The Patterns of Cervical Lymph Node

Metastases From Squamous Carcinoma

of the Oral Cavity
Jatin P. Shah, MD, FACS, Frank C. Candela, MD,
and Anil K. Poddar, MD

A retrospective review of the records of 501 previously untreated patients from

January 1,1965 through December 31,1986with squamous cell carcinoma of the
oral cavity was undertaken to ascertain the prevalence of ipsilateral neck node
metastases (NM) by neck level. The 501 patients underwent 516 radical neck
dissections. Patients were grouped by clinical neck status at the time of neck
dissection: elective dissection (ED) in the NO neck, immediate therapeutic dissection
(LTD) in the N+ neck, and subsequent therapeutic dissection (STD) in the neck
observed which converted clinically to N+. Pathologically identified NM occurred
34% of the time in ED, 69% in ITD and 90% in STD. The sensitivity, specificity, and
overall accuracy of the clinical exam was 70%, 65%, and 6870,respectively.
Detailed analysis was performed for each group based on the primary site. This
revealed a prevalence of NM in level IV of 3% (five of 167) for ED versus 17% (49/
+
296) for ITD STD (P < 0.001). Tongue, retromolar trigone, and cheek did not have
NM in level V in any group. The prevalence of NM in level V for floor of mouth or
gum primaries was < 1% (one of 109)in ED versus 6% (ten of 167)in ITD STD (P +
< 0.03).These data support the trend toward selective limited neck dissection in
both NO and N+ patients. Further, they provide the foundation for planning of
future prospective trials to assess the efficacy of modifications in the extent of neck
dissection. Cancer 66109-113,1990.

T HE ROLE OF LIMITED NECK DISSECTION in the man-

agement of regional lymph nodes for squamous cell
carcinoma of the oral cavity remains controversial. Rad-
ical neck dissection, since itsdescription by Crile in 1906,'
has been the traditional approach for excision of cervical
metastases in the neck. Classical radical neck dissection
offers comprehensive clearance of all cervical lymph node
groups, i.e.,level I through level V (Fig. 1). More recently,
modifications in the extent of lymphadenectomy, es-
pecially in patients with NO and N l necks, has been
championed by Bocca et al., Byers, and A pro-
spective, randomized comparison between radical neck
Presented in part at the 7 1st Annual Meeting of the American Radium
Society, St. Thomas, US Virgin Islands, April 15-19, 1989.
From the Head and Neck Service, Memorial Sloan-Kettering Cancer
Center, New York.
Address for reprints: Jatin P. Shah, MD, FACS, 1275 York Ave., New
York, NY 10021.
Accepted for publication March 13, 1990.
dissectionsand limited neck dissectionsin oral squamous
cell carcinomas is, however, lacking. In preparation for
such a comparative trial we present retrospective data to
demonstrate the prevalence and patterns of regional
lymph node metastases from primary oral squamous cell
carcinoma in patients who underwent radical neck dis-
sections.
Materials and Methods
From January 1, 1965 through December 31, 1986
there were 1234 patients with oral squamous cell carci-
noma registered on the Head and Neck Service at Me-
morial Sloan-Kettering Cancer Center (New York). A
retrospective review of the pathology reports of 501 con-
secutive patients with primary oral cavity squamous cell
carcinoma was undertaken to ascertain the prevalence
and distribution of neck node metastases by neck level.
Histologic variants of squamous cell carcinoma, i.e.,ver-

109
110 CANCERJuly I 1990 Vol. 66

rucous, spindle cell, etc., all were included in the study.

Radical neck dissections synchronous with the surgical
excision of the primary, and those done for subsequent
neck metastases were analyzed. The remaining 733 pa-

tients were excluded because of a lip primary, partial neck
dissection, multiple primaries, preoperative chemother-
apy, or preoperative radiation therapy. Fifty-five patients
were excluded because of inadequate data.
There were 5 16 radical neck dissections in the 50 1 pa-
tients. In 15 patients (3%), simultaneous bilateral neck
dissections were performed. In 452 patients (90%)the ex-
act count of lymph nodes pathologically examined was
reported and was 39 on average. In the remaining 10%
of patients, lymph node counts were reported as “nu-
merous,” or “matted” or “multiple.” Patients were con-
sidered to have nodal disease and counted as a positive
level when one or more nodes were reported to contain
tumor.
Patients were grouped by clinical neck status at the
time of neck dissection (Fig. 2). The three clinical groups
are (1) elective dissection in the clinically NO neck
(n= 192);(2) immediate therapeutic dissection in the clin-
ically N+ neck (n=206); and (3) subsequent therapeutic
dissection in the patients whose clinical exam progressed
to N+ during observation (n= 103). Each clinical group
was analyzed by specific primary site: oral tongue, floor
of mouth, gum, retromolar trigone, and cheek. The single
v
clinical N’+’
Subsequent
RND n=192 Therap. RND Therap. RND
(192 RND) n=103 (105) n=206 (219)
FIG. 2. Population distribution by clinical N status at first presentation.
patient with a hard palate primary was excluded from
subgroup analysis. The primary site and nodal staging
was done according to American Joint Committee on
Cancer (AJCC).5
Compilation of data was performed using the DBase
111-Plusprogram (Ashton-Tate, Torrance, CA). Statistical
analysis was performed using the chi-square technique.
Results
The population of 501 patients was composed of 357
male (71%) and 144 female (29%)patients. Age ranged
between 17 and 95 years (mean and median of 60 years).
The primary site distribution is shown in Table 1. Oral
tongue, floor of mouth, and gum accounted for 90% of
the primary sites. Most (65%)of the patients studied were
clinical Stage I1 or 111 (Table 2). The ‘T’ stage for each
clinical group is shown in Table 3.

NO neck (Elective Radical Neck Dissection)

One hundred ninety-two patients NO at presentation
underwent an immediate elective radical neck dissection.
Clinically occult, metastatic squamous cell carcinoma was
present in 65 of the 192 patients (34%).The overall prev-
alence of nodal metastases by neck level is shown in Table
4. Neck level IV nodes were involved in only 3% (six of
192) of elective neck dissections. The prevalence of level
IV involvement ranged from 2% to 6% depending on the

TABLE1. Population Distribution by Primary Site

Oral tongue 182 36%

Floor of mouth 162 33%
Gum 1 04 21%
Retromolar trigone 26 5%
Cheek 26 5%
Hard palate 1
FIG. 1. Cervical lymph node levels used for description of clinical and
50 1
pathologic findings.
No. 1 METASTASES
ORALCAVITY-NODAL Shah et al. 111

TABLE2. Population Distribution by Clinical Stage* TABLE4. Prevalence of Nodal Metastases by Neck Level

Stage No. Percent Elective Immediate Subsequent

RND 34%+ RND 69%+ RND 90%+
I 94 19% Level 65+/ 192 I52+/2 19 94+/105
.~
I1 146 29%
111 182 36% I 20% 48% 42%
IV 79 16% I1 17% 38% 54%
111 9% 31% 39%
50 1 IV 3% 15% 16%
V 0.5% 4% 2%
* TNM staging: Amencan Joint Committee on Cancer 1988.
RND: radical neck dissection.

primary site (Table 5). Of the six patients with level IV

involvement, three (50%)had nodal metastases at other
levels. Level V was involved in only one patient with a
gum primary who had all other neck levels also involved
with occult nodal metastases (Table 5).
N+ neck (Immediate Therapeutic
Radical Neck Dissection)
There were 2 19 radical neck dissections done in 206
patients who presented with a clinically N+ neck. In 69%
( 152/219) of neck dissections metastatic disease was con-
firmed histologically. The prevalence of nodal metastases
by neck level is shown in Table 4. Level IV was involved
with nodal metastases in 15% of therapeutic neck dissec-
tions. The prevalence ranged from 7% to 29% depending
on the primary site (Table 6). Only floor of mouth and
gum primaries had nodal metastases at level V, each oc-
curring 6% of the time. All of the patients (eight) with
level V involvement had nodal metastases at other levels.
The prevalence of level V metastases for tongue, retro-
molar trigone, and cheek primaries was zero (Table 6).
N+ neck (Subsequent Therapeutic
Radical Neck Dissection)
One hundred five radical neck dissections were done
in 103 patients initially staged NO. After surgical treatment
of the primary tumor, they developed a N+ neck requiring
therapeutic neck dissection. In this group, 90%of dissected
necks (94/105) had histologic confirmation of nodal me-
tastases. The median time for conversion of the initial
NO neck to N+ was 7 months. The prevalence of nodal
metastases by neck level is shown in Table 4. Level IV
was involved in 20% with tongue primaries and 12%each
with floor of mouth and gum primary sites (Table 7).
Only two patients with a floor of mouth primary had level
V involvement, both of whom also had other levels in-
volved by nodal metastases.
Accuracy of the Clinical Examination
The sensitivity, specificity and overall accuracy of the
clinical exam of the neck was 70%, 65%, and 68% re-
spectively. We found the sensitivity of the clinical exam
for nodal metastases in level V for floor of mouth and
gum lesions to be only 20%. There is a statistically sig-
nificant difference in the prevalence of nodal metastatic
disease between NO and N+ necks at level IV (P< 0.001)
and level V (P< 0.03) (Tables 8 and 9).
Discussion
The traditional approach for surgical treatment of neck
node metastases from primary squamous cell carcinoma
of the oral cavity has been a classical radical neck dissec-
tion. Since its description by Crile in 1906,' and after its
routine practice by Hayes Martin and others, it has be-
come the mainstay of surgical treatment for decades.6Al-
though the operation provides comprehensive clearance
of lymph nodes at all neck levels, it results in significant
functional and esthetic morbidity. Therefore, modifica-
tions of the classical radical neck dissection are proposed
to avoid its attendant morbidity.24 Bocca et al. and others
TABLE5. Prevalence by Primary Site: Elective
Radical Neck Dissection ( 192)

Tongue FOM Gum RMT Cheek

Level (18+/58) (15+/57) (20+/52) (7+/16) (5+/9)
TABLE3. T Status by Clinical Groups
I 14% 16% 27% 19% 44%
Elective Immediate Subsequent I1 19% 12% 21% 12% 11%
No N+ N+ 111 16% 7% 6% 6% 0
T (n = 192) (n = 206) (n = 103) IV* 3% 2% 4% 6% 0
1 29 31 64
vt 0 0 2% 0 0
2 111 99 36 FOM: floor of mouth RMT: retromolar trigone.
3
4
50
2
62
14
3
0
+
* Three of six associated with other levels (I, 11, 111).
t All five levels f.
112 CANCERJuly 1 1990 Vol. 66

TABLE6. Prevalence by Primary Site: Immediate TABLE8. Prevalence at Level IV*

Radical Neck Dissection (219)
Tongue FOM Gum Prevalence
Tongue FOM Gum RMT Cheek ~~

Level 42+/63 69+/98 27+/36 6+/1 13+/15 Elective 2/58 1/51 2/52 51167
No 3% 2% 4% 3%
I 30% 52% 56% 51% 80% Immediate TD 13/63 12/98 7/36
I1 41% 33% 39% 71% 40% N+ 21% 12% 19% 491296
111 38% 34% 19% 29% 1% Subsequent TD 13/66 2/17 2/16 17%
IV 21% 12% 19% 29% 7% Nf 20% 12% 12%
V* 0 6% 6% 0 0
FOM: floor of mouth; T D therapeutic dissection.
FOM: floor of mouth; R M T retromolar trigone. * Major sites, P < 0.00 I .
* Eight of eight associated with other + levels.

base their beliefs on anatomic studies demonstrating the
whole lymphatic system of the neck to be contained within
an aponeurotic By detaching this envelope
from underlying vital structures, lymph nodes and lym-
phatic channels can be extirpated without sacrifice of other
major neck structures. Bocca et aL2 believe that adequate
"cancerologic radicality" can be achieved by a functional
neck dissection. Indeed, recurrence rates were similar
when they compared results of their functional neck dis-
sections with historic controls undergoing radical neck
dissection.
Support is now mounting for a similar approach to
patients with limited nodal metastases at first presenta-
ti~n.''-'~Byers et u I . , ~in several retrospective studies from
the M. D. Anderson Hospital, have demonstrated the
modified neck dissection(s) to be adequate therapy for NO
or N 1 necks. Further, they advocate supraomohyoid neck
dissection for primary tumors of the oral cavity (and oro-
pharynx) in NO or N1 patients, if the surgeon is careful
enough to fully dissect level 111. However, evidence used
by Byers et al. in support of their conclusions is based on
a relatively small number of patients. With this in mind,
they found no statistically significant difference in recur-
rence rates between groups of patients undergoing supra-
omohyoid neck dissection, with or without radiation
therapy, compared with patients undergoing functional
neck dissection.
We believe that our retrospective study is unique
in that it represents a large number of patients with pri-
mary squamous cell carcinoma of the oral cavity, all of
whom underwent classical radical neck dissection. Thus,
detailed pathologic analysis of all cervical lymph node
levels was available in all patients. It is important to stress
that our study deals with the prevalence and patterns of
nodal metastases only. Since all nodal groups were excised
in all patients, the question of neck failure rates beyond
the dissected nodal groups does not arise. Further, only
a select group of clinical NO patients were treated with
elective neck dissection. Clearly, the exact reason for dis-
section in these patients is unknown. Presumably, they
represented patients at increased risk of harboring micro-
metastases, in the judgment of the surgeon.
In our clinical NO patients, if a supraomohyoid dissec-
tion had been performed, instead of a radical neck dis-
section, seven patients (3.5%) would have had nodal me-
tastases left behind: six of seven in level IV and one in
level V (Table 5). Only three of seven patients had isolated
level involvement outside of the supraomohyoid triangle
(level I, I1 or 111). These three represent only 1.5% of the
whole NO group. Clearly, the supraomohyoid dissection
would have removed the majority of lymph nodes har-
boring occult metastases in our NO patients. If level I, I1
or 111 was grossly positive at the time of surgery, or if
frozen section revealed nodal metastases, further dissec-
tion of levels IV and V in these 65 patients with positive
nodes, would have removed nodal metastases in four (6%).
This figure parallels the reported incidence of Byers et aL4
of local recurrence in 5% of pathologically NO patients

TABLE7. Prevalence by Primary Site: Subsequent

Radical Neck Dissection (n = 105*) TABLE
9. Prevalence at Level V*

Tongue FOM Gum RMT Cheek Tongue FOM Gum Prevalence

Level 58+/66 15+/11 15+/16 2+/3 2+/2
Elective 0158 0157 1/52 l/l09t
I 33% 59% 50% 33% 100% No 0% 0% 2% 1%
11 5990 41% 63% 33% 5090 Immediate TD 0163 6/98 2/36
I11 45% 24% 19% 66% 50% N+ 0% 6% 6% 10/167t
IV 20% 12% 12% 0 0 Subsequent TD 0166 2/17 0/16 6%
vt 0 12% 0 0 0 Nf 0% 12% 0%

FOM: floor of mouth; RMT: retromolar trigone. FOM: floor of mouth; TD: therapeutic dissection.
* Excludes hard palate, n = 1. * Major sites.
t Two of two associated with other + levels. t Excludes tongue, P < 0.03.
No. 1 METASTASES
ORALCAVITY-NODAL
undergoing a supraomohyoid neck dissection for oral
cavity or oropharyngeal primarie~.~ Further, a recent study
by Spiro et a/.,'' employing supraomohyoid neck dissec-
tion in clinically NO necks, revealed a 3.1% rate of neck
failure outside the operated field in pathologically NO pa-
tients.
If an anterolateral neck dissection (levelsI-IV) had been
performed on our clinically N+ patients, nodal metastases
would have been left behind in ten of 324 patients (3.2%)
at level V (Tables 6 and 7). All ten patients had nodal
disease at other levels. If the patients with histologically
confirmed nodal metastases at levels I through IV were
identified at the time of operation (246 necks with patho-
logically positive nodes) and a completion dissection were
carried out, then ten of 246 (4%)would be found to harbor
nodal metastases, beyond the limits of anterolateral neck
dissection.
Conclusion
Based on our series of 50 1 patients undergoing 5 16
radical neck dissections, we found neck levels I, 11, and
111to be at greatest risk of nodal metastases from primary
squamous cell carcinoma of the oral cavity. Level IV was
rarely involved (3%) in the NO patients, but was more
commonly involved (15%-16%) in the N+ groups (P
< 0.001). Fifty percent3 of the NO patients with level IV
nodal metastases had involvement of levels I or I1 or 111.
Neck level V was rarely involved in any major clinical
group, occurring only once in the NO group and in ten
patients in the N+ groups. Level V involvement was al-
ways associated with nodal metastases in other neck levels.
We believe that our data support the current trend toward
selective limited neck dissection in both NO and N+ pa-
tients.
Specifically, the supraomohyoid dissection for NO, and
the anterolateral neck dissection for N+ necks should re-
move the majority of nodal metastases. Identification of
positive nodes in the operating room should alert the sur-
geon regarding the potential for nodal disease at levels IV
- Shah ef al. 113
or V. The relative value of further nodal clearance in these
patients cannot be answered by our data. The addition of
adjuvant postoperative radiation therapy to the dissected
neck has been shown to reduce the rate of local recur-
rence. l4 Whether selective neck dissection and postoper-
ative radiation therapy is equally effective as radical neck
dissection and postoperative radiation is an issue which
remains unanswered. Clearly, prospective, randomized
trials comparing the morbidity and neck recurrence rates
in patients with oral cavity squamous cell carcinoma un-
dergoing selective limited neck dissection versus classical
radicala neck dissection is indicated.
REFERENCES
1. Cnle GW. Excision of cancer of the head and neck. JAMA 1906;
47: 1780- 1786.
2. Bocca E, Pignataro 0, Oldini C, Cappa C . Functional neck dissec-
tion: An evaluation of review of 843 cases. Laryngoscope 1984; 94:942-
945.
3. Byers RM. Modified neck dissection: A study of 967 cases from
1970-1980. Am JSurg 1985; 150414-421.
4. Byers RM, Wolf PF, Ballantyne AJ. Rationale for elective modified
neck dissection. Head Neck Surg 1988; 10:160-167.
5. American Joint Committee on Cancer. Manual for Staging of Can-
cer, ed. 3. Philadelphia: JB Lippincott, 1988; 33-35.
6. Martin H, Del Valle B, Ehrlich H, Cahan WG. Neck dissection.
Cuncer 1951; 4:441-499.
7. Suarez A. El problema de las metastasis linfaticas del cancer de
laringe and hipofaringe. Rev Otorrinolaringol 1963; 22:83.
8. Teatini GP, Zampano G. Lo svuotamento funzionale laterocer-
vicale: Basi anatomiche. Nuovo Arch It Otol 1976; 4:159-176.
9. Truffert P. Le Cou, les Aponeuroses, les Loges. Pans: Arnette, 1922
(in French).
10. Spiro JD, Spiro RH, Shah JP, Sessions RB, Strong EW. Supra-
omohyoid neck dissection: A critical assessment. Am J Surg 1988; 156:
286-289.
11. Lingeman RE, Helmus C, Stephens R, Ulm J. Neck dissection:
Radical or conservative. Ann Otol 1977; 861737-744.
12. Skolnik EM, King FY, Friedman M, Golden TA. The posterior
triangle in radical neck surgery. Arch Otolaryngol 1976; 102:l-4.
13. Sharpe DT. The Pattern of lymph node metastases in intra-oral
squamous cell carcinoma. Br J Plast Surg 1981; 34:97-101.
14. Vikram B, Strong EW, Shah JP, Spiro RH. Failure in the neck
following multimodality treatment in advanced head and neck cancer.
Head Neck Surg 1984; 6~724-729.