A retrospective analysis of all patients with SDC was performed at our institution. A list of SDC patients between January 2005 and December 2012 from an institutional cancer registry identified 16 patients. Patients who had salivary gland surgery and histologically proven SDC were included in analysis. All patients received postoperative radiotherapy; patients who failed to finish scheduled radiotherapy were excluded. We excluded 1 patient who received 16.2 Gy of radiotherapy because patients who fail to finish the scheduled radiotherapy were excluded from analysis. He received salvage re-irradiation after recurrence in postoperative tumor bed and successfully achieved disease control, but lung metastasis occurred later.

Of the 16 patients identified, 15 eligible patients were analyzed. The characteristics of the 15 patients are in Table 1. The median age was 61 years (range, 40 to 71 years) with 13 (87%) older than 50 years, and 12 patients (80%) were men. Twelve patients (80%) had a tumor in the parotid gland, 2 (13%) in the submandibular gland, and 1 (7%) in the maxillary sinus. Nine patients (60%) had T3 or T4 disease, and 9 (60%) had positive nodal disease. Seven patients (44%) had a positive resection margin, 8 (62%) had perineural invasion, 5 (42%) had lymphovascular invasion, and 6 (38%) had extracapsular extension of metastatic lymph nodes.

Pretreatment evaluation included a complete history and physical examination. Laboratory studies included a complete blood cell count and serum chemistry profile. For stage workup, patients underwent imaging studies, such as computed tomography (CT) and/or magnetic resonance imaging (MRI). Positron emission tomography, whole body bone scans, and chest CT were performed for systemic evaluation.

Patients received surgery and postoperative radiotherapy. Postoperative radiotherapy was administered after wound recovery from surgery using ≥6 MV photons generated by a linear accelerator. Either 3-dimensional conformal radiotherapy (3D-CRT) or intensity-modulated radiotherapy (IMRT) was used. Contrast-enhanced CT simulation with thermoplastic device was mandatory and a mouthpiece to remove the tongue from the radiation field was used according to the radiation oncologist's preference. Gross tumor volume (GTV) included residual tumor volume and metastatic lymph nodes. Clinical target volume (CTV) was defined as GTV plus a 2-cm margin or postoperative tumor bed plus a 1-cm margin and encompassed the elective nodal area. In all patients, elective nodal irradiation (ENI) target volume was the nodal areas of the operated neck including level II, level III, and level IV neck nodes. We used 3D-CRT, multifield technique, and cone-down technique. Total radiation dose was 63 Gy if the patient had risk factors, such as gross residual tumor or extracapsular extension of metastatic lymph nodes; if a patient had no risk factors, the total radiation dose was 59.4 Gy. For N0 patients, after a dose of 45 Gy in 25 fractions, CTV was reduced to the postoperative tumor bed plus a 1-cm margin and ENI of level II and level III to deliver an additional boost dose of 14.4 Gy. Using IMRT, a simultaneously integrated boost technique was used so the GTV received a dose of 2.12 Gy per fraction. CTV1 encompassed the elective nodal area of level II and level III received 1.8 Gy per fraction; CTV2 encompassed the elective nodal area of level IV and received 1.7 Gy per fraction. The total fraction number of IMRT was 30. Total doses were 63.6 Gy for GTV, 54 Gy for CTV1, and 51 Gy for CTV2. Total IMRT dose was similar to total dose of 3D-CRT in biologically equivalent doses. CTV was modified to reduce the dose to organs at risk, such as the spinal cord, optic apparatus (optic nerves and chiasm), mandible, pharynx, larynx, and esophagus. The maximal dose constraints were below 45 Gy for the spinal cord, below 55 Gy for the optic apparatus, and below 70 Gy for the mandible. Other organs at risk, such as the larynx, pharynx, and esophagus, were also delineated and set for doses as low as possible. For the planning target volume, a 0.3-cm margin was applied to the CTV considering patient motion and daily setup error.

Four patients received concurrent weekly chemotherapy of cisplatin (40 mg/m²). One patient aged 70 years received a dose of 25 mg/m² because he had poor general performance status. Cisplatin was administered on day 1. Indications of chemotherapy were gross residual tumor, positive resection margin, and extracapsular extension of metastatic lymph nodes. After scheduled postoperative radiotherapy, no adjuvant chemotherapy was given.

Each patient was examined weekly to monitor treatment-related toxicities through all phases of treatment or more often if clinically indicated. After radiotherapy completion, patients visited the hospital every 3 months for the first 2 years, every 6 months for the next 3 years, and annually thereafter. Follow-up studies consisted of blood tests including complete blood count, serum chemistry profile, and imaging studies, such as MRI or CT to assess treatment response. Response was evaluated using the Response Evaluation Criteria in Solid Tumors after completion of the therapy. Acute toxicities were recorded according to the National Cancer Institute's Common Toxicity Criteria ver. 3.0.

Local failure was defined as increase in size of the residual tumor or a new lesion in the postoperative tumor bed, or failure in the radiation field. Regional failure was defined as the occurrence of a new lesion in the regional lymph node area or soft tissues adjacent to the nodal area. Distant failure was defined as metastatic lesion in distant organs, such as the lung, mediastinum lymph node, or brain.

All statistical analysis used SPSS ver. 20.0 (IBM, Armonk, NY, USA). Locoregional failure-free survival (LRFFS), distant failure-free survival (DFFS), progression-free survival (PFS), and overall survival (OS) were calculated from the date of diagnosis to the event and were plotted using the Kaplan-Meier method. Differences in survival between risk factors were tested using a log-rank test. A p-value <0.05 was considered statistically significant.

Treatment characteristics of patients are summarized in Table 2. Median total radiotherapy dose was 60 Gy (range, 52.5 to 63.6 Gy) and median fractional dose was 1.8 Gy (range, 1.8 to 2.25 Gy). All patients received postoperative tumor bed irradiation and elective nodal irradiation of level II, III, IV; 11 patients (73%) were treated with unilateral neck irradiation and 4 (27%) were treated with bilateral neck irradiation. Total resection of the primary site was performed for 12 patients (80%), and 3 patients (20%) underwent tumor resection only. Neck dissection was done for 10 patients (63%) and 4 (25%) underwent weekly cisplatin-based concurrent chemoradiotherapy. Of 10 patients with neck dissection, 7 received modified radical neck dissection, 2 had supraomohyoid neck dissection, and 1 had lateral neck dissection.

The median follow-up time was 38 months (range, 24 to 105 months) for all patients and 38 months (range, 25 to 105 months) for patients who were still alive at the time of analysis. Fig. 1 shows survival analysis for SDC patients. Four-year rates were 86% for LRFFS, 51% for DFFS, 46% for PFS, and 93% for OS.

Patterns of failure are in Fig. 2. Local failure was observed for 2 patients (13%), and distant failure was observed for 7 (47%). In contrast to the relatively high rate of distant failure, local or regional failure was less common. Of 2 patients who had local failure, one had SDC at the parotid gland and received total parotidectomy. Pathologic stage was pT3N0M0 with positive perineural invasion. The patient had supra-auricular skin recurrence at 8 months after postoperative radiotherapy, apparently within the radiotherapy field. Retrospective review of the radiation field indicated that the radiation dose to the superficial skin was insufficient. After wide resection of the skin at the recurrent disease and ear covering with an anterior lateral thigh flap, the patient recovered and had no evidence of disease at last follow-up. Another patient who initially had a tumor at the left parotid gland with pathologic stage pT3N2bM0, perineural invasion, lymphovascular invasion, and extracapsular extension of metastatic lymph nodes experienced local recurrence in the internal auditory canal. Recurrent disease at the internal auditory canal occurred just outside the border of the radiotherapy field. The patient had synchronous local and distant failure with lung metastasis. Although salvage re-irradiation was attempted, the patient died from lung metastasis progression. Table 3 shows distant failure sites. The lung was the most commonly involved distant metastasis site.

Risk factors for survival are summarized in Table 4. Positive resection margin showed only a trend toward improved PFS (71% vs. 15%, p = 0.071). Only 2 patients had distant metastasis with positive resection margins, whereas 2 had local failures and 4 had distant metastases with negative resection margin. Age, sex, T stage, N stage, perineural invasion, lymphovascular invasion, extracapsular extension, radical surgery, concurrent chemoradiotherapy, and radiotherapy dose were not associated with LRFFS, DMFS, PFS, or OS.

No grade 3 or 4 acute toxicities were observed. The most frequent toxicities were grade 2 oral mucositis and grade 2 radiation dermatitis.