The Laryngoscope

© 2018 The American Laryngological,

Rhinological and Otological Society, Inc.

Does Anatomic Subsite Influence Oral Cavity Cancer Mortality?

A SEER Database Analysis

Zachary Farhood, MD; Matthew Simpson, MPH; Gregory M. Ward, MD; Ronald J. Walker, MD;
Nosayaba Osazuwa-Peters, BDS, MPH, CHES

Objective: To determine if there are differences in mortality from oral cavity squamous cell carcinoma (OCSCC) based on
oral cavity (OC) subsites.
Methods: Using the Surveillance, Epidemiology, and End Results Program (SEER) 9 database, patients with sequence
number 0 or 1 squamous cell OCSCC were analyzed by OC subsite for 5-year cause-specific mortality (CSM) from OCSCC. Pro-
portional hazards regression determined the association between 5-year CSM and OC subsites while controlling for treatment
modality, stage, and demographic characteristics using hazard ratios. Significance was set at alpha = 0.05.
Results: 20,647 OC patients were included in the regression analysis. The most commonly diagnosed sites were floor of
mouth (34.4%) and oral tongue (34.3%). Floor of mouth, upper gum, and retromolar trigone were associated with lower CSM
compared to oral tongue. Not receiving surgery and receiving radiation were associated with increased CSM, and CSM
increased with cancer staging when distant or regional disease was compared to localized disease. Also, patients diagnosed at
60 years or older and black patients had increased CSM.
Conclusion: Among OCSCC patients, those with oral tongue cancer are more likely to experience CSM than patients with
floor of mouth, upper gum, and retromolar trigone cancer. It is important to understand these mortality related differences in
the management of OCSCC patients. Understanding subsite-specific mortality may benefit prognosis counseling of OCSCC
patients and elicit subsite-directed research as a means to improve outcomes.
Key Words: Oral cavity, cancer, subsite, mortality, squamous cell carcinoma.
Level of Evidence: NA
Laryngoscope, 129:1400–1406, 2019

INTRODUCTION Anatomically, the OC can be divided into the follow-

Cancer of the oral cavity (OC) has an incidence of
approximately 4.3 of 100 thousand cases1 and an
observed overall 5-year survival rate of 56%.2 It was pre-
viously estimated that in 2017 there would be 32,670 new
cases in the United States.3 Fortunately, there has been
a decrease in the incidence of OC cancer and an increase
in survival rate.4 In spite of this, OC remains one of the
10 most common cancers in the United States for men
when grouped with oropharyngeal cancer and the most
common cancer in the head and neck.3 In addition, the
incidence of tongue cancer has been increasing among
women.5 The most common histologic diagnosis is OC
squamous cell carcinoma (OCSCC), which represents
more than 90% of cases.6
From the Department of Otolaryngology–Head and Neck Surgery,
Saint Louis University, Missouri, U.S.A
Editor’s Note: This Manuscript was accepted for publication on July
16, 2018.
Presented at the American Academy of Otolaryngology–Head and
Neck Surgery Annual Meeting, September 10–13, 2017, Chicago, Illi-
nois, U.S.A.
The authors have no funding, financial relationships, or conflicts of
interest to disclose.
Send correspondence to Zachary Farhood, MD, Department of
Otolaryngology–Head & Neck Surgery, Saint Louis University School of
Medicine, 3635 Vista Ave. FDT-6, St. Louis, MO 63110. E-mail: zach.
farhood@health.slu.edu
DOI: 10.1002/lary.27490
ing subsites: oral tongue, floor of mouth, lower gum,
upper gum, buccal mucosa, hard palate, and retromolar
trigone. These subsites tend to have relatively unique
lymphatic drainage pathways and require varied surgical
and reconstructive approaches.7 Several factors of
OCSCC have been associated with poorer outcomes
(e.g., cervical metastases, extranodal extension, tumor
thickness, adverse histologic features).4,8 Subsites classi-
cally have varying propensities for locoregional spread.8
Furthermore, anatomic location predisposes certain
tumors to more advanced tumor (T) staging (e.g., bone
invasion).9 The goal of the current study was to deter-
mine if OC subsite predicted survival in SCC patients,
with the null hypothesis being that there is no difference
in survival of OCSCC based on subsite of the primary
tumor.
METHODS
Data Source
Data were abstracted from the Surveillance, Epide-
miology, and End Results (SEER) 9 database.10 SEER
9 includes patient data from registries in nine areas of
the United States collected from 1975 to 2013. The result-
ing sample encompasses about 9.4% of the total
U.S. population.11 Because SEER is publicly available
and the data have been deidentified, this study was

Laryngoscope 129: June 2019 Farhood et al.: Oral Cancer Subsite Survival
1400
exempt from consideration by the Saint Louis University
Institutional Review Board.
We included patients who were diagnosed with a
malignant first primary OCSCC without recurrence or
second primary. OCSCC was defined using the Interna-
tional Classification of Diseases for Oncology, Third Edi-
tion, site recodes C02.0–02.3, C03.0–03.1, C04.0–04.9,
C05.0, C06.0, and C06.2; and histologic types 8050 to
8076, 8078, 8083, 8084, and 8094. Patients were excluded
if the only documentation of their cancer was from a
death certificate or autopsy, if cause of death was
unknown, or if patients were listed as alive but had no
survival time. Data were abstracted using SEER*Stat
version 8.3.4 (Surveillance Research Program, National
Cancer Institute, Bethesda, MD).
Measures
The outcome of interest was death from first primary
OCSCC. SEER provides a cause-specific death classifica-
tion variable that considers patients’ number of tumors,
OC subsite of first cancer, and comorbidities to reduce
cause of death misclassification. Analyses used this vari-
able as the outcome.12 SEER also provides patients’ sur-
vival time in months from date of diagnosis to date of last
contact.13 For these analyses, follow-up time for patients
ranged from month of OCSCC diagnosis to month of
death or 5 years postdiagnosis.
The primary independent variable was OCSCC sub-
site (buccal mucosa, floor of mouth, hard palate, lower
gum, oral tongue, retromolar trigone, upper gum). Covari-
ates in multivariate models included age at diagnosis
(< 60 years old, 60 years old or older), sex (female, male),
race (black, white, other), county-level percentage of
smokers (quintile), county-level median household income
(quintile), year of diagnosis, surgery treatment (received
surgery, did not receive surgery), radiation treatment
(received radiation, did not receive radiation), stage
(localized, regional, distant), and grade (well differenti-
ated, moderately differentiated, poorly differentiated,
undifferentiated, unknown).
Statistical Analyses
Chi-squared and independent sample t tests, as
appropriate, determined preliminary associations for
independent variables with death from any cause.
Kaplan-Meier survival curves were plotted using cause-
specific mortality from OCSC in which those who died
from a cause other than OCSCC were censored. The sur-
vival curves were stratified by OC subsite. A log-rank test
determined if there was a significant overall survival dif-
ference among patients diagnosed with OCSCC at differ-
ent subsites, and Bonferroni adjustments compared each
pairing of OCSCC subsites to understand which subsites
had significantly different survival from each other.
A multivariable proportional hazards competing
risks model, which utilized cumulative incidence func-
tions as described by Fine and Gray,14 was fitted to deter-
mine the hazard of death from OCSCC using cancer
subsite as the primary independent variable. The
Laryngoscope 129: June 2019
aforementioned covariates were also included in the
model. Death from causes other than OCSCC was consid-
ered a competing event, and those who did not die from
any cause during the follow-up time were censored. To
control for stage using I through IV staging, a multivari-
ate Fine and Gray model subanalysis of OCSCC patients
diagnosed 2004 to 2013 was performed because SEER
only provided this staging information for OC cancer
patients beginning in 2004. This model included OCSCC
subsite and the same covariates as the main Fine and
Gray model; however, the stage variable was replaced
with the updated stage variable (I, II, III, IV, unknown).
These models yielded adjusted hazard ratios (aHRs) and
their 95% confidence intervals (CIs) to determine signifi-
cant predictors of death from OCSCC.
All tests were two-tailed, and alpha was set at 0.05.
All analyses were performed using SAS version 9.4 (SAS
Institute, Cary, NC).
RESULTS
Demographics
20,647 patients were included in survival analyses.
The most common OCSCC subsites were floor of mouth
(34.4%) and oral tongue (34.3%). The cohort was predomi-
nantly male (62.1%), white (85.2%), and diagnosed with
localized (44.4%) or regional (44.4%) OCSCC (Table I).
Kaplan-Meier Survival Curves
The median follow-up time for the cohort was
57 months. The 2-year survival probability for the entire
cohort was 66%, and the 5-year survival probability was
49%. When stratifying the cohort by OCSCC subsite, an
overall survival difference among subsites was found (log-
rank P < 0.001). Bonferroni adjustments indicated that
oral tongue cancer had significantly better survival than
all other subsites (P < 0.001), although survival was only
noticeably different between oral tongue and lower gum
cancer for about the first 2.5 years after diagno-
sis (Fig. 1).
Proportional Hazards Models
When comparing other subsites to oral tongue, floor
of mouth (aHR = 0.92, 95% CI 0.86, 0.98), retromolar tri-
gone (aHR = 0.82, 95% CI 0.75, 0.90), and upper gum
(aHR = 0.84, 95% CI 0.72, 0.97) had lower hazard of
death from OCSCC after controlling for covariates.
Patients ages 60 and older were 21% more likely to die
from OCSCC compared to patients younger than
60 (aHR = 1.21, 95% CI 1.15, 1.27), and were 9% less
likely to receive surgery compared to patients younger
than 60 (relative risk = 0.91, 95% CI 0.90, 0.93). Black
patients were 31% more likely to die from OCSCC com-
pared to white patients (aHR = 1.31, 95% CI 1.20, 1.43).
Patients who did not receive surgery were 113% more
likely to die from OCSCC compared to those who received
surgery (aHR = 2.13, 95% CI 2.00, 2.27), whereas
patients who did not receive radiation were 26% less
Farhood et al.: Oral Cancer Subsite Survival
1401
 TABLE I.
Demographic Characteristics, SEER 9 OCSCC Patients 1975–2013.
Alive (n = 10640) Dead (any cause) (n = 10,007) Total (n = 20647) P Value

Age at diagnosis, n (%) < 0.001

< 60 5023 (47.2%) 3314 (33.1%) 8337 (40.4%)
60+ 5617 (52.8%) 6693 (66.9%) 12310 (59.6%)
Sex, n (%) < 0.001
Female 4304 (40.5%) 3531 (35.3%) 7835 (38.0%)
Male 6336 (59.6%) 6476 (64.7%) 12812 (62.1%)
Race, n (%) < 0.001
Black 642 (6.0%) 1115 (11.1%) 1757 (8.5%)
White 9240 (86.8%) 8358 (83.5%) 17598 (85.2%)
Other 688 (6.5%) 524 (5.2%) 1212 (5.9%)
Unknown 70 (0.7%) 10 (0.1%) 80 (0.4%)
County-level current smoker percentage, n (%) < 0.001
Lowest quintile 2579 (24.2%) 2222 (22.2%) 4801 (23.3%)
2nd quintile 3813 (35.8%) 3512 (35.1%) 7325 (35.5%)
3rd quintile 1510 (14.2%) 1399 (14.0%) 2909 (14.1%)
4th quintile 999 (9.4%) 900 (9.0%) 1899 (9.2%)
5th quintile 1732 (16.3%) 1967 (19.7%) 3699 (17.9%)
Unknown 7 (0.1%) 7 (0.1%) 14 (0.1%)
County-level median household income, n (%) < 0.001
Lowest quintile 729 (6.9%) 664 (6.6%) 1393 (6.8%)
2nd quintile 933 (8.8%) 765 (7.6%) 1698 (8.2%)
3rd quintile 2444 (23.0%) 2668 (26.7%) 5112 (24.8%)
4th quintile 965 (9.1%) 946 (9.5%) 1911 (9.3%)
5th quintile 5565 (52.3%) 4960 (49.6%) 10,525 (51.0%)
Unknown 4 (0.0%) 4 (0.0%) 8 (0.0%)
Subsite, n (%)
Buccal mucosa 748 (7.0%) 770 (7.7%) 1518 (7.4%) < 0.001
Floor of mouth 3414 (32.1%) 3685 (36.8%) 7099 (34.4%)
Hard palate 236 (2.2%) 298 (3.0%) 534 (2.6%)
Lower gum 935 (8.8%) 859 (8.6%) 1794 (8.7%)
Oral tongue 4088 (38.4%) 2987 (29.9%) 7075 (34.3%)
Retromolar trigone 791 (7.4%) 1078 (10.8%) 1869 (9.1%)
Upper gum 428 (4.0%) 330 (3.3%) 758 (3.7%)
Surgery, n (%)
Received surgery 9567 (89.9%) 6847 (68.4%) 16414 (79.5%) < 0.001
Did not receive surgery 947 (8.9%) 2862 (28.6%) 3809 (18.5%)
Unknown 126 (1.2%) 298 (3.0%) 424 (2.1%)
Radiation, n (%)
Received radiation 3508 (33.0%) 5566 (55.6%) 9074 (44.0%) < 0.001
Did not receive radiation 6985 (65.7%) 4178 (41.8%) 11,163 (54.1%)
Unknown 147 (1.4%) 263 (2.6%) 410 (2.0%)
Stage at diagnosis, n (%)
Localized 6217 (58.4%) 2941 (29.4%) 9158 (44.4%) < 0.001
Regional 3700 (34.8%) 5465 (54.6%) 9165 (44.4%)
Distant 304 (2.9%) 1022 (10.2%) 1326 (6.4%)
Unstaged 419 (3.9%) 579 (5.8%) 998 (4.8%)
Grade, n (%)
Well differentiated 3018 (28.4%) 2102 (21.0%) 5120 (24.8%) < 0.001
Moderately differentiated 4518 (42.5%) 4380 (43.8%) 8988 (43.1%)
Poorly differentiated 1090 (10.2%) 1702 (17.0%) 2792 (13.5%)
Undifferentiated 31 (0.3%) 56 (0.6%) 87 (0.4%)
Unknown 1983 (18.6%) 1767 (17.7%) 3750 (18.2%)

OCSCC = oral cavity squamous cell carcinoma; SEER = Surveillance, Epidemiology, and End Results Program.

Laryngoscope 129: June 2019 Farhood et al.: Oral Cancer Subsite Survival
1402
likely to die from OCSCC compared to those who did First, the 1975 to 2013 data describes staging on a basis
receive radiation (aHR = 0.74, 95% CI 0.69, 0.79). as local, regional, and distant metastasis, whereas the
Patients with moderately differentiated (aHR = 1.19, 95% 2004 to 2013 set reflects the American Joint Committee
CI 1.11, 1.27), poorly differentiated (aHR = 1.53, 95% CI on Cancer I to IV staging system. In addition, the smaller
1.41, 1.67), and undifferentiated (aHR = 1.53, 95% CI sample size from the 2004 to 2013 data subanalysis com-
1.08, 2.17) OCSCC were more likely to die from OCSCC pared to the 1975 to 2013 may also be a confounding fac-
than patients with well differentiated OCSCC (Table II). tor that distinguishes these two sets (5,130 vs. 20,647
In the primary Fine and Gray model, patients with patients).
regional (aHR = 2.57, 95% CI 2.41, 2.75) and distant
(aHR = 4.17, 95% CI 3.78, 4.61) cancer were more likely
to die from OCSCC compared with patients with localized
TABLE II.
stage cancer. In the 2004 to 2013 subanalysis, 5,130 Fine and Gray Proportional Hazards Competing Risk Model, SEER
patients had data suitable for analysis, and 1,371 experi- 9 OCSCC Patients 1975–2013.
enced cause-specific mortality (CSM). Patients with stage
Adjusted 95% CI 95% CI
II (aHR = 2.29, 95% CI 1.82, 2.88), III (aHR = 4.30, 95% Hazard Ratio Lower Upper
CI 3.42, 5.41), and IV (aHR = 6.24, 95% CI 5.06, 7.69) (aHR) Bound Bound
OCSCC were more likely to die from OCSCC compared
Age at diagnosis (ref = < 60)
with patients with stage I cancer (Table III). Significant
60+ 1.21 1.15 1.27
differences by subsite were not seen with this data except
with hard palate, which showed a higher risk of death. Sex (ref = female)
Aggregate tumor, node, and metastasis (TNM) stages, Male 0.98 0.93 1.04
however, were analyzed across subsites. Several statisti- Race (ref = white)
cally significant differences were found, as listed in Black 1.31 1.20 1.43
Table IV. Other 1.12 1.00 1.24
County-level current smoker
percentage (ref = lowest
quintile)
DISCUSSION
Per current National Comprehensive Cancer Net- 2nd quintile 1.02 0.95 1.10

work guidelines, treatment of OCSCC is guided by stage; 3rd quintile 1.08 0.98 1.18
subsite-specific recommendations do not exist.15 The 4th quintile 1.00 0.89 1.13
results of this study indicate that CSM of SCC differs by 5th quintile 0.95 0.87 1.05
OC subsite. In the overall population sample, upper/lower County-level median household
gum, retromolar trigone, and floor of mouth SCCs showed income (ref = lowest quintile)
better survival rates compared to oral tongue. 2nd quintile 0.92 0.81 1.05
A subanalysis of the data, using the data from 2004 3rd quintile 0.97 0.87 1.09
to 2013, was performed and produced differing results. In 4th quintile 0.94 0.82 1.07
this subanalysis, hard palate demonstrated a greater risk 5th quintile 0.95 0.84 1.06
of death compared to oral tongue, but other subsites did Year of diagnosis 1.00 0.99 1.00
not have statistically significant differences. The discrep- Subsite (ref = oral tongue)
ancy in these results may have various explanations:
Buccal mucosa 1.02 0.92 1.14
Floor of mouth 0.92 0.86 0.98
Hard palate 1.08 0.92 1.26
Lower gum 0.93 0.84 1.03
Retromolar trigone 0.82 0.75 0.90
Upper gum 0.84 0.72 0.97
Surgery (ref = received surgery)
Did not receive surgery 2.13 2.00 2.27
Radiation (ref = received radiation)
Did not receive radiation 0.74 0.69 0.79
Stage (ref = localized)
Regional 2.57 2.41 2.75
Distant 4.17 3.78 4.61
Grade (ref = well differentiated)
Moderately differentiated 1.19 1.11 1.27
Poorly differentiated 1.53 1.41 1.67
Undifferentiated 1.53 1.08 2.17

Fig. 1. Kaplan-Meier survival curves for oral cavity squamous cell aHR = adjusted hazard ratio; CI = confidence interval; OCSCC = oral
carcinoma stratified by subsite. [Color figure can be viewed in the cavity squamous cell carcinoma; ref = reference; SEER = Surveillance, Epi-
online issue, which is available at www.laryngoscope.com.] demiology, and End Results Program.

Laryngoscope 129: June 2019 Farhood et al.: Oral Cancer Subsite Survival
1403
 TABLE III. African Americans are more likely to present late in the
Fine and Gray Model, SEER 9 OCSCC Patients 2004–2013. clinic, present with more advanced disease, are less likely
to have health insurance, and are less likely to receive
Adjusted 95% CI 95% CI
Hazard Lower Upper definitive treatment.11–21 These factors might explain the
Ratio (aHR) Bound Bound poorer survival rates that we observed.
Patients who did not have surgery and patients who
Age at diagnosis (ref = < 60)
received radiation were more likely to experience CSM
60+ 1.35 1.20 1.52 than their counterparts. Patients with unresectable dis-
Sex (ref = female) ease likely forwent surgery and may have been treated
Male 0.89 0.79 1.00 with palliative radiotherapy; this is not specified in SEER
Race (ref = white) based on how data is collected. Additionally, higher
Black 1.09 0.90 1.33 staged disease with poor prognostic factors on pathology
Other (American Indian/Alaska 1.01 0.82 1.23 thus requiring adjuvant radiotherapy may explain why
Native, Asian/Pacific Islander) those who received radiation did worse. For example, it
County-level smoking percentage has been previously shown that perineural invasion (PNI)
(ref = lowest quintile) portends worse disease-free intervals and locoregional
2nd quintile 1.04 0.88 1.23 control even in the pathologically negative neck.22 When
3rd quintile 1.02 0.81 1.28 patients have PNI, lymphovascular invasion, and extra-
4th quintile 1.12 0.85 1.48 capsular spread together on histopathologic examination,
5th quintile 0.91 0.69 1.20 the survival rate drops precipitously.23
County-level median household This is not the first study to examine OCSCC charac-
income (ref = lowest quintile) teristics based on subsite. For example, Liu
2nd quintile 0.85 0.65 1.12 et al. demonstrated varying incidences of OCSCC by sub-
3rd quintile 0.85 0.67 1.06 site based on Asian ethnicities in a California state regis-
4th quintile 0.87 0.67 1.11 try. The authors posited that this was due to cultural
5th quintile 0.82 0.63 1.08 differences in tobacco use methods among these ethnici-
Year of diagnosis 0.97 0.95 0.99
ties.24 A separate study examined outcomes of oral ton-
gue cancer compared to other OC subsites and found that
Subsite (ref = oral tongue)
oral tongue cancer patients had fewer postoperative com-
Buccal mucosa 1.15 0.92 1.43
plications and reduced hospital costs and length of stay.25
Floor of mouth 1.05 0.91 1.22
Patients were also more likely to be younger and less
Hard palate 1.35 1.02 1.80 likely to be African American. These results corroborate
Lower gum 0.95 0.77 1.17 those found in our study that some subsites are worse
Retromolar trigone 1.00 0.82 1.24 than the oral tongue, possibly due to biological or epide-
Upper gum 0.95 0.71 1.26 miological factors such as age and ethnicity.
Surgery (ref = received surgery) Effects of treatment outcome by subsite have also
Did not receive surgery 2.99 2.61 3.42 been studied. Specifically, Zelefesky et al. studied the use
Radiation (ref = received radiation) of postoperative radiotherapy in floor of mouth and oral
Did not receive radiation 1.14 0.99 1.31
tongue SCC in 51 patients.26 The authors observed sev-
eral differences in the natural course of these two dis-
Stage (ref = I)
eases. Oral tongue patients tended to fail local control,
II 2.29 1.82 2.88
whereas floor-of-mouth patients developed distant meta-
III 4.30 3.42 5.41
static recurrence. Although overall and disease-free sur-
IV 6.24 5.06 7.69 vival were the same for both groups, oral tongue patients
Grade (ref = well differentiated) experienced a significantly shorter time to death after
Moderately differentiated 1.43 1.21 1.69 recurrence. A follow-up study also demonstrated poor
Poorly differentiated 1.59 1.30 1.94 local control of patients with oral tongue SCC and posi-
Undifferentiated 2.75 1.17 6.44 tive margins compared to other subsites.27 These studies
Unknown 0.88 0.71 1.11 highlight potential differences in the way that these
tumors behave.
CI = confidence interval; OCSCC = oral cavity squamous cell carci- Oral tongue was used as a reference point to com-
noma; ref = reference; SEER = Surveillance, Epidemiology, and End Results
Program. pare other subsites in this study. Two reasons for this
include: 1) oral tongue cancer is quite common relative to
other subsites; and 2) it has been previously used as a ref-
Age greater than 60 years and African American erence in published literature. Bell et al. examined out-
race were associated with poorer outcomes in this study. comes of the oral tongue compared to all other subsites
Oral cavity cancer has previously been described as a dis- and found no difference in survival; this may have been
ease of the sixth and seventh decade, and prognosis usu- due to a small sample size (n = 215) or to collating all
ally worsens with age.16,17 Similarly, previous studies other subsites together (and thus losing statistical
have shown that African Americans have higher mortal- power). The authors also noted that the oral tongue
ity rates.18,19 There is evidence in the literature that cohort had more early stage disease compared with all

Laryngoscope 129: June 2019 Farhood et al.: Oral Cancer Subsite Survival
1404
 TABLE IV.
Significant Findings of Cause-Specific Mortality Based on Stage and Subsite.
Subsite Versus Reference Adjusted Hazard Ratio
Hard palate vs. oral tongue at stage = II 2.231
Hard palate vs. retromolar trigone at stage = II 2.602
Buccal mucosa vs. hard palate at stage = III 0.33
Floor of mouth vs. hard palate at stage = III 0.212
Floor of mouth vs. lower gum at stage = III 0.519
Floor of mouth vs. oral tongue at stage = III 0.502
Floor of mouth vs. retromolar trigone at stage = III 0.416
Floor of mouth vs. upper gum at stage = III 0.427
Hard palate vs. lower gum at stage = III 2.447
Hard palate vs. oral tongue at stage = III 2.368
Buccal mucosa vs. lower gum at stage = IV 1.475
Floor of mouth vs. lower gum at stage = IV 1.46
Floor of mouth vs. retromolar trigone at stage = IV 1.42
Floor of mouth vs. upper gum at stage = IV 1.457
CI = confidence interval.
others.28 Montero et al. proposed nomograms to stratify
patients with OCSCC. In their study of 1,617 patients,
floor of mouth and lower gum exhibited lower CSM proba-
bility compared to oral tongue, whereas upper gum had a
higher locoregional recurrence-free probability.29 This
study did not control for overall TNM staging but did fac-
tor in tumor size, invasion of nearby structures, and clini-
cal nodal status.
Strengths of this study include the use of a well vali-
dated database, with a large sample size allowing for a
robust analysis. These data span four decades, which
allows for trend analysis and following patient cohorts for
extended surveillance. Also, CSM standards in SEER are
quite well established. Limitations include inability to
manipulate data and variables (e.g., chemotherapy and
radiation doses) given the retrospective nature of SEER.
Also, critical risk factor information is missing in SEER.
For example, OC cancers are usually associated with
tobacco and alcohol use. It could well be that it is the dose
and intensity of tobacco use that has the greater influence
over how these cancers present and not the subsite. How-
ever, SEER does not address that question because it
does not report tobacco and alcohol information, except
for smoking use at a county level. Finally, CSM may be
underreported because deaths from chemotherapy and
radiation have nonzero mortalities.
CONCLUSION
Salient differences exist between the cause-specific
mortality of OCSCC when subsite is taken into account.
Despite the anatomic proximity and NCCN guidelines
grouping, confirmation is needed to determine if these
subsites should be treated identically. The data from this
study not only supports the findings of previous studies
but may also serve as a springboard for future investiga-
tions regarding anatomic barriers and nodal drainage
patterns based on subsite. Further investigations would
seek to better delineate prognostic factors, improve our
Laryngoscope 129: June 2019
95% CI Lower Bound 95% CI Upper Bound
1.028 4.839
1.03 6.573
0.15 0.723
0.098 0.457
0.276 0.976
0.33 0.764
0.243 0.712
0.191 0.954
1.054 5.683
1.176 4.771
1.001 2.174
1.112 1.917
1.076 1.875
1.007 2.109
understanding of the pathophysiology of these tumors,
and ultimately improve patient counseling and manage-
ment of disease treatment and expectations. This may
best be done in a large single or multi-institutional set-
ting to allow data absent from SEER to be reported.
BIBLIOGRAPHY
1. Oral Cavity Cancer: Professional Version - Page 3 of 12. American Head &
Neck Society. Available at: https://www.ahns.info/resources/oral-cavity-
cancer/3/#_Toc400976516. Accessed June 21, 2017.
2. Oral Cavity Cancer: Professional Version - Page 4 of 12. American Head &
Neck Society. Available at: https://www.ahns.info/resources/oral-cavity-
cancer/4/#_Toc400976516. Accessed June 21, 2017.
3. Siegel RL, Miller KD, Jemal A. Cancer Statistics, 2017. CA Cancer J Clin
2017;67:7–30.
4. Montero PH, Patel SG. Cancer of the oral cavity. Surg Oncol Clin N Am
2015;24:491–508.
5. Saba NF, Goodman M, Ward K, et al. Gender and ethnic disparities in inci-
dence and survival of squamous cell carcinoma of the oral tongue, base of
tongue, and tonsils: a surveillance, epidemiology and end results
program-based analysis. Oncology 2011;81:12–20.
6. Barnes L, Eveson J, Reichart P, Sidransky D. World Health Organization
Classification of Tumours: Pathology and Genetics of Head and Neck
Tumours. Vol 9. 3rd ed. Lyon: IARC Publications; 2005.
7. Mukherji SK, Armao D, Joshi VM. Cervical nodal metastases in squamous
cell carcinoma of the head and neck: what to expect. Head Neck 2001;23:
995–1005.
8. Shah JP, Gil Z. Current concepts in management of oral cancer—surgery.
Oral Oncol 2009;45:394–401.
9. Trotta BM, Pease CS, Rasamny JJ, Raghavan P, Mukherjee S. Oral cavity
and oropharyngeal squamous cell cancer: key imaging findings for staging
and treatment planning. Radiographics 2011;31:339–354.
10. Surveillance Epidemiology, and End Results (SEER) Program. Incidence -
SEER 18 Regs Research Data + Hurricane Katrina Impacted Louisiana
Cases, Nov 2015 Sub (1973-2013 varying). Available at: https://seer.
cancer.gov/data-software/documentation/seerstat/nov2015/. Accessed April
18, 2017.
11. SEER*Stat databases: November 2015 submission. 2015. Available at:
https://seer.cancer.gov/data/seerstat/nov2015/. Accessed April 18, 2017.
12. SEER cause-specific death classification. Available at: https://seer.cancer.
gov/causespecific/. Accessed April 18, 2017.
13. Calculation of survival time fields. Available at: https://seer.cancer.gov/
survivaltime/SurvivalTimeCalculation.pdf. Accessed April 20, 2017.
14. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a
competing risk. J Am Stat Assoc 1999;94:496–509.
15. NCCN Clinical Practice Guidelines in Oncology. National Comprehensive Can-
cer Network May 8, 2017; Version 2. 2017. Available at: https://www.nccn.
org/professionals/physician_gls/pdf/head-and-neck.pdf. Accessed September
24, 2017.
16. Silverman S, Jr. Demographics and occurrence of oral and pharyngeal can-
cers. The outcomes, the trends, the challenge. J Am Dent Assoc 2001;132
(suppl):7s–11s.
Farhood et al.: Oral Cancer Subsite Survival
1405
17. Osazuwa-Peters N, Massa ST, Christopher KM, Walker RJ, Varvares MA.
Race and sex disparities in long-term survival of oral and oropharyngeal
cancer in the United States. J Cancer Res Clin Oncol 2016;142:521–528.
18. Osazuwa-Peters N, Christopher KM, Hussaini AS, Behera A, Walker RJ,
Varvares MA. Predictors of stage at presentation and outcomes of head
and neck cancers in a university hospital setting. Head Neck 2016;38-
(suppl 1):E1826–E1832.
19. Goodwin WJ, Thomas GR, Parker DF, et al. Unequal burden of head and
neck cancer in the United States. Head Neck 2008;30:358–371.
20. Mahal BA, Inverso G, Aizer AA, Bruce Donoff R, Chuang SK. Impact of
African-American race on presentation, treatment, and survival of head
and neck cancer. Oral Oncol 2014;50:1177–1181.
21. Daraei P, Moore CE. Racial disparity among the head and neck cancer pop-
ulation. J Cancer Educ 2015;30:546–551.
22. Chinn SB, Spector ME, Bellile EL, et al. Impact of perineural invasion in
the pathologically N0 neck in oral cavity squamous cell carcinoma. Otolar-
yngol Head Neck Surg 2013;149:893–899.
23. Lin YT, Chien CY, Lu CT, et al. Triple-positive pathologic findings in oral
cavity cancer are related to a dismal prognosis. Laryngoscope 2015;125:
E300–E305.
Laryngoscope 129: June 2019
1406
24. Liu L, Kumar SK, Sedghizadeh PP, Jayakar AN, Shuler CF. Oral squamous
cell carcinoma incidence by subsite among diverse racial and ethnic popu-
lations in California. Oral Surg Oral Med Oral Pathol Oral Radiol Endod
2008;105:470–480.
25. Li R, Fakhry C, Koch WM, Gourin CG. The Effect of tumor subsite on
short-term outcomes and costs of care after oral cancer surgery. Laryngo-
scope 2013;123:1652–1659.
26. Zelefsky MJ, Harrison LB, Fass DE, et al. Postoperative radiotherapy for
oral cavity cancers: impact of anatomic subsite on treatment outcome.
Head Neck 1990;12:470–475.
27. Zelefsky MJ, Harrison LB, Fass DE, Armstrong JG, Shah JP, Strong EW.
Postoperative radiation therapy for squamous cell carcinomas of the oral
cavity and oropharynx: impact of therapy on patients with positive surgi-
cal margins. Int J Radiat Oncol Biol Phys 1993;25:17–21.
28. Bell RB, Kademani D, Homer L, Dierks EJ, Potter BE. Tongue cancer: Is
there a difference in survival compared with other subsites in the oral
cavity? J Oral Maxillofac Surg 2007;65:229–236.
29. Montero PH, Yu C, Palmer FL, et al. Nomograms for preoperative predic-
tion of prognosis in patients with oral cavity squamous cell carcinoma.
Cancer 2014;120:214–221.
Farhood et al.: Oral Cancer Subsite Survival