1 BACKGROUND

Salivary duct carcinomas (SDCs) are highly malignant tumors. They have a low incidence, accounting for approximately 1%–3% of malignant salivary gland tumors, but are highly invasive [1, 2]. SDCs can occur in all salivary glands, the parotid gland being the most commonly involved, accounting for approximately 80% of all SDCs, followed by the submandibular gland, which accounts for approximately 15% [2, 3]. Early stage major salivary duct carcinomas (MSDCs) are prone to relapse, distant metastasis, peripheral nerve invasion, and extravasation; they accordingly have a poor prognosis and are frequently lethal [4, 5].

MSDCs more frequently develop lymph node (LN) metastases than do other salivary gland tumors [1, 6]. LN metastasis is generally considered a risk factor for long-term survival of patients with head and neck carcinomas [3, 7]. Accurate LN staging is therefore important for selection of postoperative therapy and evaluation of prognosis. The term “ratio of positive LNs” (PLNR) denotes the ratio of number of positive LNs (PLNN) to the total number of examined LNs (ELN); it has been proposed as an independent predictor of survival outcomes in patients with multiple cancers [8, 9]. Although the clinical value of PLNN and PLNR in head and neck carcinomas has been established [10-12], their prognostic roles in patients with MSDCs have not been separately studied.

In this study, we explored and analyzed the clinical characteristics and survival outcomes of patients with operable MSDCs. To the best of our knowledge, this is the first large-scale cohort analysis of the prognostic significance of PLNN and PLNR in patients with MSDC.

2 METHODS

2.1 Data collection

The clinicopathological characteristics and survival information of patients with MSDC from 2004 to 2016 were retrieved from the Surveillance, Epidemiology, and End Results (SEER) database. The following inclusion criteria were applied: (1) diagnosis of invasive ductal carcinoma; (2) location of primary tumor in a major salivary gland, including the parotid, submandibular, and major salivary, and not otherwise specified glands; (3) no evidence of distant metastasis; and (4) had undergone excision of the primary lesion. Patients with duplicated information and those for whom definite T, N, or M stages had not been supplied were excluded. The flowchart of patient selection is shown in Figure 1. To further explore the clinical value of LN parameters, we selected ELN ≥ 1 and PLNN ≥ 1 to be the main criteria for further screening the enrolled population. Because all studied data were extracted from a public SEER database, no ethics approval or informed consent was required.

3 RESULTS

3.1 Patient characteristics

Overall, 290 patients who had undergone surgery for MSDC between 2004 and 2016 were enrolled in the SEER database (Table 1). The median age at diagnosis was 68 years and 72.1% of the patients were men. More than half of the patients had LN metastases, patients with N1, N2, and N3 stage disease accounting for 12.1%, 44.5%, and 1.4% of all patients, respectively. All patients had undergone surgery for the primary lesion, 226 (77.9%) had received radiotherapy, and 75 (25.9%) had received chemotherapy. The patients' basic characteristics are summarized in Table 1. The median follow-up time was 48 months and the median OS 73 months for the whole study cohort. The 1-, 3-, 5-, and 8-year OS rates were 93.2%, 72.2%, 53.4%, and 44.3%, respectively (Figure 2a).

TABLE 1. Patient characteristics.

Characteristic	Total (n = 290)	%
Sex
Female	81	27.9
Male	209	72.1
Age at diagnosis (year)
≤ 68	150	51.7
> 68	140	48.3
Race
White	225	77.6
Black	33	11.4
Others	32	11.0
Insurance
Insured	243	83.8
No/Unknown	47	16.2
Marital status at diagnosis
Single/Widowed/Separated	63	21.7
Married	189	65.2
Unknown	38	13.1
Primary site
Parotid gland	233	80.3
Submandibular gland	39	13.4
Major salivary gland, NOS	18	6.2
Grade
I–II	36	12.4
III	122	42.1
IV	78	26.9
Unknown	54	18.6
Laterality
Left - origin of primary	121	41.7
Right - origin of primary	160	55.2
Not a paired site	9	3.1
T stage
T1	66	22.8
T2	63	21.7
T3	88	30.3
T4	73	25.2
N stage
N0	123	42.4
N1–3	167	57.6
AJCC stage
I	48	16.6
II	34	11.7
III	49	16.9
IVA/B	159	54.8
SEER stage
Localized	78	26.9
Regional	121	41.7
Distant	91	31.4
Radiotherapy
No	64	22.1
Yes	226	77.9
Chemotherapy
No	215	74.1
Yes	75	25.9

• Abbreviations: AJCC, American Joint Committee on Cancer; NOS, not otherwise specified; SEER, Surveillance, Epidemiology, and End Results.

3.4 Identification of the optimal cutoff point for PLNN and PLNR

Next, the clinical value of LN parameters in patients with PLNN ≥ 1 and ELN ≥ 1 was further explored. The optimal cutoff point for PLNN was 16 and the maximum χ² value was 4.0032. The 5-year OS rate of patients with PLNN ≤ 16 was 42.8% and that of patients with PLNN > 16 was 15.4% (p = 0.046). Moreover, the optimal cut-off point for PLNR was 0.48 and the maximum χ² value was 7.3602. The 5-year OS rate was 46.8% for patients with PLNR ≤ 0.48 and 26.3% for patients with PLNR > 0.48 (p = 0.007) (Figure 3).

4 DISCUSSION

MSDCs are one of the most aggressive salivary gland tumors, with high rates of local recurrence and metastasis [13]. Surgery with or without adjuvant therapy is the primary treatment for MSDC [14, 15]. More than half of MSDC patients reportedly have LN involvement at the time of initial diagnosis; the rate of occult LN metastasis can reach 24.4% [16]. Neck LN dissection aimed at completely removing occult metastatic LNs should therefore be performed during the initial surgery in patients suspected of having highly malignant MSDCs, whereas selective neck LN dissection combined with adjuvant radiotherapy and chemotherapy should be performed in patients with low-grade malignant tumors [17].

Notably, the status of LN metastasis is closely associated with the prognosis of patients with head and neck carcinomas [18-20]. In clinical practice, the accuracy of the PLNN is limited by the number of ELNs; an adequate number of ELNs may to some degree reduce the risk of residual LNs and LN micro-metastases. The higher the number of ELNs, the higher the possibility of pathological examination revealing LN metastasis; that is, the lower the possibility of missed diagnosis and the more accurate the pN stage. A variety of factors can lead to inadequate ELNs, such as a small surgical field, inadequate technology, inexperienced operators, small LNs, and inaccurate detection methods [21, 22]. Such factors limit the accuracy of prognoses based only on LN metastasis status. Too few ELNs can result in underestimation of the actual number of LN metastases, which in turn can directly affect the accuracy of pN stage [23]. In addition, incorrect assessment of the pN stage further affects the selection of postoperative adjuvant therapy and the accuracy of determination of a patient's prognosis.

PLNR, the ratio of PLNN to the total number of ELNs, reflects both of these prognostic factors and thus has great prognostic value. For the same number of LN metastases, the PLNR decreases in parallel with increasing number of ELNs. Conversely, for the same number of ELNs, the PLNR increases with increasing number of LN metastases, indicating that the tumor is prone to metastasize. Therefore, PLNR may provide a more accurate estimate of the pN stage than either PLNN of ELN alone in patients with head and neck carcinomas.

The main highlight of our study is that we first identified a significantly increased risk of death in patients with than without LN metastases and then focused on the subgroup of patients with LN metastases to further explore the potential value of PLNN and PLNR. The optimal cut-off values for PLNRs used in recently published studies evaluating the prognostic value of PLNR in patients with head and neck carcinoma have been inconsistent owing to differences in pathological type, sample size, inclusion and exclusion criteria, and statistical methods [24-26]. The cohort of the present study was limited to patients who had undergone surgery for MSDC. After adjustment for other variables (hazard ratio: 1.868, 95% confidence interval: 1.138–3.067, p = 0.014), we established an optimal cut-off value for PLNR of 0.48 and found that PLNR was closely associated with OS. These data suggest that PLNR is more reliable and accurate than other variables in guiding prognosis and helping to select high-risk patients for active treatment to prevent recurrence.

Our study had some limitations. We included only patients with complete clinicopathological data, which reduced the heterogeneity of samples in certain cities. In addition, the SEER database does not provide information about patient comorbidities, Eastern Cooperative Oncology Group scores, or tumor margins, lack of which may have affected the results of the present study.

5 CONCLUSION

We found that advanced T stage, submandibular gland location, and unpaired lesions were associated with LN involvement. PLNR can reflect both the effects of LN dissection and PLNN, and its prognostic value in patients with MSDC exceeds that of PLNN. The clinical application of PLNR requires further investigation.

AUTHOR CONTRIBUTIONS

Lixi Li: Conceptualization (equal); data curation (equal); formal analysis (equal); funding acquisition (equal); investigation (equal); methodology (equal); project administration (equal); resources (equal); supervision (equal); writing—review & editing (equal). Di Zhang: Conceptualization (equal); data curation (equal); formal analysis (equal); funding acquisition (equal); investigation (equal); methodology (equal); project administration (equal); resources (equal); software (equal); supervision (equal); validation (equal); visualization (equal); writing—original draft (equal).

CONFLICT OF INTEREST STATEMENT

The authors declare that they have no competing interests.

ETHICS STATEMENT

Not applicable.

CONSENT FOR PUBLICATION

Not applicable.