INTRODUCTION
Clinical studies relating to the submandibular salivary gland are less common compared to that of the parotid gland. The latter receives more attention since parotid gland lesions are far more common, and surgery of the parotid gland is technically more challenging due to the intimate anatomical presence of the facial nerve (motor) that courses through the gland substance, thus generating more technical interest, sheer expertise and ongoing study. (1) Other nerves in proximity to the parotid gland are the greater auricular nerve (sensory) from the cervical plexus and the auriculo-temporal nerve (sensory) from the fifth cranial nerve.
In literature from the African continent, there are 3 reported studies on the spectrum of submandibular gland pathology. (1,2) Of note is that majority of clinical swellings (these present with a short history and are extremely painful & tender) in the submandibular gland are due to mechanical obstructions of the duct with secondary bacterial infection and an acute suppurative inflammatory response.
Worldwide literature concurs with the rarity of salivary gland neoplasms, accounting for less than 3% of all head and neck neoplasms. (1–5) Not surprisingly, the largest of the salivary glands, viz the parotid gland is the most frequently affected, accounting for 64-80% of cases; followed by the second largest salivary gland, viz the submandibular gland, accounting for 7-15% of cases. (3,5,6) Tumours of the submandibular salivary gland account for less than 1% of all head and neck neoplasms. (4) In general, males and females are equally affected by submandibular gland neoplasms, with the peak incidence in the fifth to sixth decades of life. (4)
Historical reports suggest that the majority (up to 80%) of neoplasms arising in the submandibular gland are malignant. (4–6) However, this figure is based on reports emanating from the western world. The prevalence rates from around the world seem to vary. In Asian and Black populations, the incidence of malignancy is as low as 10%. (2,6,7) The South African experience is that the rate of malignancy of submandibular gland tumours is much lower. (2)
METHODS
This report is a retrospective clinical audit of patients over 18 years of age, who had submandibular gland resections for neoplasia at two of the largest public hospitals in Johannesburg, South Africa i.e., the Chris Hani Baragwanath Academic Hospital and Charlotte Maxeke Johannesburg Academic Hospital. The period of review was over 7 years, from 1st January 2005 to 31st December 2011. The operating room surgical procedures register was used as a primary reference tool to identify eligible patients. The hospital's ward admissions register was used as a secondary reference to identify patients whose operating room records were lacking in certain specific details. All patients who were clearly identified from the operating room register were recorded as part of the initial sample population. This information was then used to interrogate and upload the relevant information from the database of the National Health Laboratory Services. Patients identified exclusively from the secondary reference were cross checked against their recorded laboratory findings and were included in the study population. Reports with poor or incomplete information were excluded from the study sample. Data collected included age, gender, race and histo-pathological diagnosis. Data was analysed using standard statistical methods. The study was approved by the Human Research Ethics Committee of the University of the Witwatersrand.
RESULTS
A total of 61 patients that had surgical excision of the submandibular gland were identified. Of these, 29 patients were excluded because they had non-neoplastic disease; three patients were excluded because they were younger than 18 years of age; and 3 were excluded because there were no histopathological results available. Thus, the study sample included 26 patients.
Age, Gender and Race
The age range of the study patients was from 22 to 65 years, with a mean (SD) age at surgery of 42.5 years (14.5). The majority (53.9%, n=14) of patients were aged 21-40 years. The mean age of Black patients (40.8 years) was not significantly different from that of White patients (51.75 years), p=0.29. Additionally, there was no significant difference between the mean age of patients with benign (40.3 years) versus malignant disease (46.6 years), p=0.29. (Figure 1).
Females represented 46% (n=12) and males 54% (n=14) of the sample.
Twenty-two patients of the study cohort (84.6%) were Black; the remainder (n=6) were White. No other racial denominations were represented. Of the Black patients, 12 (54.6%) were male, and 10 were female (45.4%). Amongst White patients, 2 patients were male and 2 were female.
Benign vs Malignant
In total, 17 (65.4%) patients had benign disease, while the remaining 9 patients (34.6%) had malignant disease. Of the 17 patients with benign tumours, 15 (88.2%) were Black and 2 (11.8%) White. Of the Black patients, 10 were male and 5 were female. White patients were represented equally across gender (1 male and 1 female). Thus, in total, 6 females (50% of total number of females) and 11 males (78.56% of total number of males) had benign disease.
In patients with malignant disease, 7 (77.8%) were Black, while 2 (22.2%) were White. Proportionally, Blacks with malignant disease account for 31.8% (n=7), and Whites 50% (n=2). Of the 7 Black patients, 2 were male and 5 were female. In White patients, again there was 1 male and 1 female. In total, 3 males (21.43% of total number of males) and 6 females (50% of total number of females) had malignant disease.
Lymphomas were the most common malignant tumours (55.6%). The most common epithelial or glandular tumour was adenoid cystic carcinoma (22.2%). There was one each of squamous cell carcinoma and carcinoma ex-pleomorphic adenoma (squamous cell carcinoma within a pleomorphic adenoma). No other histological types were found. (Figure 2)
The comparison of the ratio of benign to malignant tumours between Blacks and White patients, revealed no statistical significance using Fisher's Exact test (p=0.59). Of the total of 17 benign tumours, the majority (82.4%) were pleomorphic adenomas. The other 3 were lipomas. No other histological types were encountered.
Clinical Findings
Localised pain was the presenting feature in 7 patients (27%). Of these, only 2 had benign disease; the remainder had malignant disease. This difference was found to be statistically significant (p=0.03). Of note, 3 patients (12%) had tumours which were tender to palpation; 1 had benign disease and 2 had malignant disease. Tenderness to palpation was not found to be significantly associated with malignancy (p=0.27).
Adjacent peripheral nerve deficits were found in 3 patients i.e., predominantly a marginal mandibular nerve weakness as evidenced by drooping of the corner of the mouth due to malignant infiltration of the nerve. This finding was statistically significant (p=0.05). None of the patients had hypoglossal nerve palsies (indicated by weakness of the tongue), skin tethering, fixation to the mandible or ulceration of the skin. These clinical features are summarised in Tables 1 and 2.
DISCUSSION
Spiro (8) reported the largest case series of salivary gland tumours in 1986, at the Memorial Sloane-Kettering Cancer Centre in New York. He reviewed 2807 patients, of which 235 had neoplasia in the submandibular gland. In this series, 43% of these tumours were malignant. Further, most neoplasia occurred in the parotid gland (70%), and only 8% in the submandibular gland. (8) The mean age of patients presenting with any submandibular tumour was 42.5 years. This compares with other studies in Africa, which show a similar trend (1,2,9). A study amongst Asian patients also showed similar findings (7) Most tumours in the current study, occurred in the age group 22 – 40 years. This contrasts markedly with the international literature which quotes a peak incidence at 40 – 60 years. (8,10)
Our data suggests that submandibular tumours occur at a younger age in non-Caucasian races. There was no difference in the mean ages of males vs. females, nor of patients with benign vs. malignant disease. Our findings concur with that of another South African study. (2) Overall, males and females were almost equally represented. This is in keeping with some studies, which show no difference in the ratio of males to females. (1,2,4,10) However, other studies have shown a higher incidence in females (6,9), and as high as 3:1 in one study. (3) It should be noted that certain tumour types, particularly Warthin's tumour, have a marked sex predilection, although none of our patients presented with this entity.
Analysis of the data regarding tumour behaviour i.e., benign vs. malignant in different race groups, revealed that in black patients, males were more likely to have benign disease than females, with a ratio of 2:1. These findings are comparable with other studies (3,6).
Of the study sample of 26 patients, 17 patients (65.4%) had benign tumours, and 9 (34.5%) had malignant tumours. Overall, this is in striking contrast to the commonly quoted rates of malignancy in the general literature, of up to 80%. (1,4,8,10) When taking race into the equation, the rate of benign disease increases to 68.2% in Black patients. Other African studies have also found that the rate of malignancy is markedly lower in Black patients (2,9). Although there were very small numbers of White patients in the study cohort, the rate of malignancy in White patients was 50%, which is similar to the rate reported in the Western literature. In this study, however, the difference in malignancy rates between Blacks and Whites patients was not statistically significant.
The majority of benign tumours were pleomorphic adenomas. This is consistent with all African and Western studies (1–3,6,9). No other histological subtypes were found. This is probably due to the low number of patients in the study, but the rarity of Warthin's tumour in Black patients has been noted previously. (9)
Three patients had lipomas. Although primary lipoma of the submandibular gland has been described (14), it is exceedingly rare. It is more likely that these patients had a lipoma of the peri-glandular soft tissues, which is abundant in adipose tissue, and which was excised together with the gland.
The dominant malignant disease found mostly in Black females was lymphoma. Lymphomas arise from the abundant lymphatic tissues that are part of and in the surroundings of the parotid gland but have also been described as primarily involving the submandibular gland. (13) For this study, HIV status was not available. Since HIV is known to be associated with B-Cell lymphomas, it is possible that the high rate of lymphoma is related to the high rate of HIV-positive individuals in the South African population. In this regard, clinicians are now alerted to the possibility of a submandibular mass being caused by lymphoma, especially in an at-risk population.
The commonest epithelial-glandular malignancy encountered was adenoid cystic carcinoma. Several other studies have shown this trend. (1,3,4,11) However, Becerill-Ramírez et al reported adenocarcinoma and squamous cell carcinoma as the most common. They do, however, acknowledge that a limitation of that study was very low patient numbers (22 patients over 10 years. (6) Studies in Africa have shown a similar pattern. (2,9) Most patients did not present complaining of pain. However, when pain was present, it was invariably associated with malignant tumours. Most studies confirm that the presence of pain as a sinister feature, as pain from adenoid cystic carcinoma invariably implies nerve sheath infiltration. (1,8,10,11,16) Tenderness was present in only 3 patients. We did not find a significant difference between benign and malignant tumours, although most reports indicate that clinical tenderness tends to be associated with malignancy.
Weakness of the marginal mandibular nerve was present in 3 patients, all of whom were positive for malignant disease. Thus, peripheral nerve involvement is a strong indicator of infiltrating malignancy. This is in keeping with other studies. (1,4,11) Although none of our patients presented with it, weakness of the hypoglossal and lingual nerves has also been reported. (4,10,11) It is well known that pain, tenderness and peripheral nerve deficits are highly likely to be associated with malignancy. However, it is interesting to note that most tumours, even malignant ones, tended to present as painless, mobile masses with no sinister clinical features. Fixation to the overlying skin or to the ramus or body of the mandible, and ulceration of the skin have been reported as indicating infiltrating malignancy. (1) None of the patients in our series had these features.
LIMITATIONS
As this study constitutes a small cohort of patients, it limits its generalisability. Another shortcoming of this study is the lack of knowledge of the HIV status of the patients included in this report as HIV-infection is associated with lymphomas. Other studies of a similar nature show that HIV is high on the list of suspicion in young black patients presenting with submandibular gland or submandibular triangle masses. There are numerous difficulties in correcting these shortcomings most notably that the study population is highly mobile and therefore not readily available for follow up.
CONCLUSIONS
The data presented in this study show that black patients in South Africa have a lower rate of submandibular gland malignancy than reported in the Western literature. However, our findings are in keeping with other studies in Africa. Furthermore, neoplasms in Black patients were found to occur in the younger age group as compared to other international studies. This is particularly true of malignant tumours. Pleomorphic adenoma and adenoid cystic carcinoma were the commonest epithelial benign and malignant tumours, respectively. Localized persistent pain and peripheral nerve weakness or palsy were confirmed as being associated with infiltrating malignant tumours.
