Keywords
patterns of lymph node metastasis - oral cavity cancer - risk factors for metastasis
- regional lymph node - skip metastasis - cervical lymph node
Introduction
Oral cavity cancers are the most prevalent cancers in Indian population, particularly
in the male population and are on rising trends. This is in large part because of
ill habits of tobacco chewing, smoking, and alcohol intake.[1]
[2]
[3] These cancers particularly spread to lymph nodes in neck in an orderly manner.[1] Various investigations are done to know the status of neck nodes before commencing
treatment for oral cavity cancer such as ultrasonography (USG) of neck, CT scan, MRI
scan, PET scan, and USG-guided FNAC from the neck nodes. In various studies, it has
been found that subsites of oral cavity behave differently, as far as the anatomical
patterns of lymphatic spread and tumor biology are concerned.[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11] Most of the literature available so far is from western world where oral cancer
is less prevalent than our country. This study is an attempt to collect information
on oral cavity cancers with respect to the tumor factors such as subsite, T stage,
depth, histological differentiation, etc. and their relation to incidence and pattern
of nodal metastases in neck.
Material and Methods
This study was conducted in the Department of Surgical Oncology and Department of
Pathology at a tertiary care cancer hospital of northwest India. It is a prospective
observational study.
The inclusion criteria were—245 patients of carcinoma buccal mucosa, anterior two-thirds of tongue, hard
palate, oral surface of soft palate, floor of mouth, vestibule, and alveolus.
The exclusion criteria were—patients who had received preoperative chemotherapy or radiotherapy and patients
with recurrent disease.
Patients of both sexes, all ages, presenting with carcinoma of oral cavity who were
subsequently planned for treatment were included in the study. Data were evaluated
with respect to subsite, T stage of cancer, tumor grading, lymphovascular and perineural
invasion by preoperative clinical findings, radiological findings, and histopathological
evaluations of the resected specimen and correlated with nodal spread in neck with
respect to number of nodes, level of spread, and size of nodes.
All the patients underwent required hematological and radiological investigations
and the resected specimen was processed for histopathological analysis. All the data
was then entered in a performa designed for the study.
Method of Surgery
All the patients underwent oncological resection which includes wide local excision
and neck dissection. The extent of neck dissection was dictated by spread of disease
as appreciable by preoperative clinical findings and radiological findings and was
based on existing protocols for management of oral cavity cancers.
All the levels of nodal spread as defined anatomically were removed separately and
sent for histopathological examination as separate specimens.
Specimen of resected tumor and nodal dissection was histopathologically evaluated
by Department of Pathology. Correlation was made between the tumor stage (T stage),
tumor characteristics, and pattern of nodal spread in neck with respect to number,
size, and level of nodal involvement.
Statistical Methods
All data were analyzed using SPSS 18.0 and Graphpad prism 7 software for statistical
analysis. Count data have been expressed as percentages (%). The χ
2 test was used for univariate analysis of the risk factors of cervical lymph node
metastasis. The odds ratio value (with 95% confidence interval) was used to express
the risk of cervical lymph node metastasis. p-Value of <0.05 was considered as the difference with statistical significance.
Results
In the present study of 245 cases, the most common age group of the patients suffering
from oral cancer (37.5%) was 41 to 50 years, with median age for males and females
being 45 years. Age group <40 years comprised approximately 30% of the patients, while
patients older than 50 years comprised approximately 33% of the population. Male to
female ratio was 3.8:1. Out of 245 patients, 226 (92.2%) were using tobacco in some
form. With regard to the presence of premalignant conditions, 33 patients suffered
from preoperative submucous fibrosis, 16 from leukoplakia, and there was one case
of erythroplakia. Of particular importance is the distribution of subsites within
oral cavity, as each subsite is known to have different rates of metastasis and different
patterns of spread to regional lymph nodes. As this is the primary subject of this
present paper, the topic will be discussed in detail in the following comments. Among
the 245 patients, 116 (47.3%) had history of smokeless tobacco, while 54 (22.04%)
had history of beedi smoking. In patients with history of beedi smoking, the most
common primary site of primary tumor was buccal mucosa, the next most common being
floor of mouth, tongue, palate and lower alveolus, in descending order of frequency.
Similarly, in patients presenting with history of smokeless tobacco, the most common
sites of primary disease were buccal mucosa, lower alveolus, tongue, palate, and floor
of mouth in descending order of frequency.
Buccal mucosa was the most common subsite of involvement in oral cavity followed by
tongue, with 113 (46.12%) patients having their primary lesion in buccal mucosa and
58 (23.67%) in tongue, respectively. The next most common sites were lower alveolus,
floor of mouth, retromolar trigone, and palate in descending order of frequency. Patients
having pathological positive nodes were highest among patients with primary subsite
at lower alveolus (63.15%), followed by tongue (50%). Though buccal mucosa was the
most common primary site in oral cavity cancers, pathologically positive nodes were
found only in 33% of the patients. This difference in nodal metastasis as per subsite
was highly significant (Chi-square = 19.410, degrees of freedom/df = 5, p = 0.0002).
Patients were staged according to the 8th edition of AJCC staging system. The stage
distribution of disease was contrary to what is found in western countries. Majority
of cases presented in Stage IVa (n = 90; 36.7%); rest in Stage II (n = 67; 24.8%), Stage I (n = 56; 22.8%) and Stage III (n = 38; 15.5%) in descending order of frequency.
Among 94 patients with pathologically positive nodes, 22 (23.40%) had no node on clinical
examination (clinically occult metastases). All the patients with clinically occult
metastases, had nodal involvement above or at level III. Out of 139 patients with
clinically positive nodes, 5% of the patients had nodal involvement below level III,
i.e., either level IV or VB.
Patterned nodal metastasis was seen in 93.5% (130/139 patients who underwent modified
radical neck dissection) of the patients, with initial involvement of level IB/IIA,
followed by level III and subsequent involvement of level IV and V. Patterned metastasis
was seen in 96% of the patients when patients who underwent supraomohyoid neck dissection,
were considered. Only nine (6.47%) patients had aberrant/skip metastasis with three
patients having isolated level III and six patients having level IV and level V involvement
without level III involvement. Level IB involvement was seen as the most common site
of nodal metastasis in cases of oral cavity squamous cell carcinoma. It was involved
in 73 (52.51%) cases followed by level IIA which was involved in 31 (22.30%) cases.
Level IA was involved in five patients (3.59%) only. Level IV involvement was seen
in four (2.87%) patients with two having patterned and two having skip metastasis.
Level V was involved in five (3.59%) patients, one having isolated VA involvement,
rest having skip metastasis to level VB. One (0.71%) patient had both level IV and
VB involvement simultaneously. Subsite of primary carcinoma in oral cavity affected
the level of nodal involvement significantly (Chi-square = 22.69, df = 12, p = 0.03, significant).
Pattern of Cervical Node Metastases “Subsite Wise”
Amongst the 245 patients, 2.04% had involvement of lymph nodes at level Ia, 29.79%
showed involvement of level Ib, 12.65% at level IIa, 1.22% at level IIb, 5.30% at
level III, 1.63% at level IV, and 2.04% at level V.
A. Buccal Mucosa
Out of 43 cases level IB was the most common site of involvement, being involved in
33 (76.74%) patients with cancer of buccal mucosa. All the patients had a systematic pattern
of lymph node metastasis. None of the patients with buccal mucosa cancer had level
IV or V involvement.
B. Tongue
Of 39 cases, none of the patients had involvement of level IIB. One patient had isolated
level III positive lymph node. Involvement of level IV was seen in three patients
with two patients having patterned metastasis with involvement at level IIA and III
simultaneously. One patient had aberrant metastasis with involvement at level IB and
skip metastasis to level IV. Aberrant metastasis to level V was seen in three patients.
One showing isolated metastasis to level VA and one showing simultaneous metastasis
to level IB and VB. The third case also had level IV involvement along with level
IB and level VB. One patient had poorly differentiated tumor with involvement at level
IV.
C. Lower Alveolus
Among 31 cases, Level IB was involved in 16 patients (51.61%) and eight patients (25.80%)
had level IIA involvement. Skip metastasis to level IV was seen in one patient (3.22%)
who also had pathological positive lymph node at level IB. Skip metastasis to level
V was seen in two (6.45%) patients with one patient having positive lymph node at
level IA and other at level IB simultaneously.
D. Floor of Mouth
Among 10 cases, level IB involvement was seen in six patients (60%) with carcinoma
at floor of mouth and level IIA was involved in one patient (10%). One (10%) of the
patients with carcinoma floor of mouth had isolated spread to level III.
E. Retromolar Trigone
Of 10 cases, level IB metastasis was seen in four patients (40%) with three patients
(30%) having metastasis in level IIA. One patient (10%) showed isolated involvement
of level III.
Pathological “T” Stage versus Pathological “N” Stage
Most of the patients had pathological T2 stage. Patients with pathological T4 stage
had maximum percentage of patients with pathological positive cervical node metastases
as well as maximum percentage of patients with multiple positive cervical node metastases.
When pathological “T” stage was correlated to pathological “N,” there was a trend
toward increase in “N” stage with increasing “T” stage and it was statistically significant
(Chi-square = 10.424 with four degrees of freedom; p = 0.03).
Pathological “T” Stage versus Cervical Node Metastases
A total of 33.3% of the patients with early stage (pathological T1/T2) had pathological
positive lymph node metastases. 50.7% of the patients with late stage (pathological
T3/T4) had pathological positive lymph node metastases. This difference was statistically
significant (Chi-square 5.721, df = 1, p = 0.01).
Tumor Characteristics
-
Depth of invasion: Depth of invasion affected rate of cervical node metastasis significantly (Chi-square
7.895; p = 0.02). Incidence of metastasis increased with increasing depth of invasion.
-
Pathological grade: A total of 71.4% of patients with poorly differentiated/undifferentiated pathological
positive nodes were identified. For grade I and II tumors, 37.3% of patients exhibited
pathological positive lymph nodes. The difference was statistically significant. Further,
it was observed that only one patient out of seven having grade I/II tumors had nodal
involvement below level III. The difference was not statistically significant (Chi-square = 6.834;
p = 0.04).
-
Pathological T stage: 33.3% of patients with early stage (pT1/pT2) disease had pathological positive lymph
node metastasis. Of patients staged as pT3/pT4, 50.7% had positive lymph nodes. This
difference was statistically significant (Chi-square = 5.721; df = 1; p = 0.01).
-
Morphological type of growth (ulceroproliferative versus ulceroinfiltrative): 29.4% of patients with ulceroproliferative growth and 36.6% of patients with ulceroinfiltrative
lesions had pathological positive lymph nodes. The difference was statistically significant
(Chi-square = 18.515; df = 2; p <0.001).
Multivariate Cox proportional hazards analysis showed that depth of invasion, pathologic
grading, pathological T stage ,and morphological type of growth are independent predictors
for regional cervical lymph node metastasis.
Discussion
The present study involved 245 patients with primary squamous cell carcinoma of oral
cavity. Patients who underwent neoadjuvant chemotherapy or had recurrent disease,
were excluded from the study. All patients underwent wide excision of primary lesion
with neck dissection and appropriate reconstruction. The resections and neck dissections
were done as per existing protocols in available literature.[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11] A total of 139 patients underwent MRND while 106 underwent SOHND, according to level
and number of lymph nodes involved clinically and intraoperatively. Of 245 patients,
99 had free flap reconstructions. Rest of the patients underwent reconstruction by
local/pedicled flaps/split skin grafting or primary closure.
In our study cohort of 245 patients, 96% of patients had a systematic spread with
initial metastasis to level I or systematic spread with initial metastasis to level
I or level IIa, followed by metastasis to level III and below. The most common site
of metastasis was level Ib (73 patients, 29.79%), followed by level IIa (31 patients,
12.65%). Level IIb metastasis was seen in three (1.22%) patients only and level III
metastasis was seen in 13 patients (5.30%). Patterned nodal metastasis was seen in
93.5% of (130 out of 139) patients who underwent Modified Radical Neck Dissection.
Nine patients (6.47%) exhibited aberrant or skip metastasis. Metastasis to level IV was seen in four patients,
all having clinically N+ and pathological N2b disease, with two having patterned and
two exhibiting skip metastasis. Level V metastasis was seen in five patients, one
showing isolated level Va involvement, rest having skip metastasis to level Vb (four
patients exhibiting infrahyoid involvement). One patient had metastasis to both level
IV and Vb simultaneously. So approximately 5% of the patients had cervical nodal involvement
below level III. It reaches 7.3% when clinically N positive necks are taken into consideration. None of the patients with clinically
N0 neck had nodal involvement below level III.
In a study by Shah et al,[12] among a total number of 501 patients undergoing 516 radical neck dissections, 357
were male (71%) and 144 female patients (29%), such that the sex ratio was 2.4: 1.
The age of patients ranged from 17 to 95 years, with mean and median age of 60 years.
Oral tongue was the common site of involvement (36%), followed by floor of the mouth
(33%). The incidence of clinically negative lymph nodes confirmed as positive on histopathologic
examination was 34%. Metastasis to level IV was seen in 3% of the patients with clinically
N0 neck. However, in clinically node positive patients, cervical nodal involvement
below level III was present in 15 to 16% of the patients. Metastasis to level V was
confirmed only in floor of mouth and gum primaries (6% in each). In our study, amongst
a cohort of 245 patients, male to female ratio was 3.8:1. This can be related to the
fact that females in India are comparatively less prone to tobacco usage. The age
ranged from 41 to 50 years, the median age being 45 years. Thus, the median age is
much younger compared with western population. An appropriate corollary can be drawn
to explain this fact, that exposure to tobacco begins at a much younger age in Indian
population. The most common site of involvement was buccal mucosa (46.12%) followed
by tongue (23.67%). This is due to prevalence of addiction to smokeless tobacco which
is kept in the buccogingival sulcus for hours together, thus justifying the name “Indian
Oral Cancer” for cancer of gingivobuccal sulcus. The incidence of clinically negative
nodes confirmed as pathologically positive on histopathologic examination was 23.40%.
This was lesser compared with the results from western studies. This can be related
to difference in histopathological techniques in American centers, where serial sectioning
of nodes is done more commonly. Level IV was involved in 2.87% of patients which was
comparable to study by Shah et al.[12] Level V was involved in 3.59% of our patients which was comparatively much lesser.
It was possibly due to higher prevalence of tongue cancer in the compared study, which
is known to metastasize more aggressively.
In a study by Woolgar,[13] all primary sites except tongue showed patterned metastasis, with jugulo-omohyoid
nodes and/or other deep cervical nodes at level IV positive in 20% of the patients.
The primary tumor was located on tongue or base of tongue in all except two of the
24 patients with level IV disease. These two exceptions had tumors in anterior floor
of mouth. Level IV was positive due to a regular progression (overflow) of metastatic
disease in 42% of 26 sides of neck with level IV disease. In six of these, positive
nodes were also found at level V. An erratic pattern of metastatic spread accounted
for positive level IV nodes in remaining 15 neck dissections (peppering, i.e., histological
involvement of nodes at multiple levels in the absence of macroscopic metastatic focus,
seven necks; skip metastasis in eight necks). The increased prevalence of peppering
in European and American studies could again be possibly due to employment of serial
sectioning technique. In another study by Woolgar,[14] levels I and II were the most common levels to be involved in metastasis. Level
IV was involved in 9% of patients and level III in 16% patients. Such high level of
metastasis to level IV can be ascribed to higher incidence of oral tongue cancer in
these studies. Metastasis in typical pattern, as mentioned in previous literature,
was seen in 67% of the patients, with skip metastasis in 10% of the patients. A single
micrometastasis was seen in 14% patients, with contralateral metastasis in one patient
and peppering in 2% of the patients.
In a study by Shukla et al,[15] 93% of patients had a patterned nodal involvement above level III, which was comparable
to our study. This shows similarity of trends in incidence amongst patients presenting
at centers of North India. In another study by Mishra and Sharma,[16] metastasis to level IV was seen in 9% of patients with clinically N+ disease and
was not detected in any case with clinically N0 disease. Hence, though the rate of
nodal involvement below level III is very low in clinically N0 patients, it rises
in patients with clinical N+ disease.
It is well known that the rate of metastasis for oral cancer is directly related to
tumor size and infiltration depth. In a meta-analysis by Huang et al,[17] a cut-off value of 4 mm for tumor infiltration depth was defined as a predictor
for cervical metastasis in metastasized cancers of oral cavity. In our study, rate
of metastasis increased with increasing depth of invasion. Similarly, Li et al[18] presented their results of 161 patients suffering from oral squamous cell carcinoma,
in relation to T stage, depth of invasion, and pathologic differentiation. Rate of
metastasis was found to increase with increasing pathologic grade of tumor, being
established as an independent predictor on multivariate analysis. In our study, rate
of metastasis increased with increasing pathologic grade of tumor. Poorly differentiated
tumors had a higher rate of nodal spread as compared with moderately or well differentiated
tumors. Haksever et al[19] have reported similar findings.
In our study, the rate of nodal metastasis increased with increase in T stage. 33.3%
of patients with early stage (pT1/pT2) disease had pathological positive lymph node
metastasis. Of patients staged as pT3/pT4, 50.7% had positive lymph nodes. This difference
was statistically significant (Chi-square = 5.721; df = 1; p = 0.01). These results are similar to those obtained by Sharma et al,[20] who showed that incidence of metastasis increased with increasing T stage. Also,
in the present study, the rate of metastasis was correlated with morphological type
of growth. 29.4% of patients with ulceroproliferative growth and 36.6% of patients
with ulceroinfiltrative lesions had pathological positive lymph nodes. The difference
was statistically significant (Chi-square = 18.515; df = 2; p <0.001). Khwaja et al,[21] Dissanayaka et al,[22] and Siriwardena et al[23] have confirmed similar findings.
Conclusion
In our study average age of patients suffering from squamous cell carcinoma of oral
cavity was a decade earlier than the age reported in literature, probably due to early
age exposure to tobacco. The most common site involved was buccal mucosa. Patterned
lymph node metastasis was seen in 93.5% cases. Skip metastasis was seen in 4.31% cases.
Level I b was the most common site of nodal involvement for all primary subsites of
oral cavity cancer. The incidence of positive nodes on histopathological analysis
was highest in cases of lower alveolus (63.15%), followed by tongue.
