Introduction
Malignant tumours of the salivary gland are rare cancers that represent less than 1% of all newly diagnosed malignancies and less than 5% of all head and neck malignancies. Reference Bray, Ferlay, Soerjomataram, Siegel, Torre and Jemal1 Salivary duct carcinoma (SDC) is an uncommon, aggressive subtype of the salivary gland malignancy that arises from the ductal epithelial cells. It represents around 1.8% of all major salivary gland tumours and about 10% of all salivary gland malignancies. SDCs may develop de novo or following malignant transformation of a chronic pleomorphic adenoma, also called SDC ex-pleomorphic adenoma.
SDCs have been associated with poor prognosis with a high proportion showing locoregional recurrence or distant metastasis. Various risk factors for poor outcomes have been noted in prior studies. In a clinicopathological report by Jaehne et al., expression of HER2/neu and p53 was statistically linked to early local disease recurrence, distant metastasis and poor survival rates. Reference Jaehne, Röser, Jaekel, Schepers, Albert and Loening2 A retrospective review by Gilbert et al. showed that an advanced stage disease at presentation was common and correlated with poor outcome. Perineural invasion (PNI), lymphovascular invasion (LVI), extracapsular extension, higher N-stage (≥N2) and facial nerve involvement at diagnosis were other negative prognostic indicators in their study. Reference Gilbert, Sharma and Schmitt3
Optimal management of SDCs, including treatment modalities and scheduling, is poorly defined due to the lack of randomised data. Treatment given is extrapolated from studies on other SGCs and majorly follows the general principles of managing head and neck malignancies. Surgery is the primary modality of treatment where apart from resection of the primary tumour with adequate margins, selective neck dissection (levels 1–3) is recommended in high-grade histologies like SDC. Reference Sood, McGurk and Vaz4 The United Kingdom National Multidisciplinary Guidelines advise consideration of postoperative external beam radiation therapy (EBRT) in high-grade histology and high-risk histological features, but themselves disclaim the lack of good level evidence behind these guidelines. Reference Sood, McGurk and Vaz4,Reference Pons, Alves, Clément and Conessa5 Addition of concurrent chemotherapy has not shown to have any survival benefit in high-risk SGCs. Reference Amini, Waxweiler and Brower6 Chemotherapy is generally reserved for metastatic forms of the disease. Expression of various receptors in SDC like HER2/neu (20–55%), androgen receptor (AR) (43–92%) and epidermal growth factor receptor (EGFR) (60–92%) Reference Fan, Melhem, Hosal, Grandis and Barnes7–Reference Williams, Roberts, Kies, Mao, Weber and El-Naggar9 has led to opening up the avenue for molecular-targeted therapies to be considered as novel treatment strategies.
Literature from India regarding SDCs comprises of very few case reports and small case series. Reference Kumar, Kini and Bhargava10 There is a great difference between the head and neck malignancies prevalent in the west versus those in India, in terms of epidemiological profile, genetic factors, acquired risk factors, stage of presentation, access to care etc., all of which dictate the treatment outcomes. Reference Prabhash, Babu and Chaturvedi11,Reference Shah, Sharma and D’Cruz12 Consequently, management strategies adopted from Western data are often not appropriate in the Indian scenario. There is incumbent need for further studies, both retrospective and prospective, to increase our understanding of these rare but aggressive tumours. The purpose of this single-institution study was to analyse the high-risk factors, patterns of care, and subsequent outcome of the patients treated in our institute, particularly focusing on the benefit of radiotherapy (RT) in management of SDCs.
Materials and Methods
Patients with a histological diagnosis of salivary ductal carcinoma in our institution in a 9-year period between February 2010 and February 2019, who were candidates for adjuvant RT following surgical resection, were eligible for inclusion into this study. Patients with metastatic disease at diagnosis were excluded from the study. Data were collected retrospectively from electronic medical records, imaging reports, histopathological reports, surgical notes, radiation therapy treatment charts, and the treatment planning system. Radiation Therapy Oncology Group (RTOG) grading criteria was used to quantify radiation-associated toxicities. Reference Cox, Stetz and Pajak13 The patterns of failure and time to progression were recorded and analysed. A telephonic follow-up was done to complement the outcome data.
Kaplan–Meier methods were used to calculate overall survival (OS) and progression-free survival (PFS) rates. In our study, OS was defined as the interval between the date of first treatment to death from any cause. PFS was defined as the interval between the date of first treatment to recurrence or disease progression. Univariate analysis of prognostic variables was performed using a log-rank test, and Cox proportional hazards regression models were used to assess the predictors of PFS and OS on multivariate analysis. Statistical analyses were performed using IBM SPSS® software (Version 25, Armonk, New York, USA).
Results
Demographics and presentation
Sixteen patients with histologically proven SDC who underwent surgery and were candidates for adjuvant RT were identified. Of the 16 patients, 15 were men. The mean age at diagnosis was 56.75 years (range 40–79 years). The most common presenting symptom was a painless mass below the ear (10 patients) with examination findings typical of a parotid mass. Four patients developed a painful parotid swelling and two patients presented with swelling in the submandibular region. The mass was on new onset in ten patients, while six patients had a history of recurrent, long-standing swelling with a sudden increase in its size, suggesting the possibility of carcinoma ex-pleomorphic adenoma (CaExPA). Six patients had palpable neck nodes at presentation, while the rest did not have any palpable cervical lymphadenopathy.
Surgery and histopathology
The surgical procedure for the primary lesion that these patients underwent was as follows: three patients underwent superficial parotidectomy only, four underwent either upfront or completion total conservative parotidectomy, while the remaining seven underwent more radical parotidectomy and the two patients with a submandibular lesion underwent submandibular tumour radical excision. The neck was not addressed for the six patients who were clinico-radiologically node-negative, that is, cN0. The remaining ten patients with clinically suspicious neck nodes underwent appropriate neck dissection—four patients underwent ipsilateral modified radical neck dissection and six patients underwent selective neck dissection.
Pathological T-stage was reported as pT1 for one patient, pT2 for five patients, pT3 for five patients, pT4a and pT4b for four and one patients, respectively. Among the nine patients who underwent neck dissection, seven patients had cervical nodal metastasis on histopathology where one patient had a single ipsilateral node positive while multiple ipsilateral nodes (range 2–22) were positive in six patients. Extranodal extension was present in five out of the nine patients (55%) who underwent neck dissection. In addition to metastatic cervical nodes, five patients also had peri-parotid nodes with tumour metastasis. Among the other high-risk features on histopathology, LVI and PNI were both present in seven patients, while two and one patient(s) had isolated LVI or PNI, respectively. Four patients’ biopsies had been tested for HER2/neu immunohistochemistry—two were negative and two were positive.
Adjuvant therapy
All 16 patients in this study were candidates to receive adjuvant RT after surgery; however, one patient defaulted RT despite being advised the same, and one patient was found to have liver metastases prior to RT and was hence advised systemic therapy instead. The remaining 14 patients (87.5%) underwent adjuvant RT as advised. RT was delivered to the post-operative site of the primary and the ipsilateral neck nodes. The dose and techniques of RT used are tabulated in Table 1. RT was reasonably tolerated, and all the patients completed the planned course of treatment. Dermatitis was the most commonly noted acute RT-related toxicity. Ten patients experienced grade-2 dermatitis, while two patients developed grade-3 dermatitis; RT was delayed by 4 days due to the same in one of these patients. Five patients developed grade-2 mucositis which was the next most common acute toxicity, but no patient had grade-3 mucositis. There were no instances of grade-4 or 5 toxicities. Toxicities were more frequent and severe in patients who received conventionally planned Cobalt-60 teletherapy. Trismus and fibrosis were the most common late toxicities. One patient developed mandibular osteoradionecrosis one-year follow-up.
Table 1. Patient, tumour and treatment characteristics
RT, radiation therapy; CTRT, radiotherapy with concurrent chemotherapy; CT, chemotherapy; IMRT, intensity-modulated radiation therapy; 3DCRT, 3-dimensional conformal radiation therapy; Conv LA, conventional technique using linear accelerator (6MV); Co60, Cobalt-60 teletherapy.
Of the six patients who received concurrent chemotherapy during RT, four received cisplatin, one received carboplatin, while one received paclitaxel and carboplatin. Eight patients did not receive concurrent chemotherapy. Further, a systemic dose of adjuvant chemotherapy with paclitaxel and carboplatin was continued for three of these six patients, while one patient who developed recurrent disease immediately upon completion of concurrent chemoirradiation was started on palliative chemotherapy.
Outcome and patterns of progression
The median follow-up period was 25 months (interquartile range: 7–40 months). Two patients were lost to follow-up at the time of the study. The mean PFS was 21.5 months (standard deviation (SD): 17.3), and the mean OS was 24.6 months (SD—16.9). Median PFS and OS were not reached at the time of analysis. Estimated PFS and OS at 12 months were 61% and 80%, respectively. Seven out of the 16 patients had disease progression, where two patients had only local recurrence, and three patients had distant metastases alone. One patient had both local and distant relapse, while one patient had local and regional recurrence as well as distant metastases (Figure 1). Four patients died due to disease progression. All but one of the patients who had progressed did so within one year of completion of treatment. The sites of distant metastases were brain, lungs, mediastinal nodes and liver (Table 2).
Figure 1. Patterns of recurrences.
Table 2. Treatment, risk factors and outcome of all patients
Y, yes; N, no; NA, not applicable; SM, submandibular gland; conc, concurrent; adj, adjuvant chemotherapy.
*Bold face represents high-risk factors among the patient who progressed.
The univariate analysis for the prognostic factors was determined by log-rank tests. PFS was significantly worse for patients with pathological neck node positivity (p = 0·036 vs. node-negative) and with tumour deposits in peri-parotid nodes (p = 0·007). Although addition of RT did not show a statistically significant improvement in PFS (p = 0·098), both patients who did not receive RT in our cohort recurred locally. As a direct corollary of this, local control was significantly worse for patients who did not receive RT (p = 0·011). Addition of any chemotherapy, regardless of whether concurrent or adjuvant, had no significant impact on the PFS or OS. Pathological neck node positivity with nodal stage of N2 or more was the only variable that significantly correlated with worse OS (p = 0·031). Multivariate analysis using Cox regression models showed a near significant association of pathological node positivity to worse PFS (p = 0·055). The high-risk features, details of treatment, and outcome details have been tabulated (Table 2), and Kaplan–Meier curves for the survival data are depicted in Figure 2.
Figure 2. Kaplan–Meier survival curves (a) progression-free survival (PFS), (b) overall survival (OS).
Discussion
SDC, first described by Kleinsasser in 1968, is a rare and aggressive type of salivary gland neoplasm. Reference Kleinsasser, Klein and Hübner14 There have only been a handful of studies analysing the survival and prognostic factors in SDC with a reasonably large cohort of patients because of its low incidence globally and in India. To the best of our knowledge, this is the only series from India to have analysing the role of RT and the factors affecting PFS and OS in SDC.
Patient and tumour characteristics
Osborn et al. reported the patient characteristics and treatment outcomes of 495 patients with SDC from the National Cancer Database (NCDB) observed that it was more prevalent among older men. Multivariate analysis also showed an inferior OS outcome in advanced age and male gender Reference Osborn, Givi and Lee15 which correlated with the Surveillance, Epidemiology, and End Results (SEER) database. Reference Jayaprakash, Merzianu and Warren16 In our study, although the mean age was 56.7 years and the predominant population was men (94%), there was no significant correlation of age with survival. In concordance with literature, Reference Gilbert, Sharma and Schmitt3,Reference Jayaprakash, Merzianu and Warren16 the parotid gland was the most common site of primary followed by the submandibular gland. Jaehne et al. reported that majority (66%) of patients with SDC presented with a locally advanced, that is, T3/T4 tumour and 56% also had cervical lymph node metastasis at the time of diagnosis. Reference Jaehne, Röser, Jaekel, Schepers, Albert and Loening2 In our study, 62 % (n = 10) of the patients had T3/T4 disease and 56% (n = 9) had cervical nodal metastasis at the time of diagnosis, again corroborating the fact that most SDCs are present in advanced stages. Around 20% had HER2/neu positivity on histology in Jaehne et al.’s study, whereas in our study HER2/neu immunohistochemistry was tested only in four patients (25%) and was positive only in two of them (12.5%).
Treatment and outcome
SDCs have some obvious behavioural and histological differences from other head and neck epithelial mucosal malignancies, and due to the lack of randomised data, their optimal management still remains a conundrum. In general, major salivary gland tumours are primarily managed with surgery with or without neck dissection. Reference Sood, McGurk and Vaz4,Reference Mifsud, Burton, Trotti and Padhya17 In view of a high incidence of nodal metastases, selective neck dissection in clinically N0 neck and modified radical neck dissection in patients with clinically suspicious neck nodes is recommended. In a 20-year review by Stodulski et al., radical surgical approach to the primary and neck nodes was associated with superior DFS and OS. Reference Stodulski, Mikaszewski, Majewska and Kuczkowski18
The benefit of adjuvant RT to the primary site +/- ipsilateral neck is well established in literature for improving the PFS but fails to improve the OS. Qian et al. showed that the lack of adjuvant RT was associated with poor local control. Reference Qian, Di, Guo, Zheng, Ji and Wang19 In concordance to this, the only two patients who did not receive RT in our study developed local recurrence, and adjuvant RT was statistically associated with improvement in local control, albeit a statistically significant association with PFS could not be demonstrated. There was no difference in the outcome observed with the different techniques or doses of RT used in our study, but treatment with conventional cobalt-60 teletherapy was relatively associated with increased toxicity.
Going beyond locoregionally directed treatment, the literature lacks adequate data regarding the role of concurrent chemotherapy and adjuvant chemotherapy in non-metastatic SDCs. In a large NCDB database study, there was no difference in the 3-year and 5-year OS with the addition of chemotherapy to RT. Reference Osborn, Givi and Lee15 In our study as well, addition of neither concurrent chemotherapy nor adjuvant chemotherapy had an impact on survival. Two ongoing prospective randomised trials, RTOG-1008 and GORTEC-SANTAL, aim to evaluate the benefit of concurrent cisplatin with postoperative RT (both studies) and definitive RT (GORTEC-SANTAL) in high-risk SGCs, including SDCs. Reference Rodriguez, El-Naggar and Blvd20,21
In concordance with literature, distant metastases were the most common type of failure in our group of patients (five out of seven patients who progressed). Walvekar et al. previously reported that most high-grade parotid malignancies recur in the first 18 months, although only 9% (ten patients) in their study had SDC. Reference Walvekar, Filho and Seethala22 In our study, all but one recurrences (37.5%, n = 6) happened within the first 12 months. Due to the highly aggressive nature and metastatic potential, distant metastases at presentation and during the course of treatment are not uncommon in SDCs. In the study by Gilbert et al., Reference Gilbert, Sharma and Schmitt3 4% of patients had distant metastases at the time of the diagnosis, and in the study reported by Osborn et al., Reference Osborn, Givi and Lee15 as this percentage was as high as 10%. In our study, one patient (6%) was found to have liver metastases after surgery and prior to RT.
Risk factors for failure
As per the published literature, Reference Jaehne, Röser, Jaekel, Schepers, Albert and Loening2,Reference Gilbert, Sharma and Schmitt3,Reference Osborn, Givi and Lee15,Reference Jayaprakash, Merzianu and Warren16,Reference Stodulski, Mikaszewski, Majewska and Kuczkowski18,Reference Roh, Lee and Choi23–Reference Lee, Lee and Keum27 various factors have been reported to affect local control, distant metastases, PFS and OS, which have been tabulated in Table 3. In our study, the factors that significantly impacted outcome on univariate analysis are tabulated in Table 4. Multivariate analysis did not show any statistically significant association with survival that may be attributed to the limited sample size.
Table 3. Summary of risk factors reported in previous studies
Table 4. Significant prognostic variable impacting outcomes
CRT, concurrent chemoirradiation; FNI, facial nerve involvement; ECE, extracapsular extension; LC, local control; LRFS, local recurrence-free survival; DMFS, distant metastases-free survival; DSS, disease-specific survival; NS, not significant; NR, not reported.
Looking beyond adjuvant radiotherapy
In general, the role of concurrent or adjuvant chemotherapy remains dubious in salivary gland malignancy due to a lack of prospective data, but large retrospective analyses have failed to show an OS benefit in SDCs. Reference Osborn, Givi and Lee15 SDCs manifest close histologic features with invasive ductal carcinoma of the breast and a high incidence of AR expression that mimics prostatic adenocarcinoma. Reference Udager and Chiosea28 Addressing the common occurrence of distant metastases should therefore predominate the required pattern of treatment, which suggests the need for a robust strategy for adjuvant systemic therapy after local therapy.
HER2/neu overexpression is frequently seen in a high proportion of patients with SDCs; in our study, only four patients had been tested for the same and two of them were positive (50%). Although it is an unfavourable parameter by itself, it opens up the avenue of using trastuzumab as an agent for systemic therapy in these patients. A promising retrospective study by Hanna et al. showed significant improvement in PFS and OS among HER2/neu-positive patients who received adjuvant Trastuzumab along with chemotherapy and RT for high-risk non-metastatic salivary gland carcinoma (SDCs comprised 45% of the population). Reference Hanna, Bae and Lorch29 AR expression is well established in SDCs, and various studies have explored the use of androgen-deprivation therapies in metastatic or recurrent SDCs. Reference Boxtel, Verhaegh and Grunsven30,Reference Fushimi, Tada and Takahashi31 EGFR expression is a potential prognostic indicator and target, incidence of which was high (70%) in the study by Williams et al. and correlated significantly with local recurrence. Reference Williams, Roberts, Kies, Mao, Weber and El-Naggar9 Han et al., however, did not find any relationship with recurrences or distant metastases. Reference Han, Roh and Choi32 Recently, programmed cell death ligand (PDL) receptors, viz. PDL1 and PDL2 expression has been identified in SDCs, but clinicopathological studies done in Japan Reference Sato, Akiba and Kawahara33 and USA Reference Hamza, Roberts, Su, Weber, Bell and Ferrarotto34 show contradictory results on its prognostic value.
Although the role of molecular-targeted therapy in patients with non-metastatic SDC is debatable at present, future studies exploring therapeutic options in patients with advanced, high-risk SDCs are warranted to improve survival among these patients.
Limitations
Although our study is limited by the selection and treatment bias inherent to single institutional, retrospective studies of small sample sizes, we provide one of the largest series from our country describing the patterns of recurrence and prognostic risk factors for an extremely rare histology. HER2/neu status was not available in all patients. Our follow-up time was relatively short, and follow-up data for some patients in our cohort were limited by the lack of compliance to recommended follow-up. However, this was compensated to a reasonable extent by telephonic correspondence.
Conclusion
SDC is a rare, locally aggressive malignancy with a high propensity to develop distant metastases. Radical surgery followed by adjuvant radiation therapy is the cornerstone for achieving optimum local control. Poorer prognosis was observed among patients who had metastatic deposits in either cervical nodes or peri-parotid nodes on histopathology. As systemic failures are more predominant among these patients, larger prospective trials are needed to formulate an optimum strategy for choice and sequencing of first-line systemic therapy, as well as to explore newer avenues like targeted therapy, for improving survival in these patients.
Acknowledgements
BS would like to thank Dr H Thomas for technical editing, language editing and proofreading of this manuscript.
Financial Support
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Conflicts of Interest
None.
Consent for Publication
All authors have read and approved the final manuscript and give consent to the publication of the manuscript in Journal of Radiotherapy in Practice should the article be accepted by the Editor-in-chief upon completion of the refereeing process.
