Female Versus Male Oral Cavity Cancer: is There a Difference?

Oral cavity squamous cell carcinoma (OSCC) is a leading cause of death in Taiwan, and most of the patients are male. Little is known about the differences in risk factors, cancer characteristics and treatment outcomes in female patients. The study aim is to investigate the clinicopathological and outcome differences between gender in patients affected by oral cancer in Taiwan. Methods: This is a retrospective study based on data obtained between 1995 and 2019. A total of 2,046 patients were recruited for analysis. Cancer characteristics, risk factors and treatment outcomes in patients with oral cancer between genders were collected. Results: Female patients represented the 6.7% of the entire cohort of study. Females were diagnosed at an older age and at an earlier local stage compared to male patients (p < 0.001). The female patients were less exposed to cigarette, alcohol, and betel-nut (BQ) (all p-values < 0.001). Tongue (55.1%) was the most frequent subsite involved in the female group, while buccal (38.4%) and tongue (35.3%) were more likely (p < 0.001) to be associated with male gender. In tongue cancer subgroup, female patients presented less frequently extra-nodal extension compared with male patients (p = 0.040). During the follow-up period, there was no signicant difference in recurrence and overall deaths between genders. Conclusion: In Taiwan, the male to female ratio in OSCC is 14:1. The tumor subsite distribution, environment exposure and stage distribution are different between females and male. There are no differences in term of survival between female and male OSCC patients.


Introduction
Oral cancer is the sixth most prevalent cancer worldwide and determines a leading cause of death in Taiwan, especially in the male population. [1] As estimated by the Ministry of Health and Welfare of Taiwan, oral cancer has remained the top ve causes of death for years. [2] The incidence is about 7,000 cases annually in Taiwan, which led to 3,027 deaths in 2018. Among the deaths, 2,779 were male, accounting for 90% of the mortality caused by oral cancer. [2] Oral cancer was observed to affect the male more than the female globally, and the ratio was highest in Taiwan (male to female ratio: 10.5). [3] In the male patients in Taiwan, the consumption of cigarettes, alcohol and betel quid (BQ) predisposes them to develop oral cavity cancer. It is estimated that cigarettes and alcohol attributed to about 80% of oral cancer. [1] In Taiwan, the signi cant sex-related differences in the frequency of tobacco and betel quid chewing may explain the higher incidence of oral cavity SCC in males (20.81 cases per 1 million persons) than in females (2.40 cases per 1 million persons) [4] Due to the substantially lower number of female patients with oral cancer, most of the studies focused on the male population. Little is known about the differences in risk factors attributing to oral cancer and treatment outcomes in women. For female patients, the environmental carcinogens' exposure is much less frequent in BQ endemic regions such as Taiwan and India. Some studies indicated that oral cancer might have different risk factors between genders. [5,6] Nevertheless, one study conducted in Italy and Switzerland showed that tobacco smoking and alcohol consumption represents the signi cant risk factors for both women and men; furthermore, Honorato et al. reported no signi cant differences in survival rates between genders. [7,8] Therefore, Considering the lack of general agreement in the literature, this study aimed to investigate the differences in characteristics, risk factors, and treatment outcomes between females and males patients affected by OSCC in Taiwan.

Study design
This retrospective study was based on data obtained from a single medical center (Chang Gung Memorial Hospital, Linkou branch, Taiwan). The data consisted of 2,046 participants diagnosed with oral cavity cancer (involving tongue, buccal, hard palate, gingiva, and oor of mouth) between Jul 1995 and Mar 2019. All the patients received pre-operative MRI, chest X-ray, liver sonography, and bone scan/ positron emission tomography (PET). All the patients had radical surgery as their rst treatment. Clinical data were collected from medical records. The staging system was based on American Joint Committee on Cancer TNM staging system, (AJCC, 2010 edition). [9] After radical surgeries, tumors with adverse pathologic factors such as advanced tumor stage (T3 or T4), poor tumor differentiation, lymph node extranodal spread (ENE), or tumor depth ≥ 10 mm underwent adjuant radiotherapy or chemoradiotherapy. [10,11] Patients were included if they had complete information and patients with distant metastasis or recurrence carcinoma were excluded in the study.

Data collection
Tobacco, alcohol, and BQ consumption history was divided into two groups: never use and ever/current use. Cancer sites included tongue, mouth oor, lip, buccal, alveolus (gum), hard palate, and retromolar trigone. The pathologic parameters included tumour size, tumour depth, perineural invasion, lymph node metastasis. Lymph node status was further classi ed as three categories, including negative, positive lymph node metastasis without ENE, and positive lymph node metastasis with ENE. Tumor size was classi ed as "small" if the patient had tumor in T1/T2 and "large" as tumor in T3/T4. Overall cancer stage in I and II was classi ed as early cancer and stage III/IV was classi ed as "advanced". All the patients were regularly followed up after the surgery. They returned to clinic every month in the rst year,

Statistical analysis
Statistical analysis was conducted to explore the dataset and compare the baseline characteristics between males and females. To assess the differences, chi-squared (χ 2 ) test was used for categorical characteristics and t-test was used for continuous variables. All p values < 0.05 were regarded as statistical signi cance. Overall and disease-free survival were assessed by using the Kaplan-Meier method and the differences were estimated by log-rank test. All analyses were conducted in STATA (StataCorp LLC, Texas, USA) version 15. Table 1 showed the distribution of baseline characteristics. A total of 2,046 observations met the inclusion criteria and were included for further analysis. The study population consisted of 1,910 males (93.4%) and 136 females (6.6%). The male-to-female ratio was 14:1 in the study. The average age of participants was 50.7 years old (Standard deviation: 11.2, range: 24.0 to 92.0.) Most of the participants had a history of smoking (84.7%), alcohol (72.1%), and BQ chewing (79.9%). Cancer sites were more frequent in tongue (35.3%) and buccal mucosa (38.4%). More than half of participants (58.1%) were diagnosed at an advanced stage. During the follow-up period, 649 (31.7%) participants were found to have recurrence.   The occurrence of cancer sites was also observed to vary between genders with a signi cant difference (p < 0.001). The most common cancer site in the male patients was the buccal (39.8%) followed by tongue (33.9%). Unlike the male patients, more than half of female patients were diagnosed to have tongue cancer (55.1%) which was followed by buccal cancer (19.1%). The diagnosed disease overall stage was similar between genders: 41.5% was early stage in males and 47.1% in females (p = 0.206).

Patient characteristics
During the follow-up period, the male patients were observed to have about one-third (32.1%) recurrence rate compared to one-quarter (26.5%) recurrence rate in female patients. It did not reach statistical signi cant difference evidence (p = 0.17) between genders.
We further analyzed in patients with tongue cancer since tongue was the most common affected anatomical site both in male (33.9%) and female patients (55.1%). The disease characteristics and treatment outcomes were compared between genders.  A total of 1979 with complete information on the time of diagnosis and death and 509 deaths among them were related with oral cancer. Four hundred and eighty males (7.7%) and 29 (7.5%) females expired during follow-up. On the other hand, 624 patients with recurrences had been observed. Half of the recurrence occurred within 8.79 years after being treated. Of the patients with recurrence, 589 were male and 35 were female. The frequency of recurrence among the female population was 9.4% and 10.5% in the male. Figure 1 showed Kaplan-Meier OS curves and Fig. 2 showed the DFS curves between genders There is weak evidence of a difference in OS between females and males during the follow-up period (log-rank test: p = 0.491). There was also no difference in the DFS between genders (log rank test: 0.280). In tongue cancer, the most frequent site of OSCC in female, the DFS was not different between female and male (p = 0.336). However, OS in female tongue cancer was better than male but not reached statistical signi cance (p = 5-year OS: 91.6% vs 87.1%, p = 0.074). Multivariate analysis adjusting age, tumor status and lymph node metastasis in DFS and OS showed no gender differences (DFS: p = 0.715 and OS: p = 0.170).

Discussion
The study included 2,046 patients in a single medical center and was wide in case number for analysis. However, it overlapped different staging systems (6th, 7th and 8th AJCC editions). We adopted the 7th edition throughout this study. In the study, more than 90% of the participants were male. They were more likely to be diagnosed with oral cancer at a younger age and with late tumor status compared with female patients. This age difference can be explained two folds: rst, in betel endemic area, patients develop OSCC at a younger age, compared to other populations not exposed at this risk factor [1] [12] Second, the female patients were rarely exposed to cigarette, alcohol, and BQ, unlike the male patients, therefore, other factors, as poor oral hygiene, inadequate dental status and chronic irritation, which represents independent risk factors for OSCC, irrespective of tobacco and alcohol consumption, may have played a role in the carcinogenesis [13]. Not smoking, not drinking patients has tend to presents at the extremes of age [14]. It is reasonable that with less carcinogen exposure, female population develop oral cancer later in their lives.
Tongue (55.1%) was the most frequent present anatomical site of oral cancer in female patients; this nding well correlates with the current literature; several studies had reported a higher involvement rate of this subsite, especially in non-smoker non-drinker women with HPV-negative OSCC [15]. On the other hand, we notice a slightly higher median age of our cohort, compared to other study [16]. We interpreted this discrepancy as to the results of the differences in the epidemiological basin assessed. Most of the studies are conducted in western countries, and low OSCC prevalence areas, as a matter of fact, Lin et al [17] which investigated the clinicopathological features in the Taiwanese population affected by OSCC have obtained similar results regarding the female patients' age. Furthermore, recent ndings, carried by Foy et al. in France have advocated a possible relationship between the development of OSCC in nonsmoking, non-drinking patients and viral infection, hypothesizing that changes in western country sexual behaviors may lead to an increased incidence of herpes virus in the oral cavity, especially HSV-2, similarly to what has already been described in HPV-positive oropharyngeal squamous cell carcinoma. Because viral genome integration has not been detected in non-smoking and non-drinking OSCC, a "hit and run" viral mechanism involving epigenome deregulation could therefore play a key role at early steps of oral carcinogenesis in this population of patients. [18] In our cohort, males OSCC were more likely to occur in buccal (38.4%) and tongue (35.3%) locations. The buccal mucosa is the most affected site in people with BQ chewing history due to frequent mucosa irritation. Since most of the male patients (79.9%) had BQ chewing history, buccal cancer was more likely to be found in the male participants. The tumor stage while diagnosed differs between genders. Most (74.3%) of the females had early tumor stage while diagnosed with oral cancer whereas the male patients only had about 60% early tumor stage. However, the lymph node status and the overall stage did not vary between genders. When we did a further analysis in patents with tongue cancer, we found that female patients were diagnosed at an older age and with early tumor size. Besides, in lymph node involvement, the females had fewer ENE if they had lymph node involvement compared to the males (N + ENE+, male: 24.9%; female: 13.3%). In non-smoking and non-drinking OSCC, the frequency of lymph node metastasis and ENE was less frequent than the smoking and drinking group. Although non-signi cant, the patients that smoked cigarettes and drank alcohol tend to have higher risks of lymph node ENE. [19] It could be related with the mechanisms of tumorigenesis. Those oral cancers related with smoking and BQ could harbour more genetic alterations from the environment carcinogens exposure and can be easier to metastasize or have ENE. The other possibility is the tongue cancers in male are more advanced in stage. The accumulated mutations with tumor stage make tumor cells aggressive in behaviour easier to metastasize.
During the follow-up period, there was no evidence of a difference in recurrence and deaths between genders. Since oral cancer was prevalent in the male population, most of the literature investigated the characteristics, risk factors, treatment, and outcomes of oral cancer in the male population. Few studies investigated this disease in females. We reviewed the studies about oral cancer between genders. Honorato et al. [8] compared the prognostic differences in oral cancer between genders in Brazil. The data showed similar consumption habits of alcohol and cigarette between genders with our population, whereas the anatomical sites of cancer differ from our experiences. It would be due to the lack of using BQ. The tongue was the most affected site in both males and females. [8] Due to the BQ consumption culture in Taiwan, buccal is the most affected site in oral cancer patients in Taiwan, unlike the results showed in Brazil. A study in European countries presented the risk factors of oral and pharyngeal cancer in women in Italy and Switzerland. [7] The study showed that women shared the same major risk factors such as habits of using cigarette and alcohol as men.
The rate of HPV infection in oral cavity cancer was about 13.4% while the infection rate was 36.7% in oropharyngeal cancer in the United States. [20] The HPV infection rate in OSCC is lower than the oropharyngeal cancer. In female OSCC patients, the frequency of HPV infection (9.6%) was much lower than the male (15.8%). In the large cohort study, female OSCC patients had better OS regardless of HPV status. It is the reason we speculated that the survival rates in our study would not be signi cantly in uenced as the HPV status was not determined in in our analysis.
In our study, there were 2,046 participants included, while only 136 patients were female. However, this is one of the largest study recruited the most number of female OSCC underwent surgery as primary treatment. HPV-related head and neck cancer were noted to be increasing in recent years especially in the younger population and those without exposure to tobacco and alcohol. Recently, with the increasing use of smoking in female population. This study provides an epidemiologic data in female OSCC in an BQ endemic area. Further cohort studies at different time periods is mandatory to compare the trends for female OSCCs. AI, CTL, CKT and CJK collected the data. YCL, HTC and CKY analysed the data. All the authors read and approved the nal manuscript.