Characteristics and Treatment of Primary Mediastinal Seminomas: A Single-center Report of 27 Cases


 Background: The low incidence of primary mediastinal seminomas has precluded the development of clinical trials on mediastinal seminomas. We evaluated the characteristics, treatments, and prognosis of patients with primary mediastinal seminoma. Methods: We retrospectively collected data on the clinicopathologic characteristics, treatments, toxicities, and survival of 27 patients from a single center between 2000 and 2018. Mediastinal lymph node staging for lung cancer and Masaoka staging for thymic neoplasms were used for disease characterization. Survival was assessed using the Kaplan-Meier method. Univariate analysis was performed using the log-rank test. Results: The median age was 28 (13-63) years. The most common symptoms were chest pain (29.6%), cough (25.9%), and dyspnea (22.2%). Twenty-four patients showed tumor invasion into adjacent structures. Seven and two patients were diagnosed as having lymph node metastasis and distant metastasis, respectively, whereas 48.1% of patients were diagnosed as having Masaoka stage IIIb disease. Sixteen patients (59.3%) had undergone radiotherapy, whereas 25 (92.6%) had undergone chemotherapy. The most widely used chemotherapy regimens were bleomycin, etoposide, and cisplatin. The median follow-up period was 32.23 (2.7-198.2) months. At 5 and 10 years, the rates of local regional relapse-free survival were 90.9% and 90.9%; overall survival, 100.0% and 80.0%; progression-free survival, 86.4% and 86.4%; distant metastasis-free survival, 95.2% and 95.2%; and cancer-specific survival, 100.0% and 100.0%, respectively. Conclusions: Primary mediastinal seminoma was frequently diagnosed in patients with tumor invasion into adjacent structures. Chemotherapy was the most widely used treatment. The disease was sensitive to chemoradiotherapy, and the prognosis was favorable.

This low incidence has precluded the development of randomized clinical trials on mediastinal seminoma, and present knowledge is based on case reports and very small studies, mostly with sample sizes of 1-16 patients. Furthermore, previous studies included patients with other germ-cell subtypes despite the many distinctive features of seminomas and nonseminomas [3][4][5] and therefore it is di cult to draw de nite conclusions from those studies.
Thus, in this study, we investigated the clinicopathologic characteristics, treatments, and prognosis of patients with primary mediastinal seminomas from a single center.

Patient Selection
This retrospective study was approved by the institutional ethics committee of our institution. We examined the medical reports of patients treated at the National Cancer Center, Beijing, China, between January 2000 and December 2018. We identi ed 30 patients with a pathologic diagnosis of mediastinal seminoma in the database. However, one patient was excluded because of incomplete data and two were excluded for having mixed germ-cell neoplasms. Thus, 27 patients were nally included.

Clinicopathologic Variables
Data regarding patient demographics, symptoms, tumor size, history of smoking and alcohol use, invasion status, treatment protocols, and survival were collected. For all cases, physical examination and chest and abdominal computed tomography (CT) were performed before treatment. Ultrasound of the testicles was also performed in all male patients to rule out gonadal involvement.
The sporadic incidence of primary mediastinal seminomas has also contributed to the preclusion of development of a staging system. To describe the extent of invasion and explore the prognosis of patients with primary mediastinal seminoma, we adopted the Masaoka staging system, which is widely used for another mediastinal tumor, that is, thymic neoplasms [6]. To characterize the invasive sites of the mediastinal seminoma, we reviewed the primary CT scans and described the invasive regions according to the International Association for the Study of Lung Cancer (IASLC) mediastinal lymph node system [7].

Outcomes and Statistical Analyses
Tumor response was initially assessed by a senior radiologist and a radiation oncologist and then con rmed by certain investigators for 1 month after treatment, according to the Response Evaluation Criteria in Solid Tumors (RECIST), version 1.1. Treatment toxicities were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.0. Overall survival (OS) was de ned as the time from diagnosis to death, and progression-free survival (PFS) was de ned as the time from diagnosis to disease progression or death. Local-regional relapse-free survival (LRFS) was de ned as the time from diagnosis to local-regional recurrence, whereas distant metastasis-free survival (DMFS) was de ned as the time from diagnosis to any new distant metastasis. Cancer-speci c survival (CSS) was de ned as the time from diagnosis to cancer-induced death. Survival curves were plotted using the Kaplan-Meier method. Univariate analysis was performed using the log-rank test and included the following variables: Eastern Cooperative Oncology Group performance status score (ECOG PS), sex, age, Masaoka stage, histology, great vessel (aorta, pulmonary artery, pulmonary vein, or brachiocephalic vein) invasion, R0 resection, radiotherapy, and chemotherapy; p<0.05 was considered statistically signi cant. All statistical analyses were conducted using SPSS 23.0 software (IBM Corp., Armonk, NY).

Clinical Characteristics
The median age was 28 (13-63) years. The median maximum primary tumor diameter was 9.9 (3.3-15) cm. The most common symptom was chest pain. Station 3A was the most common site of invasion. Adjacent tissue invasion was also very common. Most patients were diagnosed with Masaoka stage III-IV disease. Details of patient characteristics are listed in Table 1.  (Table 2) revealed high positivity rates for PLAP, OCT3/4, and SALL4. Treatments Surgery, radiation, and chemotherapy were administered in 13, 16, and 25 patients, respectively. For threedimensional conformal radiotherapy or more advanced techniques, the following target volume delineation principles were adhered to. The gross tumor volume (GTV) included the primary tumor and was determined by thoracic CT. The clinical target volume (CTV) included the GTV plus a 5-mm margin and regions of invasion. The planning target volume was created by adding an additional 5-mm margin to the CTV. Among the 25 patients who received chemotherapy, 22 received bleomycin, etoposide, and cisplatin (BEP). Details were listed in table 3. Toxicities Grade 3 toxicities were observed in four patients (14.8%) ( Table 4). Meanwhile, grade 3 hematological and nonhematological toxicities were observed in four patients (14.8%) and 2 patients (7.4%), respectively. Anemia RP, radiation-induced pneumonitis.

Prognostic Factors
As none of the patients died from seminoma at the last follow-up, univariate analysis of OS and CSS was not performed. On univariate analysis, patients with superior vena cava syndrome (SVCS) showed a better PFS. Sex was also associated with PFS. Furthermore, patients with distant metastasis at rst diagnosis were more likely to have a new distant metastasis than those without distant metastasis at rst diagnosis ( Fig. 2A-C).

Discussion
Mediastinal seminomas are di cult to depict because of their rarity. In this study, we investigated a relatively large number of patients with primary mediastinal seminomas. Seminomas are different from other common thoracic malignancies [8,9] in that neither smoking nor alcohol use is a high-risk factor for the disease. Indeed, 81.5% of the patients in our study were nonsmokers and 96.3% reported no history of alcohol use.
Seminomas mostly occur in men, usually young patients. In our study, the median age was 28 years, which is consistent with that in previous reports (28-34 years) [3,10]. Due to their slow growth, most seminomas are bulky when diagnosed. The median maximum diameter of the primary tumor (9.9 cm) is consistent with previous ndings (8-12 cm) [3,10,11]. The tumor may extend to the mediastinum, leading to compression of adjacent structures and invasion, especially into the great vessels in the mediastinum, such as the SVC and aorta. In our study, 37% of patients were found to have SVCS, which is consistent with the ndings of previous reports (10-57%) [3,4,10,11]. However, there was no mention of invasion into the aorta in these previous studies, which might cause di culties in R0 resection.
In this study, the perivascular station was the most common invasion site (100%), followed by the paraaortic station (70.4%) and subaortic station (51.9%). These results are similar to those of previous studies, which found that mediastinal seminomas are usually located in the anterior mediastinum and in front of the aorta [11].The other common invasion sites were the bilateral lower paratracheal station (40.7%,40.7%) and bilateral upper paratracheal station (40.7%, 40.7%).
The results regarding the IHC characteristics of patients varied. The positivity rate for PLAP was 70.7% in a previous study [12]. In our study, all 18 patients who underwent the PLAP test were PLAP-positive. The positivity rates of OCT3/4, SALL4, and CD117 were also high. This suggests that PLAP could be the most remarkable marker for mediastinal seminoma.
Previous studies have reported elevated β-hCG levels in 0-85.7% of patients with primary mediastinal seminoma [3][4][5]10]. In our study, 51.8% of patients showed increased β-hCG levels, which is similar to the ndings of a previous study [11]. Such elevated levels might be attributed to tumor enlargement. Meanwhile, serum LDH was not a typical marker of the disease, which is consistent with previous results [3,11].
There is no established staging system for mediastinal seminomas, and the testicular seminoma staging system cannot be used either. However, mediastinal seminomas seem to share some homogeneous characteristics with thymic neoplasms. Both are prevascular tumors and occur in the anterior mediastinum, both are with rare lymph node metastasis and both are associated with a good prognosis.
On the basis of these common aspects, we adopted the Masaoka staging system, which is widely used for thymic neoplasms, to evaluate the status of mediastinal seminomas [6,11]. We found that 88.9% of patients were diagnosed as having Masaoka stage III-IV disease. Lymph node metastasis and distant metastasis, on the other hand, are not as common as great vessel invasion. A previous study found that lymph node metastasis occurred in 2.6-38% of patients [10]. Although these ndings indicate no signi cant differences in the prognosis of patients with different Masaoka stages, we found a trend that patients with stage I-II disease exhibited higher DMFS and PFS.
In the univariate analysis, patients with SVCS demonstrated better PFS, which contradicted our current notion that tumor invasion into the great vessels is a worse prognostic factor. Possible reasons are: rst, the occurrence of SVCS symptoms led the patient to seek medical consultation, resulting in an accurate diagnosis; second, even though great vessel invasion makes R0 resection di cult, the disease is sensitive to chemoradiotherapy.
Various treatments for mediastinal seminoma aim for complete cure rather than just symptom relief.
Generally, most patients undergo chemotherapy and radiotherapy and receive BEP, as do patients with testicular seminoma. R0 resection is di cult to perform because of tumor invasion into adjacent mediastinal structures, with only 12.5% of patients undergoing such procedure in previous studies [10]. In our study, 33.3% of patients underwent R0 resection, showing good prognosis postoperatively. The postoperative disease control rate was reported as 90-100% in Chinese patients [12]. However, these ndings indicate that R0 resection might not be associated with survival in mediastinal seminoma, unlike in testicular seminoma, probably because of a favorable prognosis and sensitivity to chemoradiotherapy. Similar to testicular seminoma, mediastinal seminoma also demonstrates a high sensitivity to chemotherapy and radiotherapy. In this study, 92.6% and 59.3% of patients received chemotherapy and radiotherapy, with response rates of 85.0% and 100%, respectively, compared to those (100% and 90%, respectively) reported previously [12]. Furthermore, response rates for surgery alone, radiation alone, and chemotherapy alone were reported as 90.0%, 80.0%, and 83.3%, respectively [12]. Thus, the disease is sensitive to all three traditional treatments. The other response rates reported were: surgery+radiation, 91.7%; surgery+chemoradiotherapy, 90.0%; chemoradiotherapy, 100%; and surgery+chemotherapy, 100%. 12 Our study also used different modes of combined therapies, and all of them achieved favorable results.
Unlike routine chemotherapy regimens, radiation is delivered in different doses (25.2-56 Gy). In one study, the patients received 2 Gy × 30 fractions [11]. Comparing this nding with our nding revealed no signi cant difference in either survival or response rate, which may be due to high chemoradiosensitivities. Furthermore, a dose of 45 Gy might be a reasonable choice when considering the patient's quality of life as well as reducing the toxicities in long term.
In this study, the toxicities were tolerable. Hair loss was the most common toxicity probably because of the use of VP-16. Due to the special location of this disease and the delivery of bleomycin, we monitored for radiation-induced pneumonitis (RP). Only ve patients were diagnosed with grade 1 RP, and no severe RP was observed because of both the reasonable radiation dose and the utilization of modern radiation techniques. Although there have been no cardiac-related adverse events documented, a long-term followup for cardiac toxicities is necessary because the heart is one of the adjacent organs.
In 2015, our institution carried out a retrospective study to investigate the clinical characteristics and outcomes of patients with primary malignant mediastinal non-seminomatous germ-cell tumor (MMNSGCT). Compared with that study, our study achieved better OS (100% vs. 49.2%) and PFS (100% vs. 32.8%) [13]. The result of the comparison is also consistent with that of previous reports [3,14,15]. A series of small combined studies also compared the OS of the two different types of mediastinal germcell carcinoma. The studies showed that patients with seminomas achieved a better 5-year OS than those with nonseminomas (87.0%-100% vs. 36.7%-83.0%), although not all the studies showed statistical signi cance due to limited sample size.
In previous small-scale studies and case reports, the 5-and 10-year OS of patients with primary mediastinal seminomas ranged from 87% to 100% and from 75% to 100%, respectively [2][3][4]10,11,15]. One study also showed a 5-year LRFS of 82.1%. These ndings are consistent with our ndings. None of the patients in our study died of seminoma at the last follow-up. Despite the cumulative 10-year risk of testicular malignancy of 10.3% after a diagnosis of extragonadal germ-cell tumor [16], no study patient showed testicular invasion or metastasis at the last follow-up. Thus, the prognosis of patients with primary mediastinal seminoma was generally good. Local relapse and distant metastasis were low after treatment.
To our knowledge, this is the largest study focusing on mediastinal seminomas. We found an association between PFS, DMFS, and Masaoka stage. Most of the patients in this study underwent modern radiation methods and could represent modern real-world data. Our ndings could provide a basis for future treatment delivery in patients with primary mediastinal seminomas.
The study has a few limitations. First, different treatment regimens comprising various therapeutic agents were used with no de nite guidelines. Second, certain components of the IHC test were not possible in some samples because of the long investigation period and deterioration in storage conditions. Finally, we did not have enough time to evaluate late toxicities due to the limited follow-up period.

Conclusion
In conclusion, this study revealed that mediastinal seminomas were frequently diagnosed as large tumors, were in the anterior mediastinum and prevascular region, and always invaded the great vessels.
We also found that primary mediastinal seminomas were sensitive to chemoradiotherapy and patients with the disease could achieve a good prognosis, with moderate toxicities. Moreover, the treatment outcomes of seminomas were like its gonadal counterpart. Future studies with longer follow-up periods are required to further assess late toxicities from treatments.

Consent for publication
Written informed consent for publication was obtained from all participants.

Availability of data and materials
The data and materials were available by sending email to the corresponding author.

Competing interests
The authors declare that they have no competing interests. Masaoka stage and local regional relapse-free (A), distant metastasis-free (B), and progression-free (C) survival