Salivary duct carcinoma (SDC) was a highly aggressive subtype of salivary gland malignancy according to the WHO classification(13). Previous studies(14, 15) have shown that this subtype of patients presented with a high rate of distant metastasis in the clinic. In this study, we retrospectively collected the data of patients with SDC who underwent PORT and analyzed the prognostic value of the number of positive lymph nodes (PLN) and other risk factors for the DMFS. Moreover, we constructed the risk stratification which may be useful to identify the patients at high risk of poor survival.
Primary tumor resection was essential for the treatment of the head and neck cancer according to the National Comprehensive Cancer Network (NCCN) guidelines(6), while it could provide accurate diagnosis and treatment for the major salivary gland carcinomas. Regardless of the stage of SDC, postoperative radiotherapy (PORT) was strongly recommended because that adjuvant radiotherapy could significantly improve the local-regional control rate for the SDC (P =0.004)(7, 14). However, distant metastasis remained the major cause affecting the survival of the patients after PORT. But, few studies have investigated the prognostic predictors for distant failure with more than 50 cases due to its rarity. The largest study including the patients from the SEER database with 228 cases(16) showed that older age, large tumor size, and lymph node involvement were significantly associated with poor prognosis for the SDC. Nevertheless, factors involving distant metastasis were not reported in that study.
Regional lymph nodes played a significant role as a predictor of disease metastasis in most solid tumors. In breast cancer, the sentinel lymph node was the initial region for the tumor spreading to other organs. The tumor-positive nodes were a significant prognostic predictor for local regional and distant recurrence in breast cancer(17). The SDC shared a similar histological type with breast cancer, and nodal metastasis could predict the disease metastasis in the early time. Two retrospective studies have shown that advanced N stages (N2-3) were correlated with decreasing disease-free survival (DFS) and DMFS for the SDC(15, 18). In addition, the positive lymph ratio has also been reported as an indicator for poorer DMFS in the major salivary gland carcinoma(19, 20). And, T. J. Roberts(21) has illustrated the superior predicting value of positive lymph number (PLN) than lymph node ratio in head and neck cancer. In the major SGCs, the number of positive lymph nodes was associated with worse OS (P < 0.001) (22). However, different subgroups of the SGCs appeared with heterogeneous histological types and clinical phenotypes. Of note, the SDC was a highly aggressive subtype of malignancy compared with other SGCs, and the highest proportion of nodal involvement was observed among the major histological types in SGCs (54% vs. 24%, SDC vs. others)(23). In another study from the Netherlands, increasing PLN was associated with worse OS and DMFS in a non-linear way(24). For cases with 0 nodes being the reference, significantly worse DMFS was observed for those patients with 3-15 lymph nodes (P=0.007). Nevertheless, the study also included those patients with initial metastasis (stage M1) and those without radiotherapy. In our opinion, PORT was the standard approach for the SDC without distant metastasis. Thus, we analyzed the cohort after radical PORT and found that the risk of distant metastasis increased linearly by the PLN (Figure1.A). Moreover, we believe that several other pathological features (such as PNI, LVI) could predict the DMFS of the SDC, which were also not reported in that paper(24).
In our findings, the advanced T stage was an independent predictor for poor OS and DMFS in the multivariate analysis, which was consistent with the results from several other studies of salivary gland carcinomas following PORT(25, 26). Despite the factors of the TNM stage, PNI was correlated with decreasing DMFS in the univariate and multivariable analysis of the SDC(P=0.003). The PNI represented the metastatic ability through peripheral nerves and could predict the high risk of distant metastasis for patients with SGCs(27). The probability of PNI was higher in the SDC than that of other SGCs(28). However, few reports have included PNI into the analysis because of the very low incidence. Another study with 35 cases has illustrated the significant prognostic value of PNI for OS in the SDC(P=0.005)(8). While 48.4% of the patients were confirmed as PNI by pathological examination in our study, which indicated that the SDC appeared with high pathological aggressiveness. Furthermore, the patients with positive PNI were reported with significantly higher risk of distant recurrence(P<0.05).
For effective utility in the clinic, we combine the independent predictors to figure the risk score for every individual. Furthermore, the cohort was divided into two risk groups, whereas significant worse OS and DMFS were observed in the high-risk group and the administration of consolidation therapy may be beneficial. Several studies have investigated the efficacy of therapeutic approaches targeting AR and HER-2(29, 30), and the evaluation of AR and HER-2 was strongly recommended in NCCN guidelines for the SDC(6). However, nearly half of the patients in our study were reported with unknown status with AR and HER-2 due to the long timeline. In addition, identification of the patients at high risk could provide a reference for the trials of immune checkpoint inhibitor therapy in the SDC(13).
However, there were several limitations of this study. Firstly, the retrospective nature of the study may lead to the selection bias of the cohort. Secondly, the data of the patients were collected over a 15-year period, which could explain the lack of pathological information in AR and HER-2 (50% patients with unknown status). Lastly, the sample size was still small as the data was based on a single institution, and external validation is needed for the risk stratification model. Nevertheless, given the rarity of incidence in SDC, our study demonstrated the clinicopathologic factors that could predict the risk of distant metastasis in a moderate number of cases. In addition, some other prognostic information such as AR, HER-2, and other potential biomarkers were to be further explored.