Discussion
Owing to the rarity and heterogeneity of salivary gland malignancies, there is a paucity of high–level, large prospective studies. As a result, little information regarding the clinicopathological factors associated with these tumours and their long-term prognosis is available.
The epidemiological findings of our study are relatively consistent with those in the existing literature. The majority of our patients were male and over the age of 50 years which is similar to previous studies(5-8){Lin, 2018 #170}. Cancers of the salivary glands are classified broadly as arising from either one of the major salivary or minor salivary glands. The most common site is the parotid gland, with an incidence ten times higher than the other glands (3). This is similar to our study, where 86% malignancies were seen in the parotid gland.
Five-year OS in the present study was 81.7% and 5-year DFS was 71%. This is in the context of over a third of patients presenting with stage III or IV cancers. In comparison, a review of the National Cancer Database (USA) reported a 5-year OS of 71%, whereas the Netherlands Cancer Registry noted OS to be 69% (9, 10). In the UK, Public Health England reported a 5-year survival rate of 78.7%(11).
When examining the clinicopathological variables, we found that age, pT classification, tumour grade, TNM stage, PNI, and LVI significantly affected DFS. This is interesting because retrospective studies of similar size have revealed the following tumour-specific variables as indicative of a poor prognosis: T classification, positive resection margin status, PNI, involved lymph nodes, distant metastases, and a fixed neck mass (6, 12, 13). Of these, PNI has been reported to be a prognostic factor for poor survival, a finding that is upheld by the present study. This may be linked to a greater risk of locoregional recurrence and distant metastases (14, 15). In our study, PNI was present in almost 40% of the cohort (a third of parotid malignancies and over half of cancers at the other sites). Few studies have reported a high incidence (28.7-48%) of PNI, albeit in smaller cohorts than ours (7, 13). Higher-grade histological subtypes were strongly associated with PNI and LVI, accounting for a poorer prognosis. In our study, contrary to others(7, 10, 11, 16), we did not find that pN classification, or the presence of ECS, affected survival. This could reflect the lower number of patients in these groups.
The resection margin in salivary gland cancers is often limited because in the presence of a functioning facial nerve and the absence of intraoperative signs of invasion the facial nerve is preserved. Sharp dissection along the perineurium allows good macroscopic resection though clearly often very close pathological margins. In the present study, DFS was comparable among patients with negative and close margins. Positive margins were associated with poorer survival, but this was not statistically significant, implying that more functional and less radical surgery may be acceptable, as long as it is followed by adjuvant therapy.
A large national cancer database review of 8,580 patients has shown that adjuvant radiotherapy improves survival regardless of TNM stage(18). To explore this further, we categorised the patients undergoing adjuvant therapy into early- and late-stage. Nearly half of our early-stage patients (48.6%) received adjuvant radiotherapy. This was because of the presence of classical poor prognostic factors such as PNI, LVI, and higher grade. Adjuvant radiation improves survival in early and advanced stage tumours. Patients receiving adjuvant chemoradiation had poorer survival, but this was likely indicative of disease characteristics meriting aggressive adjuvant therapy. The majority of these patients had high-grade tumours, a positive margin, LVI (71.4% each) and PNI (100%). Multivariate analysis showed worse DFS for the following factors: advanced stage (HR, 2.59), presence of PNI (HR, 1.1) and age over 50 years (HR, 1.31). None of the factors had a statistically significant effect, but this was expected owing to the smaller number of patients in the subgroups as well as good survival and a low number of events.
Our study has some limitations that are associated with retrospective studies, including some missing data. The heterogeneity of data resulted in some analyses that were underpowered for the tumour subtypes. Nevertheless, our study has significant merit. The follow-up period was sufficient to capture relevant survival and recurrence-related data, the study included a larger cohort size for salivary gland malignancies from a single UK centre, and all patients were treated by a uniform philosophy from a multidisciplinary team at a major teaching hospital trust.