Clinical Evaluation of T Component in the Ninth Edition TNM Classification of Thymic Epithelial Tumors

ABSTRACT

Objectives

The optimal staging system for thymic epithelial tumors (TETs) remains controversial. This study aimed to evaluate the clinical utility of the T component in the latest ninth edition of the TNM staging system.

Methods

A retrospective analysis was performed on patients diagnosed with TETs at our center from January 2001 to December 2022. Recurrence‐free survival (RFS) and overall survival (OS) were analyzed using Kaplan–Meier curves and Cox proportional hazards regression models. The predictive performance of the TNM staging systems was mainly assessed by the concordance index (C‐index) and area under the receiver operating characteristic curve (AUC).

Results

Ultimately, 545 patients with TETs were included. The ninth edition TNM staging system demonstrated superior discrimination compared to the eighth edition TNM staging in RFS outcomes, particularly for T1b versus T2 patients (HR = 3.58, p = 0.014). Prognostic accuracy metrics also favored the ninth edition, with higher AUC values (RFS: 0.83 vs. 0.824; OS: 0.726 vs. 0.685) and C‐index values (RFS: 0.823 vs. 0.816; OS: 0.743 vs. 0.685).

Conclusions

The ninth edition TNM staging system provides enhanced prognostic accuracy for RFS in patients with TETs compared to the eighth edition, although both editions have limitations in predicting OS. Further studies are warranted to refine staging approaches and improve patient management in TETs.

The ninth edition TNM system shows improved prognostic accuracy for RFS in thymic epithelial tumors, particularly in T1b versus T1a patients. The ninth edition outperforms the eighth in model fit and predictive power for both RFS and OS.

Abbreviations

CI

confidence interval

HR

hazard ratio

IASLC

International Association for the Study of Lung Cancer

MG

myasthenia gravis

OS

overall survival

RFS

recurrence‐free survival

TETs

thymic epithelial tumors

1. Introduction

Thymic epithelial tumors (TETs) are rare anterior mediastinal neoplasms characterized by high heterogeneity and a favorable prognosis, with some advanced cases also achieving long‐term survival [1]. The widely adopted Masaoka‐Koga staging system originated from a retrospective study with fewer than 100 cases in a single center. Despite its strong correlation with thymic tumor prognosis, it lacked the granularity needed to distinguish patients effectively [2, 3].

The eighth edition TNM staging system for TET was first introduced in 2014 by the International Association for the Study of Lung Cancer (IASLC) [4]. Despite validation with large datasets, several issues remain: (1) mediastinal pleural invasion is difficult to assess clinically [5]; (2) tumor diameter, a common factor in other tumor types, is not included; however, subsequent studies have confirmed the prognostic significance of tumor size in TETs [6, 7, 8, 9]; (3) lung and phrenic nerve invasions can be surgically resected to achieve R0 status and should not be equated with the invasion of other tissues and organs. In 2024, the IASLC and the International Thymic Malignancy Interest Group proposed the ninth edition of the TNM staging system for TETs, introducing significant changes to the T stage: (1) tumor size was included in the T1 stage with a cutoff value of 5 cm; (2) invasion of the lung and phrenic nerve was downgraded to T2 [5, 10].

To clinically evaluate changes in T category according to the ninth edition of the TNM staging system, we retrospectively reviewed patients with TETs who underwent surgery at our hospital over the last two decades.

2. Patients and Methods

2.1. Ethical Approval and Informed Consent

The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board (2024PHB451‐001), which also waived the requirement for informed consent.

2.2. Study Design and Data Collection

We retrospectively collected medical records of patients diagnosed with TET at our center from January 2001 to December 2022. Patients who underwent non‐R0 resection, were at T4 stage, had lymph node or distant metastasis, or had a pathological diagnosis of neuroendocrine thymic tumors were excluded. Rare thymoma subtypes and unclassifiable types were grouped as “other.” The surgical approaches comprised thoracotomy (median or lateral) and thoracoscopic thymectomy, with extended thymectomy performed for patients with immune‐related diseases like myasthenia gravis (MG). Adjuvant radiotherapy was administered as indicated by multidisciplinary discussions, guided by pathological findings. The histopathological features of the tumors were reviewed by expert pathologists to provide the diagnosis and assess the tumor grade.

2.3. Follow‐Up and Study Endpoints

Patients with TETs would undergo a chest‐enhanced CT scan every 6 months during the first two postoperative years, and then annually thereafter. The follow‐up of patients was primarily conducted via telephone, encompassing details such as postoperative adjuvant therapy methods, recurrence status and timing, locations of recurrence, and the time and cause of death. The follow‐up period started from the date of surgery and continued until death, the last contact, or May 2024. In thymic malignancies, recurrence and survival rates are not closely correlated, particularly in thymoma. Recurrence‐free survival (RFS), defined as the absence of recurrence, is the most appropriate endpoint for patients undergoing R0 resection at an early stage. In contrast, overall survival (OS), measured as death from any cause, serves as the optimal outcome measure for advanced tumors. Both RFS and OS are recognized as effective indicators for evaluating survival rates in patients with TETs. RFS was defined as the time from thymic tumor resection to the first recurrence, and OS was defined as the time from surgery to death.

2.4. Statistical Analyses

RFS and OS were analyzed using the Kaplan–Meier method. The log rank test was adopted to calculate the unadjusted p value. Univariate and multivariate Cox proportional hazards regression analyses were used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs) for OS and RFS. Values of p < 0.05 were considered statistically significant. The predictive performance of the eighth and ninth editions of the TNM staging system was evaluated using various metrics, including the concordance index (C‐index), Akaike information criterion (AIC), Bayesian information criterion (BIC), and the area under the receiver operating characteristic (ROC) curve (AUC). The robustness of the C‐index was determined through the standard error (SE). The “compareC” package was utilized to compute both the C‐index and its SE. Data analysis was conducted with R version 4.3.1 and GraphPad Prism 9.5.

3. Results

3.1. The Characteristics of the Enrolled Patients

A total of 594 patients with T1‐T3N0M0 TETs who underwent R0 resection were included (Figure 1). Forty‐nine cases were lost to follow‐up, and 545 patients were ultimately enrolled, with a follow‐up success rate of 91.8% and 70 months (range 34–105 months) for the median follow‐up time. The median age of patients was 52 years, with a nearly equal gender distribution (50.1% male and 49.9% female). According to the WHO histological classification, type AB was the most frequent (35%), and 48 patients were diagnosed as thymic carcinoma. Postoperative adjuvant therapy was primarily radiotherapy, used in nearly one‐fifth of the cases. MG was observed in 63 patients. The main characteristics of the patients were summarized in Table 1.

FIGURE 1.

Flowchart of patient inclusion and exclusion criteria.

TABLE 1.

Clinical characteristics of the enrolled patients.

Characteristic	N = 545
Age, median (range)	52 (41–61)
Sex, N (%)
Male	273 (50.1%)
Female	272 (49.9%)
Tumor size, N (%)
> 5 cm	179 (32.8%)
≤ 5 cm	366 (67.2%)
WHO classification, N (%)
Thymoma
A	49 (9.0%)
AB	191 (35.0%)
B1	102 (18.7%)
B2	111 (20.4%)
B3	25 (4.6%)
Others	19 (3.5%)
Thymic carcinoma	48 (8.8%)
Adjuvant therapy, N (%)
Radiotherapy	103 (18.9%)
Chemotherapy	11 (2.0%)
Radiochemotherapy	5 (0.9%)
Myasthenia gravis, N (%)	63 (11.6%)
Eighth edition TNM, N (%)
T1a	470 (86.2%)
T1b	16 (2.9%)
T2	10 (1.8%)
T3	49 (9.0%)
Ninth edition TNM, N (%)
T1a	335 (61.4%)
T1b	151 (27.7%)
T2	39 (7.2%)
T3	20 (3.7%)

After re‐staging based on the ninth edition, the number of T1b patients increased from 16 to 151. More than half of the patients originally staged as T3 in the eighth edition were downstaged to T2 in the ninth edition. The distribution of the TNM staging system according to the eighth and ninth editions was illustrated in Figure S1. Type AB and B1 thymomas were the predominant histological types in the T1a stage. Type A thymomas were rarely identified in T2 or higher stages. In the T3 stage, thymic carcinoma accounted for most of the histological diagnoses. Of note, a substantial proportion of patients had type AB thymoma in the T1b stage of the ninth edition. The correlation between the T component and the WHO histological classification for TETs was provided in Table S1.

3.2. Comparison of Survival Curves Between the Eighth and Ninth Edition Staging Systems

We compared survival curves between the eighth and ninth editions of the TNM classification and analyzed differences based on T‐stage modifications (Figure 2). In the eighth edition, 5‐year RFS rates for T1a, T1b, T2, and T3 stages were 97%, 93%, 75%, and 79% respectively. The ninth edition reported 5‐year RFS rates of 98%, 95%, 87%, and 65% for these stages. Notably, T1b patients in the ninth edition exhibited significantly worse RFS than T1a patients (p = 0.0085, Figure S2), contrasting with the eighth edition, where no significant difference was observed (p = 0.12). Kaplan–Meier analysis revealed that RFS rates were comparable between lung/phrenic nerve invasion and pericardial invasion, both of which had better RFS than other T3 subtypes (Figure S3). Overall, the ninth edition demonstrated improved visual discrimination of RFS outcomes.

FIGURE 2.

Kaplan–Meier curves for (A) recurrence‐free survival and (B) overall survival of T stage based on eighth edition staging system; (C) recurrence‐free survival and (D) overall survival of T stage based on the ninth edition staging system.

For OS, the eighth edition reported 10‐year rates of 95%, 82%, 68%, and 80% for T1a, T1b, T2, and T3 stages, respectively, while the ninth edition recorded rates of 96%, 92%, 70%, and 100%. Neither edition revealed significant OS differences between T1a and T1b patients. Kaplan–Meier analysis confirmed comparable OS rates between lung invasion and pericardial invasion. Overall, both the eighth and ninth editions exhibited limited capacity to discriminate OS outcomes (Figure 2).

3.3. Univariate and Multivariate Analyses Based on the Eighth and Ninth Editions of T Stage

We conducted univariate and multivariate analyses to assess factors affecting patients' RFS and OS. Both T stage and histological classification were associated with increased recurrence risk (p < 0.05). The multivariate analysis indicated that the ninth edition staging system provided more distinct staging distribution differences than the eighth edition, especially in T1b versus T1a comparisons (eighth edition: HR = 2.77, 95% CI = 0.43–10.11, p = 0.183; ninth edition: HR = 3.58, 95% CI = 1.33–10.52, p = 0.014, Table 2).

TABLE 2.

Univariate and multivariate analyses of recurrence‐free survival.

Variable	Recurrence‐free survival
	Univariate analyses			Multivariate analyses
	HR	95% CI	p	HR	95% CI	p	HR	95% CI	p
Sex (female vs. male)	1.00	0.50–1.99	0.994	1.32	0.65–2.69	0.437	1.28	0.63–2.59	0.498
Age (> 50 vs. ≤ 50)	0.57	0.28–1.12	0.112	0.57	0.27–1.15	0.117	0.65	0.31–1.33	0.243
Type
[A‐B1] ref		1.00					1.00
[B2‐B3]	3.53	1.63–7.88	0.002	2.16	0.91–5.20	0.082	2.40	1.02–5.73	0.045
[C]	6.16	2.25–15.77	< 0.001	1.60	0.51–4.81	0.408	1.83	0.54–5.88	0.316
Myasthenia gravis (yes vs. no)	0.78	0.23–2.00	0.646	0.53	0.15–1.48	0.268	0.62	0.16–1.81	0.419
Eighth edition T stage
T1a (ref)		1.00
T1b	2.77	0.43–10.11	0.183	2.49	0.36–10.20	0.261
T2	9.21	1.44–33.06	0.0034	10.50	1.59–40.33	0.003
T3	10.58	5.08–22.16	< 0.001	8.10	3.45–19.22	< 0.001
Ninth edition T stage
T1a (ref)		1.00					1.00
T1b	3.58	1.33–10.52	0.014				3.37	1.23–10.09	0.021
T2	15.02	5.58–44.16	< 0.001				11.6	4.06–35.82	< 0.001
T3	22.72	7.52–70.92	< 0.001				15.98	4.67–56.85	< 0.001
AIC					339.44			334.53
BIC					352.90			348.01
C‐index (SE)					0.816 (0.03)			0.823 (0.035)

Note: Values in bold represent statistically significant results (p < 0.05).

Multivariate analysis revealed no significant associations between gender, MG, histological subtypes, and the OS of patients. The eighth and ninth edition staging systems yielded comparable OS multivariate analysis results, with no significant survival differences between T1b and T2 stages. The T2 stage was associated with significantly reduced OS compared to T1a (p < 0.05, Table 3). In addition, age emerged as an independent prognostic risk factor for OS.

TABLE 3.

Univariate and multivariate analyses of overall survival.

Variable	Overall survival
	Univariate analyses			Multivariate analyses
	HR	95% CI	p	HR	95% CI	p	HR	95% CI	p
Sex (female vs. male)	1.25	0.54–2.97	0.598	1.49	0.63–3.56	0.365	1.42	0.60–3.36	0.420
Age (> 50 vs. ≤ 50)	2.69	1.06–8.19	0.052	2.72	0.98–7.50	0.054	3.06	1.11–8.43	0.031
Type
[A‐B1] ref	1.00			1.00			1.00
[B2‐B3]	1.46	0.54–3.63	0.427	1.47	0.54–3.98	0.452	1.61	0.60–4.31	0.347
[C]	2.45	0.55–7.85	0170	1.47	0.35–6.13	0.593	2.63	0.63–10.93	0.185
Myasthenia gravis (yes vs. no)	0.87	0.20–2.57	0.826	0.76	0.21–2.75	0.678	1.11	0.31–3.95	0.872
Eighth edition T stage
T1a (ref)	1.00
T1b	2.50	0.82–7.61	0.180	2.44	0.53–11.20	0.252
T2	6.19	1.38–27.66	0.017	5.57	1.19–26.16	0.029
Ninth edition T stage
T1a (ref)	1.00						1.00
T1b	1.99	0.73–5.40	0.171				2.52	0.91–6.98	0.076
T2	5.40	1.76–15.63	0.002				4.53	1.47–14.01	0.009
AIC					247.40			241.96
BIC					257.22			251.78
C‐index					0.685 (0.058)			0.743 (0.054)

Note: Values in bold represent statistically significant results (p < 0.05).

3.4. Comparison of Prognostic Accuracy Between the Eighth and Ninth Edition TNM Staging Systems

To evaluate the prognostic efficacy of the eighth and ninth editions of the TNM staging system, ROC curves were constructed, and corresponding AUC values were calculated (Figure 3). For predicting RFS in patients with TETs, the ninth edition TNM staging system demonstrated a higher AUC value (0.83) compared to the eighth edition (0.824). Similarly, the C‐index value of the ninth edition (0.823; SE = 0.035) was higher than that of the eighth edition (0.816; SE = 0.03). Moreover, lower AIC and BIC values in the ninth edition (AIC: 334.53; BIC: 348.01) compared to the eighth edition (AIC: 339.44; BIC: 352.90, Table 2) further indicated a better model fit.

FIGURE 3.

ROC curves comparing the eighth and ninth editions of the TNM staging system for (A) RFS and (B) OS.

For predicting OS, the ninth edition TNM staging system also showed superior performance, with an AUC value of 0.726 compared to 0.685 in the eighth edition. The C‐index value of the ninth edition (0.743; SE = 0.054) was higher than that of the eighth edition (0.685; SE = 0.058). Furthermore, the ninth edition exhibited lower AIC (241.96) and BIC (251.78) values compared to the eighth edition (AIC: 247.40; BIC: 257.22, Table 3), suggesting a better model fit. Overall, the ninth edition TNM staging system exhibited superior performance in predicting both RFS and OS.

4. Discussion

We conducted a clinical validation and evaluation of the T component of the latest ninth edition staging system based on the large thymic tumor cohort from our center. The ninth edition staging system effectively differentiated between patients classified as T1b and those classified as T1a, with T1b patients exhibiting significantly worse RFS and a trend toward poorer OS compared to T1a patients. In addition, the ninth edition demonstrated superior predictive performance for OS and RFS compared to the eighth edition. Comparisons based on the AIC, BIC, and C‐index further confirmed the efficacy of the ninth edition TNM staging system. Moreover, patients with lung and phrenic nerve involvement had RFS and OS outcomes comparable to those with pericardial involvement. The WHO histological classification was strongly correlated with RFS, and age was identified as an independent prognostic factor for OS.

Tumor size effectively reflects tumor burden and is easy to measure and record clinically. Previous studies have extensively examined the impact of tumor size on TETs, including thymomas and thymic carcinomas [6, 9, 11]. Our research obtained similar findings, demonstrating that tumors larger than 5 cm were associated with significantly poorer RFS, with a trend toward improved OS in the T1 stage. A large multicenter clinical study found that tumor size significantly impacted OS (p = 0.003) and disease‐free survival (p < 0.001) in TNM Stage I tumors in the multivariate analysis, but not in advanced tumors (TNM Stages II–IV) [9]. Therefore, subdividing the T1 stage based on tumor size is justified. Moreover, the identification of pleural invasion remains challenging, restricting its application for T1 sub‐staging [10, 12]. After incorporating tumor size into the T1 stage, 161 patients were categorized as T1b in our study. Conversely, only 16 cases showed mediastinal pleural invasion in T1, highlighting the efficacy of tumor size‐based staging in identifying high‐risk recurrence patients and allowing for closer follow‐up and monitoring of these high‐risk individuals.

Advancements in surgical techniques and perioperative care have made complete resection for lung and phrenic nerve invasion more achievable, while significantly reducing the probability of postoperative complications. Patients have experienced better survival benefits, making the downstaging of lung and phrenic nerve invasion to T2 both clinically appropriate and reflective of improved prognosis. The ninth edition of the thymic tumor staging system reclassified lung and phrenic nerve invasions as the T2 stage. Our study indicated that, compared to pericardial invasion, lung and phrenic nerve invasions had similar outcomes in terms of RFS and OS. In addition, lung and phrenic nerve invasions demonstrated a lower recurrence trend compared to the T3 stage. However, the absence of deaths in the T3 stage limited further analysis of lung and phrenic nerve invasions in relation to OS in the T3 stage.

Postoperative radiotherapy is recommended for patients with pT3N0M0 TETs due to its association with improved progression‐free survival and OS [13]. However, for Stage II patients, radiotherapy is not essential [14]. Downstaging of lung and phrenic nerve invasion enables many patients to potentially avoid this treatment. Complete resection is a robust prognostic factor for survival in TETs and is closely associated with tumor stage. Other studies have also highlighted the importance of R0 resection, which is a significant predictor of improved OS [15]. Neoadjuvant chemotherapy could enhance the rate of complete resection in patients with Masaoka Stages III–IV, resulting in an R0 resection rate exceeding 75% in those who undergo surgery [16, 17]. A more refined TNM staging system may also more effectively identify patients who could benefit from neoadjuvant therapy.

Age independently influenced the prognosis of patients with TETs, with significantly worse OS observed in those over 50 years old in both univariate and multivariate analyses in our research. A previous study has reported a lower cancer‐specific survival in patients over 70 years old [18]. Early studies have also clearly established that histological type is an independent predictor of OS in Masaoka Stages I and II thymomas. Similarly, B2, B3, and C thymomas are associated with higher recurrence rates and tumor‐related mortality compared to A‐B1 thymomas [19, 20]. A recent similar study also yielded a negative correlation between the number of invaded structures (NIS) and the RFS of patients [21]. Interestingly, a large‐scale clinical study highlighted the impact of the NIS on cancer‐specific survival in TETs. Patients with involvement of more than two structures demonstrated significantly worse cancer‐specific survival, with the NIS confirmed as an independent prognostic factor [22].

This study also has several limitations. As a single‐center retrospective cohort study, it is susceptible to selection bias and follow‐up loss. Moreover, most of the study population is at the T1 stage, with a few cases at the T2 or higher stages. In addition, no fatalities were observed among T3 patients, potentially owing to limited follow‐up duration and the small sample size (n = 20). Additional large‐scale prospective clinical studies are necessary to further validate our findings. Considering the favorable prognosis and extended survival time of TETs, ensuring sufficient follow‐up duration is essential. Moreover, this study comprised solely 50 cases of thymic carcinoma and neglected rare neuroendocrine cancers. Subsequent studies may require additional data collection for further validation.

5. Conclusions

Our study confirmed the effectiveness of incorporating tumor size into the T1 stage and downgrading lung and phrenic nerve invasion to T2 in the ninth edition of the TNM staging system, highlighting their prognostic significance. Further evaluation in larger prospective clinical cohorts is warranted to confirm these findings.

Author Contributions

Yulong Wang: conceptualization, investigation, methodology, validation, software, formal analysis, data curation, writing – original draft, writing – review and editing. Wenhan Weng: conceptualization, investigation, methodology, writing – review and editing, writing – original draft, project administration. Xin Wang: writing – review and editing, software, methodology, validation. Zhijian Liang: conceptualization, investigation. Xuedong He: conceptualization, investigation, validation, formal analysis. Jianfeng Li: conceptualization, investigation, methodology. Xiao Li: conceptualization, methodology, validation, software, formal analysis.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1: Supporting Information.

Wang Y., Weng W., Wang X., et al., “Clinical Evaluation of T Component in the Ninth Edition TNM Classification of Thymic Epithelial Tumors,” Thoracic Cancer 16, no. 22 (2025): e70189, 10.1111/1759-7714.70189.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.