Identification of predictors of long‐term survival and prognostic outcomes in thymic squamous cell carcinoma: A real‐world study

Abstract Background Although thymic squamous cell carcinoma (TSCC) is among the most prevalent forms of thymic carcinoma, there are relatively few studies on this tumor type, and its staging, optimal treatment strategies, and relevant prognostic factors remain controversial. Methods The present study analyzed 79 patients diagnosed with TSCC between January 2008 and January 2021. Kaplan–Meier curves and Cox univariate and multivariate regression analyses were used to explore factors associated with overall survival (OS) and progression‐free survival (PFS) in the overall patient cohort and patient subgroups stratified according to the TNM stage. Time‐dependent receiver operating characteristic (ROC) analyses were used to compare the TNM and Masaoka systems as predictors of patient prognosis. Results The 5‐ and 10‐year OS rates in this study were 65.5% and 49.4%, respectively, with corresponding 5‐ and 10‐year PFS rates of 52.3% and 37.9%. Survival outcomes were better for patients with early‐stage disease (p < 0.001) and patients that underwent surgical treatment (p < 0.001). Neither extent of resection (p = 0.820) nor the surgical approach (p = 0.444) influenced patient survival. In individuals with advanced disease, all forms of adjuvant therapy including radiotherapy (p = 0.021), chemotherapy (p = 0.035), and chemoradiation (p = 0.01) significantly improved patient PFS, but only adjuvant chemoradiotherapy improved patient OS (p = 0.035). When predicting the patient survival outcomes, the TNM system was slightly superior to the Masaoka system (area under the ROC curve [AUC] at 5 years: OS, 0.742 vs. 0.723; PFS, 0.846 vs. 0.816). Conclusion TSCC is an orphan malignancy with a poor prognosis. TNM staging may be superior to Masaoka staging as a predictor of TSCC patient prognosis. Surgery is the mainstay of TSCC treatment. Video‐assisted thoracoscopy (VATS) should be considered for selected patients. Multimodal therapy was associated with excellent results for patients with advanced TNM stage, particularly when surgery was accompanied by adjuvant chemoradiation.


| INTRODUCTION
Thymic epithelial tumors (TETs) are rare malignancies with an estimated incidence rate of 1.7-3.2 per million personyears. 1,2 These tumors are stratified into the thymoma and thymic carcinoma subtypes, with the latter typically being more aggressive and correlated with a poorer prognosis. 3,4 Thymic squamous cell carcinoma (TSCC) is the most common subtype of thymic carcinoma, [5][6][7][8] accounting for 73.4%-80% of thymic carcinoma cases. 6,[8][9][10][11][12] Earlier studies have tended to analyze thymic carcinoma cases as a unified whole without considering the different pathological subtypes. This has the potential of inducing bias in the findings due to differences in the characteristics of the subtypes, resulting in misleading study conclusions. 13 We are only aware of three large-scale studies that have examined the optimal treatments and prognostic factors associated with TSCC. 8,12,14 The results of these studies were inconsistent, leading to difficulties in clinical decision-making.
The present study aimed to provide more up-to-date realworld evidence regarding the characteristics and prognosis of TSCC patients, through the analysis of 79 TSCC patients from a single Chinese institution to identify predictors of overall survival (OS) and progression-free survival (PFS).

| Patients and materials
This was a retrospective, single-center, observational study of patients admitted to our hospital with a pathologically confirmed diagnosis of TSCC between January 1, 2008 and January 1, 2021. The study was terminated on August 1, 2021, allowing a follow-up period of over 6 months.
The inclusion criterion was that all patients had our hospital's histopathological diagnosis of TSCC. The exclusion criteria included: (1) the diagnosis was indefinite or made by another hospital, (2) the clinical data were incomplete, and (3) patients were lost to follow-up.
The clinical data for the patients included age, sex, year of diagnosis, concurrent diseases, tumor size, disease stage, treatments, and follow-up status. We divided the year of diagnosis into two groups from the middle to evaluate differences between surgical approaches used at different times. Tumor size was defined as the maximum tumor diameter measured via chest computed tomography (CT) imaging prior to treatment. 15 Direct tumor size measurements were not utilized, as not all patients underwent surgery, and tumor resection was incomplete in certain patients that did undergo surgery. Disease staging was based upon the 8th edition of the AJCC TNM staging system 16 and the Masaoka system. 17 Disease staging was independently confirmed by two thoracic surgeons (Z. MX and C. YY), who reviewed patient medical records and pathological reports. Any discrepancies were resolved through consultation with a third expert (L. SQ).
Subgroup analyses were performed by separating patients that had undergone surgery into early-stage (TNM stage I/ II) and advanced-stage (TNM stage III/IV) groups for a more reliable assessment of the relationship between outcomes and treatment modalities. The factors included the extent of tumor resection (complete vs. incomplete), the selected surgical approach (thoracotomy vs. video-assisted thoracoscopy [VATS]), and postoperative adjuvant therapy (chemotherapy alone, radiotherapy alone, or chemoradiotherapy).
OS was defined as the interval between biopsy or surgery and all-cause death or the most recent follow-up. PFS was defined as the interval between biopsy or surgery and documented disease progression (including tumor recurrence, enlargement, and metastasis) or death without progression. Outpatient records and telephone calls were used to assess patient follow-up status. Those patients alive at the last follow-up were censored.
The study was approved by the institutional review board of our hospital (ZS-3305). All patients provided written informed consent for the research use of their data at the time of hospital admission, and the study was thus exempt from the further need to obtain informed patient consent.

| Statistical analysis
Normally distributed continuous data are presented as means ± SD and were compared via Student's t-tests. In contrast, non-normally distributed continuous data are presented as median (range) and were compared via Mann-Whitney U-tests. For survival analyses, median values were used to transform non-normally distributed continuous variables into dichotomous variables. Categorical with excellent results for patients with advanced TNM stage, particularly when surgery was accompanied by adjuvant chemoradiation.

K E Y W O R D S
chemoradiotherapy, surgery, thymic squamous cell carcinoma, TNM system variables were analyzed using chi-squared tests or Fisher's exact test. Spearman correlation analyses were used to assess relationships between TNM stage and tumor size. The prognostic utility of the TNM and Masaoka staging systems was assessed using the time-dependent receiver operating characteristic (ROC) curves, and the area under the curve (AUC) values at 1, 3, and 5 years were compared.
Kaplan-Meier curves were used to evaluate patient OS and PFS. Prognostic variables were identified through Cox proportional hazards regression models, with variables exhibiting p-value <0.2 in the univariate analyses incorporated into the subsequent multivariate analyses.
R version 3.6.3 (The R Foundation for Statistical Computing) was used for all data analysis. The R packages timeROC, ggplot2, and survminer were used for evaluating survival. Two-tailed p-values <0.05 were considered statistically significant.

| Clinical characteristics
In total, 91 TSCC consecutive patients who were admitted to our hospital over the study period were initially recruited. Twelve of these patients (13.2%) were lost to follow-up. The study thus included the remaining 79 patients ( Figure 1). Their clinical characteristics are summarized in Table 1. The male-to-female ratio was 1.47:1, with an average age of 54.7 ± 10.6 years. The number of cases detected between 2008 and 2014 was comparable to the number of cases detected between 2015 and 2020. The most prevalent symptoms included chest discomfort, such as chest pain and chest tightness, which were experienced by more than 30% of the patients ( Figure 2). Forty patients (50.6%) suffered from concurrent diseases, of which hypertension was the most common, occurring in 21 patients (26.6%). Myasthenia gravis (MG) was only observed in 3 patients (3.8%). The treatment administered, namely, surgery (p = 0.962), chemotherapy (p = 0.721), and radiation (p = 0.678) did not differ significantly between patients with and without concurrent disease. The median tumor size was 4.6 cm, and tumor size was moderately correlated with the patient TNM stage (ρ = 0.521, p < 0.001).
As per the Masaoka system, 64.6% of the included patients had stage III/IV disease, while according to the TNM system 58.2% of patients had advanced-stage disease. The ROC curves for these two staging systems and corresponding 1-, 3-, and 5-year AUC values are shown in Figure 3.
The main surgical indications were as follows: (1) anterior mediastinal mass with suspected diagnosis of malignant tumor; (2) two or more thoracic surgeons had considered that the mass was resectable; (3) whole body examination indicated that there were no multiple systematic metastases; (4) debulking surgery was performed to control symptoms. The surgical approach was decided by the thoracic surgeon in charge. Overall, 67 patients (84.8%) underwent surgical treatment, of whom 58.2% underwent thoracotomy. VATS was the most commonly performed procedure between 2015 and 2020 relative to 2008-2014 (61.8% vs. 21.2%, p < 0.001).
Complete resection was unsuccessful in one patient among those with early-stage disease. The tumor had invaded the pericardium and the final pathological result showed positive microscopic margins in the resected pericardium. Among patients with advanced-stage disease, 12 underwent partial pneumonectomy and 6 vascular reconstructions. Pleural disseminations were observed in five patients during the surgery, resulting in incomplete resection. Debulking surgery aimed at symptom alleviation was performed on five patients. Table 2 summarizes the differences in clinical characteristics between patients with early and advanced disease. Patients with early-stage TSCC were more likely to have smaller tumors and to have undergone surgical treatment via VATS. A higher proportion of early-stage patients treated via surgery were underwent adjuvant radiotherapy, whereas advanced-stage patients were more likely to have undergone chemoradiotherapy. A total of 45 patients (57%) received chemotherapy, of which platinum drugs plus paclitaxel was the most frequently used, being administered in 38 patients (48.1%). Other chemotherapy treatments included cisplatin plus doxorubicin plus cyclophosphamide (n = 3), cisplatin plus doxorubicin plus vincristine plus cyclophosphamide (n = 2), and gemcitabine plus cisplatin (n = 2).
Kaplan-Meier curve analyses revealed that advanced disease was associated with significant reductions in patient OS  For individuals with early-stage disease, adjuvant therapy was not associated with any improvement in OS

| Prognostic factors in TSCC patients
The multivariate analysis showed that advanced TNM stage and not having undergone surgical treatment were associated with worse OS and PFS outcomes in the overall patient cohort (Table 3).
In individuals with early-stage disease, no variables were significantly associated with patient OS or PFS in the univariate or multivariate analyses (Table S1). Only the adjuvant therapy approach was eligible for incorporation into multivariate analysis for patients with advancedstage disease. The extent of resection was also assessed as a covariate. This showed that all forms of adjuvant therapy were independently predictive of longer PFS. In contrast, only adjuvant chemoradiotherapy was associated with increased OS (Table 4).

| DISCUSSION
The present study was undertaken to identify TSCCrelated features and prognostic factors through a singlecenter analysis of 79 patients with this rare disease. Together, these results offer updated real-world evidence that can guide the treatment and evaluation of this uncommon malignancy.
The analyses revealed that the TNM system was slightly superior to the Masaoka system when used to predict TSCC patient prognostic outcomes. Moreover, advanced TNM staging and not having undergone surgical treatment were independently associated with worse patient OS and PFS outcomes. No differences in patient survival outcomes were noted when comparing treatment via VATS or thoracotomy, or patients that underwent incomplete or complete tumor resection. Postoperative adjuvant radiotherapy, chemotherapy, and chemoradiotherapy were correlated with improved PFS in patients with advanced disease, but only chemoradiotherapy was independently associated with improved OS.
The primary characteristics of TSCC patients in our study were that most were male, diagnosed when middleaged, and exhibited an insidious disease onset with rare comorbidity of MG. These characteristics were consistent with those found in other studies on TSCC patients. 8,12,14 The Masaoka system is the most commonly used tool when staging TETs. However, the use of this system in assessing thymic carcinoma remains controversial. 9,12,18 For example, the single-center study noted above 12 found the Masaoka stage to be unrelated to TSCC patient OS or disease-free survival (DFS), and similar results have also been reported in an analysis of thymic carcinoma. 9 These outcomes demonstrated that the Masaoka system may be not appropriate for TSCC. Instead, since the proposal of the TNM system, several studies have found it was at least as effective as the Masaoka system, and also acted as a prognostic factor in thymic carcinoma, indicating its potential for use in TSCC. 19 In the present study, we utilized T A B L E 3 Univariate and multivariate analysis of OS and PFS in the overall patient cohort. time-dependent ROC curves. We discovered that the TNM staging system had a slightly higher AUC value than the Masaoka system. In addition, the multivariate analysis supported the TNM staging system's independent prognostic importance.

Univariate analysis on OS
Our study confirmed that surgical treatment is associated with better patient prognosis, even for individuals with advanced disease. However, data on the relevance of the extent of tumor resection remain somewhat controversial. The present study noted that complete resection was not an independent protective factor, in line with the results of Shen et al. 20 On the contrary, the retrospective analysis of Zhao et al. 12 and the SEER study performed by Yang et al. 8 both found that complete resection was predictive of improved survival outcomes in TSCC patients. Some studies of thymic carcinoma patients have also identified complete resection as an important prognostic factor. 9,10,21-23 Although the reasons for this disparity are unknown, we continue to recommend complete resection as the ultimate surgical goal whenever possible. When total resection is deemed impossible, surgery may still be advised for selected individuals.
The number of thoracic surgical procedures performed using the VATS approach has risen in recent years. Nevertheless, few studies have compared outcomes associated with VATS and thoracotomy procedures in TSCC patients. In the present analysis, there were no differences in OS or PFS between VATS and thoracotomy, and VATS was not linked with increased rates of incomplete resection (p = 0.09). Thus, VATS represents a promising choice for appropriate patients.
There is no consensus on the optimal postoperative adjuvant therapy regimen for treating TSCC. In the present study, we discovered that adjuvant radiotherapy and chemotherapy alone were associated only with extended PFS but not OS. Nevertheless, postoperative chemoradiotherapy was found to be an independent protective factor for both OS and PFS in patients with advanced-stage TSCC. Several studies 5,22,24-26 have observed better outcomes in patients who underwent postoperative radiotherapy, especially those with advanced disease. The effect of postoperative chemotherapy was unclear. The main view is that adjuvant chemotherapy may be of less significance than radiotherapy, which has been verified by some studies. 10,27 However, other studies have found that adjuvant chemotherapy could improve the prognosis of thymic carcinoma, especially in prolonging PFS. 12,28 Besides, Yang et al. 8 in their SEER study indicated that TSCC may be more sensitive to chemotherapy than other histological types. Very few research studies on adjuvant therapeutic regimens have specifically investigated the effects of chemoradiotherapy treatment. Kim et al. 29 found that adjuvant chemoradiotherapy was associated with a better prognosis in patients with Masaoka stage III thymic carcinoma. In conclusion, for advanced-stage TSCC patients, adjuvant therapy should be administered, and chemoradiotherapy, in particular, should be considered in selected patients to improve both OS and PFS.
There are several limitations to this study. For instance, because this was a single-center retrospective investigation, the results are prone to implicit bias. Furthermore, the sample size was small, necessitating further research with additional in-depth subgroup analysis. Additionally, the 12-year study period may have led to some differences in the treatment plans utilized. Besides, due to the unavailability of some data, treatment details, such as the dose and site of radiotherapy, were not recorded and compared.

| CONCLUSIONS
These results suggest that the TNM staging system may be optimal for evaluating TSCC patients. Both advanced TNM stage and not having undergone surgical treatment were herein found to be independently associated with a risk of poorer OS and PFS outcomes. However, neither resection extent nor surgical approach was an independent factor of TSCC patient survival. Therefore, VATS might be a better option for selected patients. After evaluation, surgery should be the first alternative for individuals with advanced-stage TSCC, and postoperative chemoradiotherapy should be given to improve patient outcomes. Further in-depth studies with uniform treatment guidelines will be essential to gain a more detailed understanding regarding the prognosis of patients with this form of cancer.