Efficacy of Surgery in Patients with Stage I Primary Parotid Gland Lymphoma: A Population-Based Study

Background Limited and controversial evidence is available on the efficacy of surgery for patients with stage I primary parotid gland lymphoma. Thus, we aimed to investigate whether surgery can enhance the prognosis of patients with stage I primary parotid gland lymphoma using large sample data. Methods From 1998 to 2015, we searched the Surveillance, Epidemiology, and End Results (SEER) program database and extracted information regarding patients with stage I primary parotid gland lymphoma; we classified these patients into surgery and nonsurgery groups. We calculated overall survival (OS) and cancer-specific survival (CSS) using Kaplan–Meier curves and log-rank testing. Propensity score matching (PSM) analysis was also used to further account for confounding variables before comparing the OS and CSS again. We used the COX proportional hazard regression model in both multivariate and univariate analyses. Results We enrolled 918 patients with stage I primary parotid gland lymphoma, among which 656 (71.5%) patients underwent surgery. Before PSM, the surgery group had better OS (hazard ratio (HR) = 0.673, 95% confidence interval (CI): 0.519–0.873, and p=0.003) and CSS (HR = 0.595, 95% CI: 0.403–0.879, and p=0.008) than the nonsurgery group. After PSM, surgery was still a beneficial factor for OS (HR = 0.569, 95% CI: 0.399–0.810, and p=0.002) and CSS (HR = 0.384, 95% CI: 0.220–0.669, and p=0.001). Furthermore, in univariate and multivariate analyses, total parotidectomy significantly increased OS (p=0.001 and p=0.021, respectively) and CSS (p=0.001 and p=0.037, respectively). Conclusions In summary, the prognosis of patients with stage I primary parotid gland lymphoma can be significantly improved by surgery. Moreover, total parotidectomy was a protective factor for OS and CSS before and after PSM analysis, suggesting that surgery acts as a significant component in multimodal therapy for early primary parotid gland lymphoma.


Introduction
Primary parotid gland lymphoma is a rare type of head and neck lymphoma, with a total incidence of roughly 0.3% of all cancer cases, 2%-5% of salivary gland tumors, and 5% of extranodal lymphomas [1][2][3]. Patients with primary parotid gland lymphoma often have a lower-grade lymphoma and better prognosis than those with other extranodal lymphomas [4,5]. Primary parotid gland lymphoma is described by its asymptomatic progress in the parotid gland for 4-6 months [6]. Swelling of both parotid glands, cervical lymphadenopathy, pain, and facial nerve paralysis are the other symptoms [7,8]. Te clinical outcome of parotid lymphoma appears similar to other benign parotid swellings, thus making the diagnosis difcult. Lymphomas are commonly overlooked in a new parotid mass in preoperative evaluations [9]. Te presence of autoimmune conditions, such as Sjögren syndrome, is linked to an elevated risk of lymphoma (44%). Moreover, these conditions increase the risk of lymphomas of the parotid gland, particularly marginal zone lymphoma [10].
Radiotherapy, chemotherapy, or both may be used to modify treatment to the specifc grade and histologic type; this provides a positive prognosis in most cases [11][12][13]. However, there have been controversies over the role of surgery in primary parotid gland lymphoma. Surgical excision can enhance OS [14][15][16]; however, some studies found that it did not afect patients' survival [6]. Furthermore, surgery acts only as a diagnostic tool [17][18][19].
We used population-based data from the Surveillance, Epidemiology, and End Results (SEER) database to determine whether surgery for stage I primary parotid gland lymphoma was efective. We also assessed the survival efects of various surgical methods.

Data Sources.
We used the SEER program database of the National Cancer Institute, which gathers data on cancer incidence and survival from 15 states and covers roughly 34.6% of the US population. Using SEER * stat software package 8.3.5, we extracted eligible cases from the SEER-18 registry (with modifed treatment felds). Te data were free to download from the SEER database and thus did not require informed consent from patients; we fled a request to the SEER database project and obtained approval. From 1998 to 2015, we gathered the following information from the SEER database for all patients with primary parotid gland lymphoma: demographics, pathological types, Ann Arbor stage I, treatment information (surgery, radiation, and chemotherapy), and survival. Te following patients were excluded: (1) those who did not have primary parotid gland lymphoma as the sole malignancy, (2) those who did not provide information on race or marital status, (3) those who did not have a clear surgical therapy, and (4) those who had a survival time of zero.

Statistical Analysis.
Descriptive statistics were used for the assessment of the demographic and tumor features of the patients. Categorical variables were assessed using the chisquare test. Te Cox proportional risk model was used to assess characteristics that were independent predictors of OS and CSS. Te multivariate Cox analysis included factors that were substantially linked with prognosis in the univariate Cox analysis. Moreover, the hazard ratios (HRs) and 95% confdence intervals (CIs) were computed using the Cox proportional hazards models. A Kaplan-Meier survival curve was created using this technique; the log-rank test was used to assess the diference between the survival curves.
An analysis of propensity scores on a one-to-one basis was performed using R statistical language to reduce baseline diferences between the two groups. All statistical analyses were performed using R statistical language (Version R 4.2.1). p < 0.05 was considered statistically signifcant when using two-sided tests.

OS and CSS Comparison.
A comparison of OS and CSS of patients with stage I primary parotid gland lymphoma is shown in Table 3. Before PSM, the surgery group had a better OS (HR � 0.673, 95% CI: 0.519-0.873, p � 0.003) (Figure 1 Figure 2, surgery can signifcantly improve prognosis. Specifcally, the efects of diferent surgical procedures on OS and CSS are shown in and p � 0.065, respectively). Furthermore, sex and chemotherapy had little efect on the results of individuals treated with surgery ( Figure 3).

Discussion
Te early diagnosis of primary parotid gland lymphoma is still challenging as the condition can be easily misdiagnosed or even left unnoticed because the incidence of primary parotid gland lymphoma is extremely low and the initial clinical symptoms are untypical. Te main diagnostic tool for primary parotid gland lymphoma is surgery since the accuracy of fne-needle aspiration biopsy is low and the radiological features are not obvious [20,21]. Magnetic resonance imaging, computed tomography, and ultrasound do not provide much additional information at the onset of parotid swelling [22]. Parotidectomy is highly recommended both for treating the tumor and for histologically diagnosing the tumor for further follow-up planning [2]. Depending on its grade and stage, the treatment of primary parotid gland lymphoma is determined. Te majority of parotid lymphomas are low-grade and locally confned Ann Arbor lymphomas of Stage I or II that can be treated with surgery, adjuvant radiation or chemotherapy, or both, depending on their stage and grade [14,[23][24][25][26].
Surgery is performed in most cases where systemic treatment is required; however, the procedure is categorized as a diagnostic measure [1,9,[27][28][29][30]. Conversely, surgical excision efectively reduces the tumor size and improves the prognosis. Te fve-year survival rate reached 61% for parotid lesions with a diameter of >6 cm and increased to 87%   when the largest diameter was <6 cm [31,32]. Furthermore, total excision of all known lymphomas, such as stomach lymphoma, may be curative for some subtypes of lymphomas [33,34]. Te treatment method did not afect the OS of patients with low-grade primary parotid gland lymphoma. In most cases, parotidectomy alone might be considered a curative option [24,25].
Te fndings of this study indicate that surgical treatment of stage I primary parotid gland lymphoma may provide some survival benefts. Te multivariate analysis results revealed that surgery was an independent predictor of living a longer life, in agreement with a previous retrospective analysis, wherein Feinstein et al. analyzed data collected from the SEER database concerning 2,140 patients with parotid lymphoma. Tey found that patients who underwent surgery had a 35% lower death rate than patients who did not [14]. Vazquez et al. found that there was no statistically signifcant diference in survival among groups treated with surgery, radiation therapy, or both for MALT lymphoma of the salivary glands [25]. Olivier et al. studied 35 patients diagnosed with Ann Arbor Stage I and II NHL of the parotid gland and reported 90% OS and 71% disease-specifc survival at 5 years and 10 years after surgery, with no signifcant diference between the radiation and surgery groups [35].
Furthermore, we analyzed the efect of diferent surgical procedures on the outcome. Compared with patients who did not undergo surgery, patients who underwent total parotidectomy had better OS and CSS, but patients with superfcial parotidectomy only had better OS. As opposed to previous studies, Mehle et al. found that surgery does not afect the prognosis regardless of whether it is superfcial or total parotidectomy [6]. Tis was probably due to their sample size being small (only 16 subjects), so they could not accurately assess the efects. In the case of malignant tumors without facial nerve involvement that can be completely resected during surgery, superfcial parotidectomy was proven to be an efective therapeutic method for unifocal, early-stage parotid gland lymphoma [36]. When the facial nerve is confrmed to be infltrated preoperatively or during surgery, total parotidectomy combined with resection of the facial nerve is recommended [37]. In the surgery group, the prognosis was better if the patient was younger than 65 years; therefore, age should be considered an important factor in clinical practice when considering whether surgery should be performed. Surgery is recommended for white people, unmarried individuals, and people receiving radiotherapy since they have a better prognosis. Patients diagnosed with parotid lymphoma due to MALT had better prognosis in the surgery group, consistent with previous studies. Early-stage parotid gland MALT lymphomas have an idle behavior and tend to remain localized for long periods, requiring less aggressive treatment [7].
Tere are several limitations to our study. First, because this was a retrospective study, some selection bias was inevitable; it is possible that patients in the surgery group were healthier and had lower tumor loads than those in the nonsurgery group. Second, there is no detailed information in the SEER database regarding the purpose, timing, and outcome of surgery; the recurrence rates following surgery also remain unclear. Terefore, a multidisciplinary comprehensive evaluation is essential when determining which patients should be recommended for surgery. Nevertheless, the strengths of the study are that this is the frst populationbased study to assess the efcacy of surgical therapy in   Journal of Oncology patients with early-stage primary parotid gland lymphoma. Furthermore, we drew the study population from a nationally representative dataset, which may have reduced potential selection bias to an extent. Considering that both multivariable and PSM analyses were conducted and that no signifcant changes were noted in OS, we can afrm that our fndings are valid and stable.

Conclusions
In summary, surgery can considerably improve the OS and CSS of patients with stage I primary parotid gland lymphoma, thus supporting an increased role of surgery in multimodal treatment.

Data Availability
Te data supporting the study's fndings are freely available in the Surveillance, Epidemiology, and End Results' Program at https://seer.cancer.gov/.

Ethical Approval
Te study employed deidentifed data and followed the World Medical Association's Helsinki Declaration for Ethical Human Research.