Anda di halaman 1dari 5

The Laryngoscope

VC 2009 The American Laryngological,

Rhinological and Otological Society, Inc.

Tonsillectomy vs. Deep Tonsil Biopsies

in Detecting Occult Tonsil Tumors
Joshua D. Waltonen, MD; Enver Ozer, MD; David E. Schuller, MD; Amit Agrawal, MD

Objectives: To compare the yield of detecting occult
Otolaryngologists frequently encounter patients with a
tumors by performing tonsillectomy and deep tonsil
biopsies in patients with metastatic carcinoma to the neck mass of unknown etiology. In many instan-ces, these
neck in whom a primary tumor was not evident on masses represent a lymph node with metastatic carcinoma.
physical examination, radiography, or panendoscopy. A primary tumor of the upper aerodigestive tract can be
Study Design: Retrospective study. detected by physical examina-tion and office endoscopy in
Methods: Over a 10-year period, the charts of all about 50% of cases.1,2 There remains a sizeable group of
patients diagnosed with metastatic carcinoma to the patients with occult primary tumors, comprising 3% to 5% of
neck in whom a primary tumor was not evident on head and neck malig-nancies.3 This problem remains a
examination, imaging, and panendoscopy were diagnostic and therapeutic challenge to practitioners, with a
reviewed. Specifically, we compared groups of patients good deal of effort directed at identifying the primary tumor
who had undergone diagnostic tonsillectomy with those
site. Successful identification of the primary tumor obviates
who had undergone deep tonsil biopsies in search of an
the need for wide-field radiotherapy, considerably reduc-ing
occult primary tumor.
Results: One hundred twenty-two patients were the morbidity of treatment to patients.4
identified as meeting study criteria. All underwent
directed biopsies of the tongue base, hypopharynx, and Various studies have evaluated the effectiveness of
nasopharynx as well as either deep tonsil biopsies or various imaging modalities in identifying occult primary
diagnostic tonsillectomy. Deep tonsil biopsies were per- tumors, and most authors recommend some type of
formed in 95 patients. In this group, 11 primary tumors radiologic study. It has long been standard to examine such
were subsequently identified (11.6%): three in the ton-sil, patients under anesthesia with panendoscopy as well.
six in the tongue base, one in the hypopharynx, and one Should a primary tumor remain elusive after imag-ing and
in the nasopharynx. Tonsillectomy was performed in 27 endoscopy, it is common practice among surgeons to obtain
patients. Eleven primary tumors were identified (40.7%): tissue specimens from mucosal sites most likely to harbor
eight in the tonsils, two in the tongue base, and one in occult primary tumors. Biopsy sites often include the
the hypopharynx. All tonsil primary tumors identified were nasopharynx, tongue base, hypophar-ynx, and palatine
ipsilateral to the presenting neck mass. The overall yield
tonsils. These biopsies have been variably referred to as
of finding an occult primary carcinoma in the tonsil was
3.2% for deep tonsil biopsies vs. 29.6% for ‘‘blind,’’5 ‘‘random,’’6 or ‘‘directed’’7 biopsies.
tonsillectomies (P < .0002).
Conclusions: Tonsillectomy offers a signifi-cantly One particular area of controversy concerns the
higher likelihood of finding occult tonsillar tumors than optimal method to obtain tissue from the palatine tonsil
deep tonsil biopsy. region. While some authors advocate deep biopsies from
Key Words: Unknown primary, occult primary, the tonsillar fossae,8–10 recent articles have recom-mended
tonsillectomy, tonsil biopsy, head and neck carcinoma.
Laryngoscope, 119:102–106, 2009 unilateral or even bilateral tonsillectomy.11–14 We are
unaware of any study, however, comparing the results of
From the Department of Otolaryngology–Head and Neck Surgery, The tonsillectomy to random deep biopsy of the tonsils in this
Ohio State University, Columbus, Ohio, U.S.A. situation. At our institution, patients diag-nosed with
Editor’s Note: This Manuscript was accepted for publication August 22, metastatic carcinoma of the neck in whom prior
2008. examination, imaging, and panendoscopy does not reveal
Dr. Waltonen is now at Wake Forest University, Winston-Salem, North
Carolina, U.S.A.
an obvious primary tumor typically undergo directed
Presented at the Triological Society Annual Meeting, Orlando, Florida,
biopsies obtained from the nasopharynx, base of tongue,
U.S.A., May 3, 2008. and hypopharynx. Additionally, patients variably undergo
Send correspondence to Amit Agrawal, MD, Department of Otolar- either deep tonsillar biopsies or tonsillectomy at the
yngology–Head and Neck Surgery, The Ohio State University, Cramblett Hall, discretion of the attending surgeon. We therefore perform
Suite 4A, 456 W. 10th Avenue, Columbus, Ohio, 43210. E-mail: this retrospective review in hopes of comparing the results
of each of these methods in detecting occult tonsil
DOI: 10.1002/lary.20017 malignancy.

Laryngoscope 119: January 2009 Waltonen et al.: Tonsillectomy vs. Tonsil Biopsy
From January 1997 through June 2007, clinic, operative, and Distribution of Neck Metastases by Level.
pathologic records from the Head and Neck Oncology Divi-sion at
the Arthur G. James Cancer Hospital and Richard J. Solove Tonsil
Neck Overall Biopsy Tonsillectomy
Research Institute at The Ohio State University were reviewed in level (n ¼ 122) (n ¼ 95) (n ¼ 27)
patients diagnosed with metastatic carcinoma of unknown primary
origin in whom all prior evaluation including examination, imaging, I 9 8 1
and panendoscopy (prior to biopsies/ton-sillectomy) had failed to II 84 66 18
yield an obvious primary tumor location. Records from our clinic as III 41 31 10
well as those from referring physicians’ offices were included for IV 10 9 1
review. Patients were excluded if there was a prior history of head V 14 11 3
and neck cancer, if the histopathology of the presenting neck mass
NA* 9 6 3
was other than squamous cell carcinoma or poorly differentiated
carcinoma, or if they had received treatment for their cancer prior to
In many patients, more than one level were involved, either by multi-
referral. Data were analyzed using Fisher’s exact tests. Institutional ple nodes or by large solitary masses.
Review Board approval for this study was obtained from the Office *Unable to determine from records which level(s) were involved.
of Responsible Research Practices at The Ohio State University. These masses had all been excised prior to referral to our institution.

RESULTS primary tumors identified arising from the tonsil, eight

(36.4%) from the base of tongue, two (9.1%) from the
During the time period of January 1997 through June
hypopharynx, and one (4.5%) from the nasopharynx. Two
2007, we identified 122 patients meeting inclusion criteria.
tongue base primary tumors were identified contra-lateral to
All were diagnosed with metastatic carcinoma of the neck of the presenting neck mass; the rest of the tumors including
unknown primary origin in whom prior examination, imaging all tonsil primaries were found ipsilat-eral to the presenting
studies, and panendoscopy carried out at our institution had neck mass.
failed to reveal a suspicious primary lesion. All patients had
Among the 95 patients who underwent deep biop-sies
undergone at least one cross-sectional and/or metabolic
of the tonsils along with the directed biopsies of other sites,
imaging modality including computed tomography (CT) scan
11 primary tumors were identified (11.6%). Three were
of the neck (81%), magnetic resonance imaging (MRI) of the
located in the tonsils, six in the base of tongue, and one
neck (7%), whole-body 18-fluorodeoxyglucose positron
each in the hypopharynx and nasopharynx.
emis-sion tomography (18-FDG PET) scan (26%), and/or
whole-body PET-CT fusion study (21%). Of the 122
Among the 27 patients who underwent tonsillec-tomy,
patients, 90 (74%) were male and 32 (26%) were female.
a total of 11 primary tumors were identified (40.7%),
The ages ranged from 31 to 86, with a median age of 56
significantly more than the tonsil biopsy group (P ¼ .013).
years. Histopathology of the presenting neck mass for the
Eight tonsil primary tumors were detected, two in the base
vast majority of patients in this study revealed met-astatic
of tongue and one in the hypopharynx. The yield of
squamous cell carcinoma (92%), with the remainder
tonsillectomy in detecting occult tonsillar tumors (29.6%)
revealing poorly differentiated carcinoma. One patient had
was significantly higher than that obtained by deep tonsil
synchronous poorly differentiated carci-noma and
biopsy (3.2%) (P < .0002). Table
squamous cell carcinoma in the left and right necks,
II details the identification of primary tumors among these
Of the 100 patients in whom workup including biop-
Following negative panendoscopy, all patients
sies and/or tonsillectomy did not reveal a primary tumor site,
underwent bilateral biopsies of the tongue base, hypo-
three primary tumor locations were subsequently
pharynx, and nasopharynx. Additionally, bilateral biopsies of
discovered after completion of treatment: one each in the
the tonsillar fossae were performed in 95 patients, while
lung, nasal cavity (arising from an inverted papilloma), and
tonsillectomy was performed in 27 (16 unilateral, 11
tonsil. The patient with the tonsil cancer had under-gone
bilateral). The decision to perform tonsil-lectomy vs. tonsil
tonsil biopsies, which did not reveal a primary tumor.
biopsy was made according to the different preferences of
the three attending surgeons of these patients. There were
no complications attributable to the panendoscopy, biopsies,
or tonsillectomies.
Table I details the distribution of lymphadenopathy in DISCUSSION
the patients’ necks by level. There was no significant Patients with head and neck cancer quite fre-quently
difference between the two groups. present to the clinician with a neck mass as the only
symptom. Physical examination has been shown to reveal
about 50% of the primary tumors.1,2 Undoubtedly, flexible
Identification of Primary Tumor nasopharyngoscopy in the office setting has allowed
Of the 122 patients, 22 primary tumors were subse- detection of even more primary tumors, espe-cially in the
quently identified based upon the ensuing biopsies and/ or nasopharynx and hypopharynx. It is not uncommon,
tonsillectomies (18.0%). Eleven patients (50%) had however, that a primary tumor remains

Laryngoscope 119: January 2009 Waltonen et al.: Tonsillectomy vs. Tonsil Biopsy
TABLE II. lectomy in 87 patients with unknown primary tumors and
Identification of Occult Primary Tumors. identified 23 primary tumors (26.4%) in this manner.5
Mendenhall et al. performed tonsillectomy in 34
Overall Biopsy Tonsillectomy P patients and identified primary tumors in 12 tonsils (35.2%).
In 11 of those, suspicion of a primary tumor was noted on
Number of patients 122 95 27
examination or imaging prior to the tonsil-lectomy. Among
Primary tumors identified 22 11 11 P ¼ .013
the 10 patients without evidence of a lesion on examination
Tonsil 11 3 8 P <.0002 or imaging, only one primary tu-mor (10%) was identified by
Tongue base 8 6 2 tonsillectomy.7 Randall et al. describe a series of 34
Hypopharynx 2 1 1 unknown primary tumors; six patients underwent ipsilateral
Nasopharynx 1 1 0 tonsillectomies. Six pri-mary tumors (18%) were discovered
in this manner.12
McQuone et al. have also recommend tonsillectomy in
this group of patients. In a 1998 series by this group, 13 23
patients without suspicion of primary tumor by examination
unknown after careful examination as well as cross-sec- or imaging underwent unilateral (7) or bilateral (16)
tional and/or metabolic imaging. tonsillectomy. Nine of the 23 patients (39%) had squamous
Evaluation of these patients under anesthesia with cell carcinoma diagnosed via tonsil-lectomy. These authors
panendoscopy has long been considered an accepted also make compelling arguments for performing bilateral
standard of care.2,15 For those patients without suspi-cious tonsillectomies in this group of patients, 14 whereby in a
lesions on endoscopy, Barrie10 and MacComb16 were subsequent update of their series reported in 2001, the
among the early proponents of adding ‘‘blind’’ biop-sies of authors describe four cases (10% of the total group of
the nasopharynx and tongue base, although neither patients with unknown primary tumors) in which carcinoma
addressed the tonsils. Jesse described successful ‘‘random’’ was identified in the tonsil contralat-eral to the presenting
biopsies from these locations and recom-mended adding neck mass. Although in our current series we did not identify
biopsies of the tonsils and pyriform sinus as well.8 With the occult primary tumors in the tonsil contralateral to the
recognition of the tonsil region as a significant site for presenting neck mass, via either random biopsy or excision
of the contralateral tonsil, it is likely that the added morbidity
primary tumors to emerge after treatment,9,10 many of carrying out contralat-eral tonsillectomy is minimal given
subsequent series have included random biopsies of the the overall low complication rate. Additionally, another
tonsil as well. However, the over-all yield of random biopsies theoretical advant-age of performing bilateral tonsillectomy
in identifying occult primary tumor location was recognized over ipsilateral tonsillectomy alone is the prevention of a
to be quite low, ranging from 2% to 15% in patients without
confusing asym-metric physical examination in future
suspicious lesions on examination or radiography.17,18 surveillance for recurrent or second primary tumors.
The use of tonsillectomy during the search for primary The yield of identifying occult tonsil malignancy via
tumors was infrequently mentioned during the 1980s.19,20 tonsillectomy in our series was 29.6%, which is in keep-ing
Righi and Sofferman in 1995 described a series of 19 with the range of 10% to 39% reported in several prior
patients, all of whom underwent tonsillectomy ipsilateral to series. The results of the aforementioned series are
the neck mass.11 Six primary tumors (32%) were identified summarized in Table III. Combining the results of all
via tonsillectomy. Similarly, Lapeyre et al. performed tonsil-

Tonsillectomy in the Detection of Occult Primary Tumors.
Number of
Number of Number of Positive Tonsil
Patients with Patients Specimens
Unknown Undergoing Obtained via
Series Primary Tumor Tonsillectomy Tonsillectomy (%) Comment
Righi and Sofferman, 1995 19 19 6 (31.5%)
Lapeyre et al., 1997 87 87 23 (26.4%)
Mendenhall et al., 1998 130 34 12 (35.2%) Eleven had suspicion of tonsil lesion on
exam and/or imaging.
Randall et al., 2000 34 27* 6 (22.2%) *Authors could not tell if tonsillectomy
was performed or not in 21; they are
included here in the denominator.
Koch et al., 2001 41 41 16 (39.0%) Two were in contralateral tonsil and two
in bilateral tonsils.
Present series 122 27 8 (29.6%)
Total 433 235 71 (30.2%) None of these series document any
complications from tonsillectomy.

Laryngoscope 119: January 2009 Waltonen et al.: Tonsillectomy vs. Tonsil Biopsy
these previous series with our current series, 71 of 235 field radiation therapy, compel us to recommend at least
(30.2%) of patients who underwent tonsillectomies had ipsilateral tonsillectomy (over random deep tonsil biopsy)
occult tumors detected in the tonsil. In our current series, along with directed biopsies of other sites including tongue
the yield of finding occult tonsil malignancy in patients base in patients diagnosed with meta-static carcinoma to
undergoing diagnostic excisional tonsillec-tomy was the neck of unknown primary origin.
significantly higher (29.6%) than in those patients who
underwent tonsil biopsies (3.2%, P < .0002).

Questions arise as to the reason why the tonsil and

tongue base are sites commonly found to harbor occult BIBLIOGRAPHY
tumors. Randall et al. discuss the histopathol-ogy of the six 1. Jones AS, Cook JA, Phillips DE, Roland NR. Squamous car-
occult T1 tonsil cancers discovered via tonsillectomy. 12 In all cinoma presenting as an enlarged cervical lymph node.
Cancer 1993;72(5):1756–1761.
six cases, tumors were less than 1.5 cm, and each
2. Martin H, Morfit HM. Cervical lymph node metastases as the
appeared to originate in submucosal crypts of the tonsil. first symptom of cancer. Surg Gynecol Obstet 1944; 78:133.
Mancuso et al. describe nine patients with cancer of
unknown primary. CT scan was performed and detected 3. Mahoney EJ, Spiegel JH. Evaluation and management of
submucosal tumor in each patient.21 All nine patients had malignant cervical lymphadenopathy with an unknown
primary tumor. Otolaryngol Clin North Am 2005;38(1): 87–97,
normal-appearing mu-cosa overlying the tumors, examined viii–ix.
by multiple clinicians. As such, it is likely that lesions arising 4. Reddy AN, Eisele DW, Forastiere AA, Lee DJ, Westra WH,
in tonsillar crypts harbor potential for submucosal growth Califano JA. Neck dissection followed by radiotherapy or
and development of regional metastases while remaining chemoradiotherapy for small primary oropharynx carci-noma
undetectable via visual and/or endoscopic examination. with cervical metastasis. Laryngoscope 2005;115(7): 1196–
5. Lapeyre M, Malissard L, Peiffert D, et al. Cervical lymph node
metastasis from an unknown primary: is a tonsillec-tomy
Lapeyre et al. discussed the concept of ‘‘geographic necessary? Int J Radiat Oncol Biol Phys 1997;39: 291–296.
miss’’ in tonsil biopsies. In other words, obtaining a small
6. Gluckman JL, Robbins KT, Fried MP. Cervical metastatic
amount of tonsil tissue with cupped forceps has a high
squamous carcinoma of unknown or occult primary source.
likelihood of a false negative result if a small, clini-cally Head Neck 1990;12(5):440–443.
undetectable tumor is present.5 Tonsillectomy eliminates the 7. Mendenhall WM, Mancuso AA, Parsons JT, Stringer SP, Cassisi
possibility of such sampling error, with little added morbidity NJ. Diagnostic evaluation of squamous cell carci-noma
over biopsy alone. Of course, the possibility of geographic metastatic to cervical lymph nodes from an unknown head
and neck primary site. Head Neck 1998; 20:739–744.
miss exists with small tumors of other sites including the
tongue base, an area for which there is no low-morbidity 8. Jesse RH, Perez CA, Fletcher GH. Cervical lymph node me-
excisional procedure such as tonsillectomy. However, given tastasis: unknown primary cancer. Cancer 1973;31(4): 854–
the finding in our series that 11 occult tumors (50% of 859.
9. Silverman CL, Marks JE, Lee F, Ogura JH. Treatment of
identified pri-mary tumors) were detected in locations other epidermoid and undifferentiated carcinomas from occult
than the tonsils via directed biopsies, we feel that biopsies primaries presenting in cervical lymph nodes. Laryngo-scope
of apparently normal mucosal sites, particularly the base of 1983;93:645–648.
tongue, remain warranted, and this practice should be 10. Barrie JR, Knapper WH, Strong EW. Cervical nodal metas-
continued to improve overall yield of primary tumor tases of unknown origin. Am J Surg 1970;120(4):466– 470.
11. Righi PD, Sofferman RA. Screening unilateral tonsillectomy in
the unknown primary. Laryngoscope 1995;105:548– 550.
If a primary tumor remains undiscovered, treat-ment
recommendations usually include wide-field radiation to 12. Randall DA, Johnstone PAS, Foss RD, Martin PJ. Tonsil-
putative mucosal sites and bilateral necks, with or without lectomy in diagnosis of the unknown primary of the head and
neck. Otolaryngol Head Neck Surg 2000; 122:52–55.
neck dissection.6,22 Radiation treatment in this setting
results in considerable morbidity for these patients, such as 13. McQuone SJ, Eisele DW, Lee DJ, Westra WH, Koch WM.
Occult tonsillar carcinoma in the unknown primary.
permanent xerostomia.23,24 On the other hand, should a Laryngoscope 1998;108(11 Pt 1):1605–1610.
primary tonsillar tumor be identi-fied, morbidity is greatly 14. Koch WM, Bhatti N, Wiliams MF, Eisele DW. Oncologic ra-
reduced. In such cases, the ipsilateral oropharynx and neck tionale for bilateral tonsillectomy in head and neck squa-
can be irradiated, with sparing of the contralateral parotid mous cell carcinoma of unknown primary source. Otolar-
yngol Head Neck Surg 2001;124:331–333.
gland4 and likely improvements in quality of life.
15. Fitzpatrick PJ, Kotalik JF. Cervical metastases from an
unknown primary tumor. Radiology 1974;110(3):659– 663.
In summary, we find that diagnostic excisional ton-
sillectomy provides about a 30% chance of identifying occult 16. MacComb WS. Diagnosis and treatment of metastatic cervi-cal
tonsil malignancy, yet adds little morbidity over that of cancerous nodes from an unknown primary site. Am J Surg
random deep biopsies of the tonsils, which com-paratively 1972;124(4):441–449.
have very low yield of detecting primary tumor location. 17. Fried MP, Diehl WH, Brownson JR, Sessions DG, Ogura JH.
Cervical metastases from an unknown primary. Ann Otol
These findings, in concert with the likely benefit of allowing Rhinol Laryngol 1975;84:152–157.
more precise treatment plan-ning while reducing the toxicity 18. Mendenhall WM, Mancuso AA, Amdur RJ, Stringer SP, Vil-laret
associated with wide- DB, Cassisi NJ. Squamous cell carcinoma

Laryngoscope 119: January 2009 Waltonen et al.: Tonsillectomy vs. Tonsil Biopsy
metastatic to the neck from an unknown head and neck 22. Chepeha D, Koch W, Pitman K. Management of unknown
primary site. Am J Otolaryngol 2001;22(4):261–267. primary tumor. Head Neck 2003;25(6):499–504.
19. Leipzig B, Winter ML, Hokanson JA. Cervical nodal metas- 23. Colletier PJ, Garden AS, Morrison WH, Goepfert H, Geara F,
tases of unknown origin. Laryngoscope 1981;91(4):593– Ang KK. Postoperative radiation for squamous cell car-
598. cinoma metastatic to cervical lymph nodes from an unknown
20. Lee DJ, Rostock RA, Harris A, Kashima H, Johns M. Clini-cal primary site: outcomes and patterns of failure. Head Neck
evaluation of patients with metastatic squamous car-cinoma 1998;20(8):674–681.
of the neck with occult primary tumor. South Med J 24. Harper CS, Mendenhall WM, Parsons JT, Stringer SP, Cas-sisi
1986;79:979–983. NJ, Million RR. Cancer in neck nodes with unknown primary
21. Mancuso AA, Hanafee WN. Elusive head and neck carcinomas site: role of mucosal radiotherapy. Head Neck 1990;12:463–
beneath intact mucosa. Laryngoscope 1983;93(2): 133–139. 469.

Laryngoscope 119: January 2009 Waltonen et al.: Tonsillectomy vs. Tonsil Biopsy