J Oral Maxillofac Surg

61:751-758, 2003

Outcomes of Oral Squamous Cell

Carcinoma in Taiwan After Surgical
Therapy: Factors Affecting Survival
Wen-Liang Lo, DDS,* Shou-Yen Kao, DDS, DrMSc,†
Lin-Yang Chi, DDS, PhD,‡ Yong-Kie Wong, DDS, MS,§
and Richard Che-Shoa Chang, DDS, MS¶

Purpose: The study goal was to determine which clinical features correlated with 5-year survival in
patients surgically treated for oral squamous cell carcinoma (OSCC) in Taiwan.
Patients and Methods: The records of 378 OSCC patients surgically treated with or without
chemotherapy and radiotherapy were reviewed retrospectively. Their 5-year survival in relation to
age, gender, tumor site, lymph node involvement, presence of distant metastasis, staging, differen-
tiation, and risk factors, including betel quid (BQ) chewing, cigarette smoking, and alcohol con-
sumption, was analyzed.
Results: The majority of the patients were men (male-to-female ratio, 5.87:1) with the mean age of
57.1 ⫾11.7 years. Tumors occurred mainly at the buccal mucosa (BM) (100 of 378, 26.5%), gingiva
(105 of 378, 27.8%), and tongue (103 of 378, 27.2%). Neck nodal metastasis occurred frequently at
the floor of the mouth (in ⬎60% of cases), followed by the gingiva (45.7%), buccal mucosa (34%),
and tongue (20.4%), whereas early distant metastasis was rare (5.3%). There were 104 (27.5%) stage
1, 96 (25.4%) stage 2, 98 (25.9%) stage 3, and 80 (21.2%) stage 4 patients. OSCC at the BM and
gingiva was most (and at the tongue least) associated with risk factors of BQ use and smoking. The
5-year survival was 75%, 65.6%, 49%, and 30% for patients with stage I, II, III, and IV, respectively.
The size, nodal involvement, distant metastasis, staging, differentiation, and BQ use significantly
affected the survival (P ⬍ .05, Kaplan-Meier analysis). BQ use also correlated most significantly with
the younger age of occurrence of OSCC patients.
Conclusions: Our data suggest that early treatment is the key to increasing the survival of OSCC
patients. Periodic screening of high-risk populations for OSCC represents an urgent need in Taiwan.
© 2003 American Association of Oral and Maxillofacial Surgeons
J Oral Maxillofac Surg 61:751-758, 2003

Oral squamous cell carcinoma (OSCC) accounts for vival rate in many parts of the world, and severe
more than 95% of all malignant neoplasms in the oral functional and cosmetic defects that accompany treat-
cavity.1 OSCC represents a significant problem be- ment. The malignant transformation of OSCC involves
cause of its high incidence, unsatisfactory 5-year sur- a multihit process of aberrant genetic events follow-

*Fellow, Oral and Maxillofacial Surgery, Department of Dentistry,
Taipei Veterans General Hospital, School of Dentistry, National
Yang-Ming University, Taiwan, ROC.
†Associate Professor Staff, Oral and Maxillofacial Surgery, De-
partment of Dentistry, Taipei Veterans General Hospital, School of
Dentistry, National Yang-Ming University, Taiwan, ROC.
‡Associate Professor, School of Dentistry, National Yang-Ming
University, Taipei, Taiwan, ROC.
§Chief and Associate Professor, Oral and Maxillofacial Sur-
gery, Department of Dentistry, Taichung Veterans General Hos-
pital, Taiwan, ROC.
¶Professor and Chairman, Oral and Maxillofacial Surgery,
Department of Dentistry, Taipei Veterans General Hospital,
School of Dentistry, National Yang-Ming University, Taiwan,
ROC.
This study was supported by grants VGH-88-340C from Taipei
Veterans General Hospital and NSC-89-2314-B-075-113 from Na-
tional Science Council, Taiwan, ROC.
Address correspondence and reprint requests to Dr Chang:
Oral and Maxillofacial Surgery, Department of Dentistry, Tai-
pei Veterans General Hospital, No 201, Sec 2, Shih-Pai
Road, Taipei, Taiwan (112), ROC; e-mail: sykao@vghtpe.
gov.tw
© 2003 American Association of Oral and Maxillofacial Surgeons
0278-2391/03/6107-0004$30.00/0
doi:10.1016/S0278-2391(03)00149-6

751
752
ing the action of various carcinogens, which may
come from chronic use of factors such as tobacco,
alcohol, and betel quid (BQ).2-8 In addition, the inci-
dence of OSCC may vary according to race or ethnic-
ity, circumstances, and oral habits.9 In Western coun-
tries, the incidence of oral cancers is about 3% of all
malignancies,1 whereas in India, it ranks first (45%).10
Differences in the genetic and molecular factors asso-
ciated with OSCC have been proposed as the reason
for the variation in etiology and incidence of this
disease in patients of different ethnic backgrounds.11
For example, polymorphic differences in the gene
CYP1A1 have been correlated with varied incidences
of OSCC in different countries.12-20 Although cigarette
smoking and alcohol drinking are the risk factors that
cause most of the cases of OSCC in Western coun-
tries,21 BQ use and smoking are main risk factors in
southeast Asia, including Taiwan.4,5,22 An epidemio-
logic study of OSCC in Taiwan by Ko et al23 revealed
that 9.8% of males and 1.6% of females had the habit
of BQ chewing. Previous studies on the relationship
between frequency of mutation in OSCC patients and
these risk factors revealed that mutations of the ras
oncogene and p53 tumor suppressor gene occur sig-
nificantly more frequently in persons in countries
where massive amounts of BQ are consumed, includ-
ing Taiwan.24-29
Huang et al30 reported that 50% of patients with
new cases of OSCC treated in medical centers had
stage III or IV, which is associated with poor progno-
sis and survival. This may explain why, in 2000, OSCC
was the seventh leading cause of cancer mortality in
Taiwan. Although the mechanism of OSCC pathogen-
esis is still unexplored, the importance of early diag-
nosis and treatment for the survival of patients is well
recognized.
Because previous reports suggested that the ge-
netic and molecular differences in OSCCs might re-
flect the diverse etiology, risk factors, incidence, or
ethnic background of patients,11 it is therefore rea-
sonable to hypothesize that the generally agreed-on
treatment guideline or modality for OSCC given by
the American Joint Committee on Cancer might result
in different 5-year survival rates in different countries.
The purpose of this retrospective study was, first, to
determine the 5-year survival rate of our OSCC pa-
tients. Survival data were, second, statistically ana-
lyzed to determine which clinicopathologic features,
including disease characteristics and documented risk
factors (especially smoking and BQ chewing), corre-
lated with survival rate.
Patients and Methods
All patients with oral malignancies that were diag-
nosed at Taipei Veterans General Hospital (VGH)
ORAL SQUAMOUS CELL CARCINOMA SURVIVAL
from 1975 to 1996 were included in this study. Taipei
VGH serves the growing needs of both retired service-
men and their dependents in central and southern
Taiwan. The medical records of these investigated
subjects were reviewed. All of these subjects were
OSCC patients 1) who received surgical intervention
with or without adjuvant therapies (radiotherapy,
chemotherapy) and 2) from whom clinical data had
been systematically collected. The reviewed cases
were of new OSCC in patients first treated at Taipei
VGH. Lesions were classified according to the disease
site as described in the World Health Organization
International Classification of Diseases (ICD) for
Oncology. These sites included 1) the lips (ICD 140),
2) tongue (ICD 141), 3) buccal mucosa (ICD 142), 4)
alveolus and gingiva, upper and lower jaw (ICD 143),
5) floor of the mouth (ICD 144), and 6) palate (ICD
145). Surgical patients treated with radiotherapy or
chemotherapy only and patients who were followed
less than 5 years were excluded. Patients with tumors
of the salivary glands, sarcomas, or other malignancies
were excluded. Reliable information regarding de-
ceased patients during follow-up, including the date
and the cause of death, was obtained from the De-
partment of Health, Taiwan, ROC. The follow-up data
of the OSCC patients were recorded on a special chart
in the Oral and Maxillofacial Surgery Department,
Taipei VGH. The medical records were reviewed, and
details including gender, age, site, size of tumor,
lymph node (LN) involvement, distant metastasis,
staging, differentiation, risk factors of BQ chewing,
cigarette smoking, and alcohol consumption were
summarized on the special chart. A database contain-
ing 11 covariates (clinicopathologic features of OSCC
in our patients) was compiled for analysis (Table 1).
All data for statistical analysis were compiled using a
Windows-based PC and the SPSS 10.0 statistical pack-
age (SPSS Inc, Chicago, IL), including the SURVIVAL
supplementary module. Survival probabilities of those
who received surgical treatment with or without as-
sociated chemotherapy or radiotherapy were esti-
mated using the Kaplan-Meier method; survival
curves were plotted using the log-rank test. Multivar-
iate analysis with Cox’s proportional hazards linear
regression model and forward stepwise regression
was used to determine the independent prognostic
value of selected variables. In addition, the risk factors of
alcohol consumption, BQ use, and cigarette smoking
were further analyzed using random samples t test to
find the correlation with age at occurrence of OSCC.
Results
PATIENTS ENROLLED
The records of 575 patients with oral malignancies
were reviewed. A total of 197 patients were excluded:
LO ET AL 753

Table 1. CHARACTERISTICS OF 378 PATIENTS WITH OSCC WHO PARTICIPATED IN THE STUDY

Covariate Description (No. of Patients, Percentage of Total)

Gender Male (323, 85.4%); female (55, 14.6%)

Age 10 to 19 (0%); 20 to 29 (2, 0.5%); 30 to 39 (34, 9.0%); 40 to 49 (61, 16.1%); 50 to 59 (121, 32.0%);
60 to 69 (107, 28.3%); 70 to 79 (50, 13.2%); 80 to 89 (3, 0.8%); mean: 57.1 ⫾ 11.7 years old
(male: 56.9 ⫾ 12.0 years old; female: 58.1 ⫾ 9.7 years old)
Site Lip (11, 2.9%); buccal mucosa (100, 26.5%); gingiva (105, 27.8%); palate (36, 9.5%); tongue (103,
27.2%); floor of mouth (23, 6.1%)
Tumor size (T) T1 (ⱕ2 cm) (136, 36.0%); T2 (2 to 4 cm) (140, 37.0%); T3 (ⱖ4 cm) (93, 24.6%); T4 (tumor
invades adjacent structures) (9, 2.4%)
Lymph node (N) N0 (no palpable node) (257, 68.0%); N1 (ipsilateral palpable node) (71, 18.8%); N2 (contralateral/
bilateral palpable node) (35, 9.3%); N3 (palpable fixed node) (15, 4.0%)
Distant metastasis M0 (no distant metastasis) (358, 94.7%); M1 (clinical evidence of metastasis) (20, 5.3%)
(M)
Staging* I (104, 27.5%); II (96, 25.4%); III (98, 25.9%); IV (80, 21.2%)
Differentiation Well (214, 56.6%); moderate (145, 38.4%); poor (19, 5.0%)
Betel quid chewing Yes (183, 48.4%)/no (195, 51.6%)
Cigarette smoking Yes (276, 73.0%)/no (102, 27.0%)
Alcohol Yes (125, 33.1%)/no (253, 66.9%)
consumption
*Clinical staging (American Joint Committee on Cancer, 1998): I (T1 N0 M0); II (T2 N0 M0); III (T3 N0 M0/T1 N1 M0/T2 N1 M0/T3 N1 M0);
IV (T4 with any N or M/any T or N with M1).

33 had incomplete records, 113 had malignancies
other than OSCC (eg, mucoepidermoid carcinoma,
adenocystic carcinoma, malignant melanoma, sar-
coma, and major salivary gland malignancies), and
51 underwent no surgical intervention. A total of
378 cases were included, and case data from these
were collected and analyzed. Data on 11 clinicopath-
ologic features (regarded as covariates), including
gender, age, primary tumor site, size of tumor, LN
involve ment, distant metastasis, staging, differentia-
tion, BQ chewing, cigarette smoking, and alcohol
consumption, were compiled for analysis. When
the relationships between these covariates and sur-
vival rate (and other clinically associated outcomes)
were analyzed by log-rank comparison of Kaplan-
Meier survival curves, 6 of the 11 variables were
identified as significantly affecting survival (P ⬍ .05)
(Table 2).

Table 2. RELATIONSHIP BETWEEN 5-YEAR SURVIVAL AND CLINICAL FEATURES AS DETERMINED BY KAPLAN-
MEIER SURVIVAL ANALYSIS

Log Rank P
Variables* 5-Year Survival Rate (213 of 378, 56.3%) Test df Value

Gender Male (182 of 323, 56.3%), female (31 of 55, 56.4%) 1.01 1 .3158
Age Greater than mean (57.1 years old): 54.5%; less than mean 2.26 6 .8944
(57.1 years old): 45.5%; 20 to 29 (1, 50%); 30 to 39 (17,
50%); 40 to 49 (33, 54.1%); 50 to 59 (66, 54.5%); 60 to 69
(65, 60.7%); 70 to 79 (30, 60%); 80 to 89 (2, 66.7%)
Site Lip (7, 63.6%); buccal mucosa (59, 59%); gingiva (46, 43.8%); 4.84 5 .4362
palate (23, 63.9%); tongue (66, 64.1%); floor of mouth (12,
52.2%)
Tumor size (T)‡ T1 (90, 66.2%); T2 (81, 57.9%); T3 (40, 43.0%); T4 (2, 22.2%) 19.23 3 .0002†
Lymph node involvement (N)‡ N0 (169, 65.8%); N1 (27, 38.0%); N2 (14, 40%); N3 (3, 20%) 9.49 3 .0234†
Metastasis (M)‡ M0 (213, 59.5%); M1 (0, 0%) 8.40 1 .0038†
Stage‡ I (78, 75%); II (63, 65.6%); III (48, 49.0%); IV (24, 30%) 20.10 3 .0002†
Differentiation‡ Well (105, 49.1%); moderate (101, 69.7%); poor (7, 36.8%) 62.47 2 .0291†
Betel quid chewing‡ Yes (90, 49.2%)/no (123, 63.1%) 3.57 1 .0490†
Cigarette smoking Yes (152, 55.1%)/no (61, 59.8%) ⬍0.01 1 .9842
Alcohol drinking Yes (62, 49.6%)/no (151, 59.7%) ⬍0.01 1 .9746
*For definition of each variable, refer to the descriptions in Table 1.
†P ⬍ .05.
‡Significance affecting survival.
754 ORAL SQUAMOUS CELL CARCINOMA SURVIVAL

CLINICAL FEATURES OF CASES ASSOCIATED WITH

Table 4. DISTANT METASTASES, STAGING, AND
THE 5-YEAR SURVIVAL REGIONAL LYMPH NODE METASTASES IN PATIENTS
WITH ORAL SQUAMOUS CELL CARCINOMA
Demographic Data: Gender, Age, and Sites of
Tumor Lymph Node
For the 378 study patients, the male-to-female ratio Stage Metastasis
was 5.87:1 (323 men and 55 women). The mean age I ⫹ II III ⫹ IV Yes No
was 57.1 ⫾11.7 years (197 patients were ⬎57.1 years
old, 181 patients were ⬍57.1 years old). Male patient Patient no. 205 173 121 257
Metastasis no. 3 17 16 4
ages ranged from 27 to 88 years (mean age, 56.9 Percentage 1.5* 9.8* 13.2 1.5
⫾12.0 years), whereas female patient ages ranged
*␹2 test, P ⬍ .05.
from 32 to 79 years (mean age, 58.1 ⫾9.7 years). Of
the sites of OSCC involvement, 3 major anatomic
locations (gingiva, n ⫽ 105, or 27.8%; tongue, n ⫽ showed that 5-year survival significantly declined
103, or 27.2%; and buccal mucosa, n ⫽ 100, or 26.5%) with increased tumor size. The overall incidence of
accounted for the majority (81.5%) of all cases. Less regional LN metastasis was 32%. Significant difference
frequent locations included the palate (36, or 9.5%), in 5-year survival was noted between those with N0
floor of the mouth (23, or 6.1%), and lips (11, or and those with N-positive cervical nodal involvement
2.9%). The 5-year survival rate in men (56.3%) was (169 of 257, or 65.8%, and 44 of 121, or 36.4%,
very close to that in women (56.4%). There was no respectively) but not among those with N1, N2, and
significant statistical correlation between 5-year sur- N3 involvement.
vival rate and gender, age, or tumor site (Table 2). The incidence of regional LN metastasis was signif-
Nature of Tumors, Including Tumor Size (T), icantly correlated with the location of the primary
Cervical LN Involvement (N), Distant Metastases tumor; for example, palate and lip carcinoma was
(M), Staging, and Differentiation significantly correlated with low metastatic rates (P ⬍
Of the reviewed cases, 64% of tumors were greater .05) (Table 3). No patients with distant metastasis
than 2 cm on the first visit to the hospital; therefore, detected on the first visit survived 5 years; however, 4
only 36% of new cases involved a primary tumor of patients (4 of 20, or 20%) died in their third year after
less than 2 cm. About one third of the patients (121, control of their metastatic lesions failed. The rate of
or 32.0%) were found to have clinically detectable distant metastasis detection increased with tumor
and pathologically verified cervical LN metastasis, and stage (stages I and II: 3 of 205, or 1.5%; stages III and
most of these metastases (71 of 121) were located in IV: 17 of 173, or 9.8%; P ⬍ .05) (Table 4). The lung
the ipsilateral nodes. (18 of 20) was the most common site at which distant
Twenty patients (5.3%) were found to have distant metastasis was detected, followed by bone (5 of 20)
metastasis on the first visit. Distant metastases were and axillary skin (1 of 20). Of these 20 patients, 4 had
found in the lung, vertebrae, rib, and axillary skin, distant metastasis at more than 1 site (Table 5). The
which received focal radiation for tumor control. Tu- proportion of the 378 patients designated as having
mor size, cervical LN involvement, and metastasis stage I (27.5%), stage II (25.4%), stage III (25.9%), or
were found to correlate significantly with survival stage IV (21.2%) cancer was relatively evenly distrib-
(P ⬍ .05). The 5-year survival rates were 66.2% (90 of uted. In contrast to advanced stages (III and IV: 178,
136) for T1, 57.9% (81 of 140) for T2, 43% (40 of 93) or 47.1%), early stages of OSCC (I and II: 200, or
for T3, and 22.2% (2 of 9) for T4 patients, which also 52.9%) occurred in just over one half of cases. On the
basis of the patient’s history and the chief complaint
at the first visit, we estimate that the mean delay from
Table 3. PRIMARY TUMOR SITES AND REGIONAL
LYMPH NODE METASTASES
Table 5. LOCATION OF DISTANT METASTASES
Patient No. of Regional Lymph Node
Site No. Metastases (%) Distant Axillary
Location Metastases Lung Bone Skin
Lip 11 2 (18.2)*
Tongue 103 21 (20.4) Tongue 1 1
Buccal mucosa 100 34 (34) Gingiva 11 9 2 1
Alveolus/gingiva 105 48 (45.7) Buccal mucosa 5 5 2
Floor of mouth 23 8 (34.8) Palate 1 1
Palate 36 6 (16.7)* Floor of the mouth 1 1 1
Total 378 121 (32.0) Lip 1 1
Total 20 18 5 1
*P ⬍ .05.
LO ET AL 755

Table 6. RELATIVE RISK OF DEATH IN ORAL

oral habits, the most common site of OSCC was the
SQUAMOUS CELL CARCINOMA PATIENTS WITH tongue (29 of 63, or 46.0%). The difference showed
AND WITHOUT ORAL HABITS that BQ chewing has a significant influence on the
location of the OSCC site. The correlation between
Habits* Relative Risk the mean age at diagnosis of OSCC and these risk
No habit 1.00 factors was further tested using the random samples t
B⫹C⫹A 5.32 test. The mean age of OSCC patients chewing BQ,
B⫹C 2.00 smoking cigarettes, and drinking alcohol was 11.9
B⫹A 1.99 years younger than the age of those without any of
B 1.19
C⫹A 1.13 these habits (P ⫽ .004). The mean age of patients
C 1.14 chewing BQ and smoking was 7.8 years younger than
the mean age of those without these habits (P ⬍
*A indicates alcohol consumption; B, betel quid chewing; C,
cigarette smoking. .001). There also was a statistically significant trend
showing that patients who chewed BQ with or with-
out drinking alcohol were also younger than those with-
the first symptom to the first visit was about 8.6 out any risky habits (P ⫽ .065, P ⫽ .089) (Table 7).
months.
There were 214 cases (56.6%) of well-differentiated
Discussion
OSCC, 145 cases (38.4%) of moderately differentiated
OSCC, and 19 cases (5.0%) of poorly differentiated This retrospective study investigates the survival of
OSCC. The 5-year survival rates had a statistically 378 OSCC patients after surgical treatment. The rela-
significant correlation with staging of tumors: 75% (78 tionship of 11 demographic factors (including risk
of 104) in stage I, 65.6% (63 of 96) in stage II, 49.0% factors) to postoperative survival was tested. By the
(48 of 98) in stage III, and 30% (24 of 80) in stage IV. end of this study, 165 patients had died of recurring
Taken together, 5-year survival rates in those patients OSCC with associated fatal complications such as
with early stages (I and II) of OSCC (141 of 200, or massive bleeding or multiorgan failure due to late
70.5%) were significantly higher than in those with distant metastasis and cachexia.
advanced stages (III and IV) of OSCC (72 of 178, or
40.5%). A 5-year survival rate in patients with well- DEMOGRAPHIC DATA AND RISK FACTORS
differentiated OSCC (105 of 214, or 49.1%) and mod- Oral cancer occurs predominantly in males, with
erately differentiated OSCC (101 of 145, or 69.7%) the gender ratio ranging from 2 to 10:1.31,32 The
was found to be higher than in patients with poorly gender ratio in the present study was 5.87:1. The
differentiated OSCC (7 of 19, or 36.8%) (Table 2). marked male predominance among OSCC patients is a
reflection of the numbers of Taiwanese males with
RISK FACTORS OF ORAL HABITS, INCLUDING BQ high-risk factors for oral cancer, which include BQ
CHEWING, CIGARETTE SMOKING, AND ALCOHOL
CONSUMPTION
Of all cases involving oral habit risk factors, 183
Table 7. RISK FACTORS CORRELATED TO THE MEAN
(48.4%) were of BQ chewing, 276 (73%) were of AGE OF ORAL SQUAMOUS CELL CARCINOMA
smoking, and 125 (33.1%) were of alcohol use (Table PATIENTS AT DIAGNOSIS
1). When the relationship of BQ chewing, cigarette
smoking, and alcohol consumption to survival was Total Mean Age at
No. of Time of Relative
tested, the relative risk (RR ⫽ 5.32) of mortality due
Patients Diagnosis of Age† P
to OSCC in patients who simultaneously chewed BQ, Oral Habits* (%) SCC (yr) (yr) Value
smoked, and consumed alcohol was significantly
higher than that in patients without any of these No habits 63 (16.7) 58.1 — —
A 3 (0.8) 64.0 ⫹5.9 .484
habits (RR ⫽ 1) (Table 6). About one half of these B 18 (4.8) 53.1 ⫺5 .065
OSCC patients were BQ chewers (49.2%). The 5-year C 76 (20.1) 60.7 ⫹2.6 .221
survival rate of chewers (90 of 183, or 49.2%) was A⫹B 18 (4.8) 52.9 ⫺5.2 .089
significantly lower than that of nonchewers (123 of A⫹C 35 (9.3) 61.4 ⫹3.3 .218
195, or 63.1%) (P ⫽ .0490). No significant difference B⫹C 84 (22.2) 50.3 ⫺7.8 ⬍.001‡
A⫹B⫹C 81 (21.4) 46.2 ⫺11.9 .004‡
was observed in 5-year survival between smokers and Total 378 (100)
nonsmokers or between alcohol users and nonusers
(Table 2). Among BQ chewers, the predominant sites *A indicates alcohol consumption; B, betel quid chewing; C,
cigarette smoking.
of OSCCs were the buccal mucosa and gingiva (136 of †Relative to the “No habits” group.
183, or 74.3%); however, among those without risky ‡P ⬍ .05 (independent-samples t test).
756
chewing, cigarette smoking, and alcohol use.33 The
peak age of occurrence of OSCC in the present study
cohort (50 to 59 years old) is similar to that in a study
of 703 oral cancer patients reported by Chen et al33 in
southern Taiwan. The predominant sites of OSCC
occurrence in the present study were the gingiva,
buccal mucosa, and tongue. This was in contrast to
reports from Western countries, in which the tongue
and floor of the mouth were the main sites.32,34
NATURE OF THE TUMOR (TNM STAGING,
DIFFERENTIATION, AND METASTASIS)
The patients with lesions of greater than 2 cm
accounted for 75.6% of the 378 patients, which is
much greater than the proportion reported in other
studies.35 Regional LN involvement as a clinical symp-
tom of OSCC metastasis often represents the first
indicator of invasiveness.36 The percentage of OSCC
cases with initial LN involvement has been reported
to be higher in the tongue and floor of the mouth than
in other anatomic locations.37 From our data, we
found that patients with OSCC at sites in the buccal
mucosa, gingiva, and floor of the mouth had metasta-
ses in more than 30% of their regional lymph nodes.
According to the study of Probert et al,38 more than
50% of OSCC patients with distant metastasis were
tumor free postsurgically at the primary tumor site.
This suggests metastasis might have occurred subclin-
ically during or before OSCC therapy. The proportion
of patients with distant metastasis detected at the first
visit is atypical in this study (20, or 5.3%) compared
with other reports, in which it ranges from 12% to
19%.38,39 However, in the present study, the presence
of many patients who finally died because of metas-
tasis-related complications after combination treat-
ment suggests that subclinical disease was progress-
ing during the treatment period. Previous reports
suggest the lung to be the most common location of
distant metastasis, accounting for about 70% to 80% of
all metastases in OSCC patients. The occurrence of
distant metastasis in the liver and bones is also rela-
tively high and accounts for about 20% to 40% of all
metastasis in OSCC patients.38,39 In the present study,
the results show that the lungs were the main site of
distant metastasis. For screening distant metastasis,
plain films of chest radiographs, whole abdomen
sonography, and Tc-99 radionuclide whole body bone
scan are of value for ruling out early metastasis when
diagnosing primary OSCC and are also useful adjunc-
tive tools for screening during periodic follow-ups.
The present study showed that the 5-year survival
rate of OSCC patients was mainly associated with the
severity of clinical features. Of the 11 factors tested
for effect on survival in this study, 6 factors were
found to correlate significantly with survival; these
included tumor size, LN involvement, distant metas-
ORAL SQUAMOUS CELL CARCINOMA SURVIVAL
tasis, overall staging, differentiation, and BQ use. A
similar finding correlating the higher severity of tu-
mors with lower 5-year survival rates was noted in
previous reports.40
The incidence of OSCC ranges from high in India
and south Asia to low in the Western countries (3%).1
In Taiwan, OSCC is the seventh most common malig-
nancy. It is generally agreed that the epidemiology
and risk factors of this phenotypically similar, gene
mutation–related disease may vary widely. Thus, bias
should exist when comparing the overall 5-year sur-
vival of OSCC patients of different backgrounds or
from different countries. To obtain an objective com-
parison of a patient’s survival from the didactic retro-
spective studies, we might have to correct or stan-
dardize cases by matching him or her for meta-
analysis. For example, the categories of age, gender,
stage of tumors, and even risk factors might have to
be subdivided to see if groups of subjects in the same
or different studies differ regarding their survival.
However, variations in method of patient selection,
staging of disease, and standard treatment modality
used in different studies might also add intrinsic bias
to the analysis, although the difference between our
data showing a 56.3% overall 5-year survival rate and
the reported 5-year survival rates of about 50% in
other developed countries is significant.41,42
IMPORTANCE OF EARLY DETECTION OF OSCC
Our data showed that the percentage of advanced
stages (III and IV) was high (⬃47.1% of all new cases)
and that 75.6% of patients had primary lesions of
greater than 2 cm. These findings highlight the im-
portance of screening and early precancer/cancer de-
tection. Evidence obtained from patient records
showed that there were obvious delays (mean delay
from the time of symptom onset to presentation at the
medical center was ⬃8.6 months) in seeking treat-
ment. This period of delay is consistent with the
results of a study by Wang et al43 but is much longer
than delays in previous reports by Flamant et al44 and
Bruun.35 It has been reported by Guggenheimer et
al45 that the main causes of delay in seeking treatment
for oral cancer may be due to 1) less attention being
paid to the small, painless intraoral lesions, 2) cancer
phobia and unwillingness to face facts, and 3) care-
lessness of the local dentist or physician causing a
delay in referral. Another possible cause of delay is
when the patient first seeks traditional Chinese herbal
therapy rather than formal medical management. In-
terestingly, 2 previous studies regarding this issue
from Brazil and Denmark showed no association be-
tween patient delays and disease staging.46,47 Further
epidemiologic investigations are necessary to substan-
tiate the possible association between treatment delay
and disease staging.
LO ET AL
BQ USE, SMOKING, AND ALCOHOL USE AS MAJOR
RISK FACTORS IN TAIWAN
The combined effects of BQ chewing, smoking,
and alcohol drinking as risks to survival were appar-
ent in the present study. The genesis of OSCC is a
complex process involving multiple genetic and epi-
genetic alterations, which can be affected by risk
factors. However, 16.7% of our OSCC patients had no
risky oral habits. Alcohol consumption and cigarette
smoking are the most recognized major risk factors
for OSCC in developed Western countries. In high-
risk regions of the world, however, other causative
factors appear to play important roles. This was found
to be true in south Asia, southeast Asia, and New
Guinea, where BQ chewing is common and highly
correlated with the occurrence of OSCC.48-50 Our
results reveal that the high incidence of buccal and
gingival OSCC was associated with BQ chewing and
smoking.51 The vast majority of patients had at least 1
risky oral habit, which included BQ chewing, alcohol
consumption, and cigarette smoking. Experimental
data have shown that components of BQ chewing
have cancer-promoting effects.52 Patients who smoke
cigarettes, drink alcohol, and chew BQ have a 5.32-
fold increased likelihood of death compared with
those without any oral habits (Table 6). Our data
revealed that BQ use was the most significant risk
factor for early OSCC occurrence (Table 7). There-
fore, an approach for increasing the survival and pre-
venting OSCC would be to reduce the prevalence of
these habits through public-awareness programs.
In summary, it is generally accepted that cancer
mortality may be reduced if lesions are detected,
diagnosed, and treated at an earlier stage. Although an
investigation of this association was not included in
the present study, we did find that 5-year survival
rates were better in patients with the early stages of
OSCC than in those with the advanced stages. Thus,
we may conclude that the periodic screening of high-
risk populations for OSCC represents an urgent need
in Taiwan. The early treatment and the withdrawal
from risky oral habits, especially BQ chewing and
smoking, should increase patient survival.
References
1. Muir C, Weiland L: Upper aerodigestive tract cancers. Cancer
75:147, 1995 (suppl)
2. Ponder BA: Genetic predisposition to cancer. Br J Cancer
64:203, 1991
3. Ranadive KJ, Shivapurkar NM, Gothoskar SV: Betel quid chew-
ing and oral cancer: Experimental studies on hamster. Int J
Cancer 24:835, 1979
4. Pindborg JJ, Reibel J, Roed-Petersen B, Mehta FS: Tobacco-
induced in oral leukoplakic epithelium. Cancer 45:2330, 1980
5. Rao AR: Modifying influences of betel quid ingredients on
B(a)P-induced carcinogenesis in the buccal pouch of hamster.
Int J Cancer 33:581, 1984
757
6. Marshall JR, Graham S, Haughey BP, et al: Smoking, alcohol,
dentition and diet in the epidemiology of oral cancer. Oral
Oncol 28:9, 1992
7. Mashberg A, Boffetta P, Winkelman R, et al: Tobacco smoking,
alcohol drinking, and cancer of the oral cavity and oropharynx
among U.S. veterans. Cancer 72:1369, 1993
8. Degama M, Stern SJ, Martin MV, et al: Metabolic activation and
carcinogen-DNA adduct detection in human larynx. Cancer Res
54:4915, 1994
9. Ehlinger P, Fossion E, Vrielinck L: Carcinoma of the oral cavity
in patients over 75 years of age. Int J Oral Maxillofac Surg
22:218, 1993
10. Babu KG: Oral cancers in India. Semin Oncol 28:169, 2001
11. Paterson IC, Eveson JW, Prime SS: Molecular changes in oral
cancer may reflect aetiology and ethnic origin. Oral Oncol Eur
J Cancer 32B:150, 1996
12. Park JY, Muscat JE, Ren Q, et al: CYP1A1 and GSTM1 polymor-
phisms and oral cancer risk. Cancer Epidemiol Biomarker Prev
6:791, 1997
13. Hung HC, Chuang J, Chien YC, et al: Genetic polymorphisms of
CYP2E1, GSTM1, and GSTT1; environmental factors and risk of
oral cancer. Cancer Epidemiol Biomarker Prev 6:901, 1997
14. Matthias C, Bockmuhl U, Jahnke V, et al: The glutathione
S-transferase GSTP1 polymorphism: Effects on susceptibility to
oral/pharyngeal and laryngeal carcinomas. Pharmacogenetics
8:1, 1998
15. Matthias C, Bockmuhl U, Jahnke V, et al: Polymorphism in
cytochrome P450 CYP2D6, CYP1A1, CYP2E1 and glutathione
S-transferase, GSTM1, GSTM3, GSTT1 and susceptibility to to-
bacco-related cancers: Studies in upper aerodigestive tract can-
cers. Pharmacogenetics 8:91, 1998
16. Tanimoto K, Hayashi SI, Yoshiga K, et al: Polymorphisms of the
CYP1A1 and GSTM1 gene involved in oral squamous cell car-
cinoma in association with a cigarette dose. Oral Oncol 35:191,
1996
17. Morita S, Yano M, Tsujinaka T, et al: Genetic polymorphisms of
drug-metabolizing enzymes susceptibility to head-and-neck
squamous-cell carcinoma. Int J Cancer 80:685, 1999
18. Sato M, Sato T, Izumo T, et al: Genetic polymorphism of
drug-metabolizing enzymes and susceptibility to oral cancer.
Carcinogenesis 20:1927, 1999
19. Katoh T, Kaneko S, Kohshi K, et al: Genetic polymorphisms of
tobacco- and alcohol-related metabolizing enzymes and oral
cavity cancer. Int J Cancer 83:606, 1999
20. Sato M, Sato T, Izumo T, et al: Genetically high susceptibility to
oral squamous cell carcinoma in terms of combined genotyp-
ing of CYP1A1 and GSTM1 genes. Oral Oncol 36:267, 2000
21. Elwood JM, Pearson JC, Skippen DH, et al: Alcohol, smoking,
social and occupational factors in the etiology of cancer of the
oral cavity, pharynx and larynx. Int J Cancer 34:603, 1984
22. Chen YK, Huang HC, Lin LM, et al: Primary oral squamous cell
carcinoma: An analysis of 703 cases in southern Taiwan. Oral
Oncol 35:173, 1999
23. Ko YC, Chiang IA, Chang SJ, et al: Prevalence of betel quid
chewing habit in Taiwan and related sociodemographic fac-
tors. J Oral Pathol Med 21:261, 1992
24. Field JK: Oncogenes and tumour-suppressor genes in squa-
mous cell carcinoma of the head and neck. Eur J Cancer
28B:67, 1992
25. Kuo MYP, Jeng JH, Chiang CP, et al: Mutations of ki-ras onco-
gene codon 12 in betel quid chewing-related human oral squa-
mous cell carcinoma in Taiwan. J Oral Pathol Med 23:70, 1994
26. Donehower LA, Bradley A: The tumor suppressor p53. Biochim
Biophys Acta 1155:181, 1993
27. Chang F, Syrjanen S, Tervahauta, et al: Tumourigenesis associ-
ated with the p53 tumour suppressor gene. Br J Cancer 68:653,
1993
28. Greenblatt MS, Bennett WP, Hollstein, et al: Mutations in the
p53 tumor suppressor gene: Clues to cancer etiology and
molecular pathogenesis. Cancer Res 54:4855, 1994
29. Slootweg PJ: Suppressor protein p53 and its occurrence in oral
tumours. Curr Top Pathol 90:179, 1996
30. Huang CS, Kuo SS, Kuo IS: Oral cancers in Taiwan: Literature
review. Chin Oral Maxillofac Surg 5:13, 1994
758
31. Sankaranarayanan R: Health care auxiliaries in the detection
and prevention of oral cancer. Oral Oncol Eur J Cancer 33B:
149, 1997
32. Pinholt EM, Rindum J, Pindborg JJ: Oral cancers: A retrospec-
tive study of 100 Danish cases. Br J Oral Maxillofac Surg 35:77,
1997
33. Chen YK, Huang HC, Lin LM, et al: Primary oral squamous cell
carcinoma: An analysis of 703 cases in southern Taiwan. Oral
Oncol 35:173, 1999
34. Oliver AJ, Helfrick JF, Gard D: Primary oral squamous cell
carcinoma. A review of 92 cases. J Oral Maxillofac Surg 54:949,
1996
35. Bruun JP: Time lapse by diagnosis of oral cancer. Oral Surg Oral
Med Oral Pathol 42:139, 1976
36. Merino OR, Lindberg RD, Fletcher GH: An analysis of distant
metastasis from squamous cell carcinoma of the upper respi-
ratory and digestive tracts. Cancer 40:145, 1977
37. Lindberg R: Distribution of cervical lymph node metastases
from squamous cell carcinoma of the upper respiratory and
digestive tracts. Cancer 29:14469, 1972
38. Probert JC, Thompon RW, Bagshaw MA: Patterns of spread of
distant metastases in head and neck cancer. Cancer 33:127,
1974
39. Dennington ML, Caster DR: Distant metastasis in head and neck
epidermoid carcinoma. Laryngoscope 90:196, 1980
40. Holm LE, Lundquist PG, Ruden BI, et al: Combined preopera-
tive radiotherapy and surgery in the treatment of carcinoma of
anterior two-thirds of the tongue. Laryngoscope 93:792, 1983
41. Boyle P, Macfarlane GJ, Maisonneuve P, et al: Epidemiology of
mouth cancer in 1989: A review. J R Soc Med 80:724, 1990
ORAL SQUAMOUS CELL CARCINOMA SURVIVAL
42. Shah JP, Lydiatt W: Treatment of cancer of the head and neck.
CA Cancer J Clin 45:352, 1995
43. Wang YI, Chang SY, Hwu BC: The factors of diagnosis and
treatment delay in the late stage of cancers in pharynx, larynx
and oral cavity. Chin Med J 49:86, 1992
44. Flamant R, Hayem M, Lazar P, et al: Cancer of tongue: A study
of 904 cases. Cancer 17:377, 1964
45. Guggenheimer J, Verbin RS, Johnson JT, et al: Factors delaying
the diagnosis of oral and oropharyngeal cancers. Cancer 64:
932, 1989
46. Jiwaksju Jom Franco EL, Torloni H, et al: Lateness of diagnosis
of oral and oropharyngeal cancer: Factors related to the tumor,
the patient and health professionals. Oral Oncol Eur J Cancer
30B:167, 1994
47. Wildt J, Bundgaard T, Betzen SM: Delay in the diagnosis of oral
squamous cell carcinoma. Clin Otolaryngol 20:21, 1995
48. Ng KH, Siar CH, Ramanathan K, et al: Squamous cell carcinoma
of the oral mucosa in Malaysia: Any change? SE Asian J Trop
Med Pub Health 16:602, 1985
49. Chattopadhyay A: Epidemiologic study of oral cancer in eastern
India. Ind J Dermatol 4:59, 1989
50. Thomas SJ, MacLennan R: Slaked lime and betel nut cancer in
Papua New Guinea. Lancet 340:577, 1992
51. Chen CH: An epidemiological study of oral squamous cell
carcinoma in southern Taiwan. J Formosan Dent Assoc 10:268,
1987
52. Lin LM, Chen YK, Lai DR, et al: Cancer-promoting effect of
Taiwan betel-quid in hamster buccal pouch carcinogenesis.
Oral Dis 3:232, 1997