Salivary gland carcinoma is relatively uncommon, leading to limitations in information regarding its epidemiological evaluations. This population-based study provides an overview of the incidence of malignancies in the major and minor salivary glands in Shanghai from the year 2003 to 2012.
In the present study, the crude incidence of SGC in Shanghai ranged annually from 11.37 to 16.07 per 1,000,000 person-years, averaging an ASR of 7.99 per 1,000,000 person-years. There was no significant difference in the ASRs of the first five and the next five years of the study period.
In contrast to other malignancies, the incidence rate of SGC has shown little variation among countries in the world. Sweden reported an SGC incidence rate of 1.32 cases per 100,000 inhabitant years during the time period between the years 1960 and 1989. The figure included both major and accessory SGCs, and showed no significant difference between the examined 10-year periods [9]. Luukkaa and his colleagues conducted a nation-wide evaluation of SGC in Finland with histological revision during the years 1991–1996 [4]. The incidence rates were in accordance with the data released by the Finland Cancer Registry, with ASRs ranging from 0.6 to 1.0/100,000 person-years during the years 1958–1998, displaying little change over the period [10]. Denmark has recently reported an ASR for SGC of 0.8/100,000/year, which is a little higher than the figure previously reported in the country [11]. The differences were probably owing to the improved standards of reporting and data-merging that improved completeness. In the Netherlands, the annual incidence of major SGC is approximately 0.7/100,000 [12]. Patrick et al. have reported annual incidence rates ranging from 0.83 to 1.38/100,000 population for both major and minor SGC during the years 1988 to 2007 in Nottingham of England [5]. There has been a modest increase in the incidence of malignant parotid neoplasms over the 20-year period in this region. According to the National Cancer Registries of England, the incidence rate for major SGC during 1990 to 2006 was between 0.7 and 0.8/100,000 population/annum. In the US, Pinkston and his group reported an incidence rate of 0.9 per 100,000 for major SGC, between 1968 and 1989 in the city of Jefferson [13], in agreement with data released by the Surveillance, Epidemiology, and End Results (SEER) program for the period spanning the years 1973 to 1992 [14]. Thus, the incidence rates among these countries have been comparable, at about 1.0/100,000 person-year.
The sex distribution of SGC in the present study suggests a higher incidence in men than that in women, with the M/F ratio of 1.10. A similar ratio has been reported in a Finland study [4], which is a little higher than that reported by studies from Sweden [9] and Denmark [11]. Guntinas-Lichius et al. described the epidemiology of head and neck cancer in Thuringia (Germany) and showed that the M/F ratio for newly diagnosed SGC was 1.64 during the period between 1996 and 2005 [2], and the incidence in men increased significantly from 0.50 to 1.57 per 100,000 persons/year over the 10-year period. Pinkston et al. reported an M/F ratio of 2.55 for SGC in the US, but the number of the cases was small [13]. Studies of the SEER database have reported that the M/F ratio for SGC was 1.15 with no change between the years 1973 and 1992 [14], while it increased to 1.29 during the period between 1992 and 2006 [1]. The male predominance in SGC incidence was once suggested to be associated with viral infection [15, 16], though this has not been confirmed by recent studies [1, 14]. In our series, the M/F ratio showed variations when analysed by anatomic subsites. SGCs of the submandibular gland showed male predominance, while those of the minor glands showed female predominance. The differences in sex distribution were consistent with those reported by Ostman [9] and Bjørndal [11]. This suggests that there are different risk factors for SGCs originating from different glands.
The median age for newly diagnosed SGC was 58 years old in our series. Women were diagnosed at an earlier age than were men, in agreement with that reported by Luukkaa [4] and Sun [14]. Other European studies have reported higher median ages at initial diagnosis, ranging from 60 to 63 years [11, 17,18,19], while a Brazilian study has reported an age peak for SGC incidence in the 70s [20]. In contrast, Pinkston [13] and Spiro [21] have reported a median age of 56 and 54, respectively. The aetiology of SGC at different ages-of-onset needs further investigation.
In the present study, the distribution of tumour sites was similar to that reported by other population-based studies, which revealed that over half of the SGC occurred in the parotid gland; the palate was the most common site for minor SGC; and the sublingual gland had the lowest incidence of salivary malignancies. The percentage of minor SGC in total SGCs has been reported to be 19% in Finland [4], 22.6% in Sweden [9], 24.2% in the Netherlands, and 27.9% in Denmark [11], while we reported a slightly higher incidence of 28.3%. Whether these differences are based on geographical variation, needs further evaluation.
There are some potential limitations of the study. Most importantly, the pathological diagnoses were not well-specified in 21.3% of cases. SGC presents histopathological challenges, since it displays great heterogeneity and the histological criteria are updating in the latest WHO classification. Another reason is that the SCR collects cancer reports from about 190 hospitals in Shanghai and pathological diagnosis with SGC subtyping is not mandatory for reporting. Another limitation in the study is that some SGC cases may be missed in this study. This may lead to an underestimation of the incidence. However, the reporting system to SRC is being improved in order to minimize the missing data.