Comprehensive Treatment for Major Salivary Gland Carcinoma Based on Intensity-Modulated Radiotherapy with or without Radical Surgery

The present study aimed to determine a treatment strategy and Intensity-Modulated Radiotherapy (IMRT) target volume for major salivary gland carcinoma (SGC). The following primary tumor sites were identied: parotid gland in 61 (69.3%) patients, submandibular gland in 21 (23.9%) patients, and sublingual gland in six (6.8%) patients. Lymphoepithelial carcinoma (LEC) was the most common tumor subtype that accounted for 23.9% of cases. A total of 80 (90.9%) patients received radical surgery combined with postoperative radiotherapy. Eight patients (9.1%) received denitive radiotherapy: six patients with advanced-stage disease received induction chemotherapy (IC) combined with concurrent chemoradiotherapy (CCRT), and two patients with early-stage disease received CCRT. Complete response was observed in these eight patients after treatment completion. The median follow-up time of all patients was 42 months (range: 4–129 months). No patient developed local recurrence. The 5-year overall survival, regional failure-free survival, distant metastasis-free survival, and progression-free survival probabilities were 84.1%, 95.6%, 75.3%, and 75.7%, respectively. Distant metastasis was observed in 18 (20.5%) patients, followed by regional 2 (2.3%) recurrence. Permanent facial nerve injury was conrmed in 31 patients by follow-up. None of the patients experienced facial nerve paralysis in the denitive radiotherapy group.


Background
Salivary gland carcinomas (SGCs) are malignant neoplasms that account for approximately 1.0-8.5% of all head and neck cancers [1] [2]. Recent studies have shown that the incidence of SGC is increasing every year, while the age of onset is decreasing [3] [4]. Guidelines from the National Comprehensive Cancer Network (NCCN) recommend surgery as the primary treatment for SGCs and postoperative radiotherapy for patients with advanced-stage disease or high-risk factors [5]. The major complications of surgery include facial nerve injury, with reported incidence of 12-40% [6].
According to previous literature, lymphoepithelial carcinomas (LECs) are malignancies that have morphological features similar to those of undifferentiated nasopharyngeal carcinomas (NPCs) and occur in areas outside the nasopharynx, such as in the salivary gland, lung, and stomach [7][8] [9] [10]. NPCs are sensitive to chemotherapy and radiotherapy, and a radical effect can be achieved by chemoradiotherapy [11]. However, there are few reports on major salivary LECs. Radical surgery is also the primary treatment for this tumor subtype according to the NCCN guidelines. In addition, LECs often invade the facial nerve due to a high malignancy degree. For this reason, some surgical scholars believe that facial nerve preservation will affect the safety of surgery and thus advocate for the removal of the involved facial nerve to reduce local recurrence [12]. In short, surgery will likely cause permanent damage to the facial nerve in addition to dis gurement, which might affect patients' quality of life and self-con dence.
Intensity-modulated radiotherapy (IMRT) has become the standard treatment technique for head and neck cancer. However, there are few studies on target volume delineation for SGCs. Previous recommendations have determined the IMRT target volume based on the "tumor bed", "surgical bed", or even "parotid bed" [13] [14] [15]. However, our prior study suggests that these recommendations may not consider individual subtleties associated with the exact location of the primary lesion [16]. They also did not make full use of the IMRT advantage, which is that the high-dose areas conform closely to the three-dimensional shape and scale of the tumor.
Based on these considerations, the present retrospective study summarized the clinical characteristics of SGCs and analyzed failure patterns in patients treated with IMRT to provide a reference for individualized SGC treatment.

Methods And Materials
Patients A total of 96 patients with SGCs were retrospectively evaluated between August 2009 and August 2020. The patients were restaged according to the 8th edition of the American Joint Committee on Cancer (AJCC) staging system for major salivary glands. The exclusion criteria were as follows: (1) evidence of distant metastasis before treatment, secondary malignancy, or both; (2) non-epithelial tumors. The ethics committee of Sun Yat-sen University Cancer Center approved the study protocol.

Diagnosis
All patients underwent a comprehensive exam and evaluation that included computed tomography (CT) or magnetic resonance imaging (MRI) of the head and neck, chest radiography, abdominal ultrasonography, emission computed tomography, or positron emission tomography-computed tomography. Some patients also underwent color Doppler ultrasound imaging of the salivary glands. The nal diagnosis was based on histopathology or cytopathology results. Due to the histological nding similarity, distinguishing LECs from lymph nodes containing NPC metastases was especially important. Nasopharyngoscopy was conducted in all patients in order to make a correct diagnosis.

Chemotherapy
The induction chemotherapy (IC) regimen was a combination of taxanes, cisplatin, and uorouracil (TPF), comprising intravenous infusion of docetaxel at a dose of 50 mg/m 2 on day 1, intravenous infusion of cisplatin at a dose of 60 mg/m 2 on day 1, and continuous intravenous infusion of 5-uorouracil at a dose of 500 mg/m 2 /day on days 1-5 for 120 h, three times per week, for a total of four cycles. If tumor shrinkage achieved a partial response (PR) or above after two cycles of IC, patients continued IC for up to four cycles and were administered concurrent chemoradiotherapy (CCRT). Otherwise, the patients received CCRT after two cycles of IC or were reconsidered for surgery. CCRT treatment prescribed cisplatin at a dose of 30 mg/m 2 of intravenous infusion on day 1, for 4-6 cycles weekly [17].

Radiotherapy
De nitive radiotherapy: prior to radiotherapy, patients were immobilized with head-and-neck thermoplastic masks in a supine position. A CT simulation was then performed using 3-mm slices of the head and neck within 1-2 weeks after IC [17].
Postoperative radiotherapy: postoperative MRI of the head and neck was performed 3 weeks after surgery when the wound had healed. The patients were immobilized in a supine position with a head-and-neck thermoplastic mask [16].
Some patients were immobilized with a bolus to the skin if necessary.

Target volume delineation
De nitive radiotherapy: gross tumor volume of primary site/regional lymph node (GTVp/nd) was de ned as the volume of the primary tumor including lymph node metastasis. Medium risk clinical tumor volume (CTV1) was de ned as GTVp/nd plus a 0.5-to 1.0-cm margin. Low-risk CTV (CTV2) was de ned as CTV1 plus a 5-mm margin together with the regional selective lymph drainage areas. According to prior studies, ipsilateral level Ib-Va and -should be included in parotid gland cases, ipsilateral level I-Va should be included in submandibular gland cases, and bilateral level I-Va should be included in sublingual gland cases [18] [19] [20]. For patients with advanced-stage disease, the GTVp/nd was contoured according to the tumor regression after IC [17].
Postoperative radiotherapy: among the reserved tissues, those located <5 mm from the invasive tumor edge before surgery were de ned as high-risk CTV (CTV-HD); those located <10 mm away were de ned as CTV1; and those located 10-20 mm away together with the regional selective lymph drainage areas were de ned as CTV2 [16]. Examples of target volume delineation are presented in Figure 1.
Normal structures, including the mandible, brainstem, temporal lobe, oral cavity, middle ear, and spinal cord, were also contoured slice-by-slice in the treatment-planning CT scans [17].
Planning target volumes (PTV) were generated by addition of a 3-5-mm margin to all GTV/CTV values [17]. Table 1 summarizes the target volume de nitions. Table 1

Follow-up
The follow-up time was until August 2021 or the date of death. Treatment failure was con rmed by biopsy. Failure was de ned in accordance with the de nition provided by Chao et al. [21].

Statistical analysis
Estimates of overall survival (OS), regional failure-free survival (RFFS), distant metastasis-free survival (DMFS), and progression-free survival (PFS) were obtained using the Kaplan-Meier method. Statistical calculations were performed using SPSS version 25.0 (IBM SPSS Statistics for Windows, Armonk, NY, USA).

Patient characteristics
Eight patients were excluded from the analysis due to development of distant metastases during treatment or diagnosis of nonepithelial tumors. As a result, a total of 88 patients with SGCs were selected for analysis of clinicopathological pro les: 33 females (37.5%) and 55 males (62.5%). The onset age ranged from 18 to 77 years, and the median age was 45 years. The primary tumor sites were as follows: parotid gland in 61 (69.3%) patients, submandibular gland in 21 (23.9%) patients, and sublingual gland in six (6.8%) patients. LEC was the most common tumor subtype (23.9%). The type of surgery was determined by surgeons. A total of 80 (90.9%) patients received radical surgery with or without neck dissection combined with postoperative radiotherapy. Eight patients (9.1%) received de nitive radiotherapy: six patients with advanced-stage disease received IC combined with concurrent CCRT, and two patients with early-stage disease received CCRT. The details of patient and tumor characteristics are summarized in Table 2. Abbreviations: IMRT = intensity-modulated radiotherapy; IC = induction chemotherapy; CCRT = concurrent chemoradiotherapy; S = surgery; RT= radiotherapy; AR-IMRT = adaptive re-planning intensity-modulated radiotherapy * Patients received surgery in another center and their operation records could not be found.

E cacy
After completion of de nitive radiotherapy, complete response (CR) was observed in all of the patients. Table 3 summarizes clinical characteristics and outcomes of these eight patients. Cases 1-3 have been previously described by our team [22].
Updated follow-up data showed that Cases 1 and 2 were still alive without evidence of disease, while Case 3 died of distant metastasis. Case 6 was a 52-year-old man who achieved a PR after IC. After completion of de nitive radiotherapy, CR was observed via MRI, and the patient showed no evidence of disease until the last follow-up (2021-8). Details of Case 6 are presented in Figure 2.  Figure  3).
Clinical characteristics of the two patients with regional failure are summarized in Table 4. The pre-RT MRI showed an insigni cantly enlarged lymph node that did not meet the diagnostic criteria in both cases, which were contoured in CTV2 (Figure 4-5).

Treatment Toxicity
Thirty-one patients who received surgery experienced peripheral facial paralysis. The relationship between facial paralysis and SGC subtypes is shown in Figure 6. LECs with facial paralysis accounted for 7/12 patients (58.3%), and other pathological types accounted for 19/63 patients (30.2%). Eight patients without surgery had an intact facial nerve and a normal appearance without reconstruction. In addition, Case 2 was a 33-year-old man who was admitted with a slowly growing mass in the left periauricular region and facial paresis. Six months after de nitive radiotherapy, the patient's facial nerve function was fully

Discussion
Histological classi cation of SGCs is very demanding and 24 subtypes have been speci ed according to the World Health Organization classi cation of malignant salivary gland tumors [23]. In 666 patients with SGCs in a study performed in the Netherlands, for which the pathology results were revised, adenoid cystic carcinoma (27%) was the most frequently diagnosed, followed by mucoepidermoid carcinoma (16%) and acinic cell carcinoma (14%) [24]. No satisfactory chemoradiotherapy method for the above tumor subtypes has been reported, while surgery was considered to be a more effective treatment [5] [25]. LEC has distinct racial and regional characteristics and is well known to occur in a limited number of patients in the localized regions of Southwest Asia, southern parts of China, and the Arctic Circle [30] [31]. In 235 patients with SGCs in a study performed in China, LECs were diagnosed in 21.2% of the cohort, which is a much higher rate than the reported average incidence of this disease in the western world [32]. Results of the present study are similar to this previous report ( There is still confusion concerning the optimal radiation target volume for SGCs. As reported previously, the present study relied on surgical principles to determine the IMRT target volume [16]. The 5-year OS, RFFS, DMFS, and PFS were 84.1%, 95.6%, 75.3%, and 75.7%, respectively. No patient developed local recurrence, and the main cause of failure within the study cohort was distant metastasis, which suggested that the method was reasonable and worthy of further research.
The present study included 14 patients with recurrent SGCs after primary surgery. In general, the rst treatment plays a major role in cancer. But perhaps because of the special anatomic location of the salivary glands, secondary operation plus postoperative radiotherapy for recurrent SGCs also showed good clinical outcomes. Consequently, patients with recurrent SGCs are expected to strive for radical treatment.
The pre-RT MRI showed an insigni cantly enlarged lymph node that did not meet the diagnostic criteria for both regional failures, which were contoured in CTV2 (Figure 4-5). According to a previous report, high rates of implicit metastasis of approximately 12-45% were observed for lymph nodes in SGCs cases, suggesting that it is very important in clinical practice to determine whether the lymph nodes have been spared or not [42] [43]. However, the optimal treatment for risky lymph nodes that do not meet the diagnostic criteria remains to be determined. Guidelines from the NCCN recommend prescription doses of 44-50 Gy and 54-63 Gy for low and intermediate risk sites of suspected subclinical spread, respectively [5]. In addition, recurrence observed in a previous study occurred in a cervical lymph node that was not signi cantly enlarged, but was probably involved, and received a radiation dose of about 64 Gy [44]. Consequently, to control the more than microscopic disease, a dose of 63-65 Gy has been irradiated for the risky lymph nodes in the following treatment.
The present study has some limitations. First, it was a retrospective study from a single center, and further prospective multicenter studies are needed. Second, patients without surgery received lesion site ne needle aspiration biopsy in our study.
According to previous reports, ne needle aspiration biopsy showed a sensitivity and a speci city of 41.7-92.8% and 93.9-98.5%, respectively [45][46] [47]. Whether intraoperative frozen sections should be performed to obtain more pathological information is need to be studied further.

Conclusions
LECs may be sensitive to chemoradiotherapy, which may achieve a radical effect and avoid unnecessary surgical injury. IC combined with CCRT is expected to become a new treatment strategy for advanced LECs. The IMRT target volume delineation according to the surgical principles may be a more promising method with good clinical e cacy that is worthy of further study.  showed that the reserved tissues included the retained parotid tissues (blue line), posterior venter of the digastric muscle, carotid sheath, styloid process and mandibular branch. (D) Posterior venter of the digastric muscle, carotid sheath, styloid process and mandibular branch were all less than 5 mm from the primary tumor edge, so they were very likely to be invaded, as delineated in the CTV-HD area. The retained parotid tissues were more than 10 mm away from the primary tumor edge, so they were delineated in the CTV2 area.  (A) Postoperative MR showed a lymph node measuring 6 mm; (B) The patient complained of a palpable node after 5 years, and regional recurrence was nally diagnosed; (C) Target volume delineation of postoperative radiotherapy; (E) Dose color wash; (F) DVH.

Figure 5
Page 19/19 (A) Postoperative CT showed a lymph node measuring 4 mm; (B) Regional recurrence was nally diagnosed by CT after 2 years; (C) Target volume delineation of postoperative radiotherapy; (E) Dose color wash; (F) DVH.

Figure 6
The relationship between facial paralysis and SGCs subtypes. Abbreviations: LEC = lymphoepithelial carcinoma

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.