar region.
Laryngoscope 1986;96:240-4.
19. Givens C, Johns M, Cantrell W. Carcinoma of the tonsil. Arch
Otolaryngol 1981;107:730-4.
20. Johnson J, Newman R. The anatomic location of neck metastasis
from occult squamous cell carcinoma. Otolaryngol Head Neck
Surg 1981;89:54-8.
21. MacComb W. Diagnosis and treatment of metastatic cervical
cancerous nodes from an unknown primary site. Am J Surg
1972;124:441-9.
22. Barrie J, Knapper W, Strong E. Cervical nodal metastases of
unknown origin. Am J Surg 1970;120:466-70.
23. Harper C, Mendenhall W, Parsons J, et al. Cancer in neck nodes
with unknown primary site: role of mucosal radiotherapy. Head
Neck 1990;12:463-9.
24. Jesse R, Perez C, Fletcher G. Cervical lymph node metastasis:
unknown primary cancer. Cancer 1973;31:852-9.
25. Maulard C, Housset M, Brunel P, et al. Postoperative radia-
tion therapy for cervical lymph node metastases from an
occult squamous cell carcinoma. Laryngoscope 1992;102:
884-90.
26. Nguyen C, Shenouda G, Black M, et al. Metastatic squamous cell
carcinoma to cervical lymph nodes from unknown primary
mucosal sites. Head Neck 1994;16:58-63.
27. Silverman C, Marks J, Lee F, et al. Treatment of epidermoid and
undifferentiated carcinomas from occult primaries presenting in
cervical lymph nodes. Laryngoscope 1983;93:645-8.
28. Carlson L, Fletcher G, Oswald M. Guidelines for radiotherapeu-
tic techniques for cervical metastases from an unknown primary.
Int J Radiat Oncol Biol Phys 1986;12:2101-10.

5th International Muscle Symposium

The symposium will be held May 19-21, 2000, in Vienna, Austria.

For further information, contact Manfred Frey, MD, Division of Plastic and
Reconstructive Surgery, Department of Surgery, University of Vienna, Medical
School, Währinger Gürtel 18-20, A-1090 Vienna, Austria; phone, 43-1-40400-6986;
fax, 43-1-40400-6988; e-mail, manfred.frey@akh-wien.ac.at; URL, http://www.akh-
wien.ac.at/plastsurg.