Oncologic outcomes and predictors of recurrence following minimally invasive and open surgery for thymoma

Abstract

Background

Thymomas are rare neoplasms of the mediastinum with an incidence of 0.15 per 100,000 person-years but represent the most common tumor of the anterior mediastinum in adults. Complete surgical removal is the mainstay treatment for thymomas. While minimally invasive surgeries (MIS) for thymoma are increasing, it remains unknown if MIS is oncologically equivalent to traditional open surgical techniques (OTs). The aim of the present study was to characterize outcomes and identify risk factors associated with recurrent disease in patients undergoing surgery for early-stage thymoma, with specific attention to the influence of surgical approach.

Methods

We conducted a retrospective cohort study using a database of thymic epithelial tumor patients evaluated at Indiana University Melvin and Bren Simon Comprehensive Cancer Center from 2005 to 2021. A total of 178 stage I–III thymoma patients who underwent surgery [OT: n=129 (72%), MIS: n=49 (28%)] were identified and reviewed to determine factors predictive of outcomes.

Results

The median age was 56.2 years and 56% were female. The median follow-up was 64 months. The stage distribution was I 39%, II 44%, and III 16%. Complete resection was achieved in 82% of patients. Multivariable analysis demonstrated that presurgical sampling through image-guided fine needle aspiration (FNA)/core [hazard ratio (HR) =2.738] and thoracoscopic biopsies (HR =2.646), MIS approach (HR =2.481), greater stage (HR =2.951), and incomplete resection (HR =2.569) were associated with inferior disease-free survival (DFS). However, no significant difference was observed in DFS or overall survival (OS) on Kaplan-Meier analysis stratified by surgical approach. Of the 14 patients who developed pleural recurrence after undergoing unilateral MIS, all pleural metastases were ipsilateral to the surgical approach.

Conclusions

Our data demonstrates that both OT and MIS have excellent outcomes with respect to OS for early-stage thymoma. For low surgical risk patients with small anterior mediastinal masses which are suspicious for thymoma, removal using an MIS approach without biopsy appears reasonable. In higher-stage tumors, an MIS approach should be utilized cautiously at centers with surgical expertise.

Highlight box.

Key findings

• In our multivariable model, resection using minimally invasive surgery (MIS) techniques, including robotic-assisted thoracic surgery and video-assisted thoracic surgeries, showed inferior disease-free survival (DFS) compared to open surgical techniques (OTs).

• MIS techniques demonstrated a trend towards a higher number of nodules at the time of initial recurrence and a strong association between the laterality of surgical approach and site of pleural recurrence.

• Preoperative image-guided and thoracoscopic biopsies were associated with inferior DFS.

What is known and what is new?

• MIS techniques are associated with improved perioperative outcomes, but it remains unclear whether MIS is oncologically equivalent to traditional OT. Current National Comprehensive Cancer Network (NCCN) guidelines discourage preoperative biopsy, but this recommendation is based on low evidence levels with only a few retrospective studies addressing this topic.

• This study adds evidence supporting NCCN guidelines against preoperative biopsy and demonstrates potential for inferior DFS with MIS approaches.

What is the implication, and what should change now?

• MIS techniques for resection of thymoma should be utilized with caution, particularly for larger or more invasive tumors. Careful patient selection for an MIS approach is critical, with strict adherence to the “no-touch” technique to minimize the risk of tumor dissemination.

• Preoperative biopsy should be avoided, when possible, as it is associated with inferior DFS.

Introduction

Thymomas are rare neoplasms of the mediastinum with an incidence of 0.15 per 100,000 person-years but represent the most common tumor of the anterior mediastinum in adults (1,2). Complete surgical removal is the mainstay treatment for thymomas (3). Resectability, tumor stage, and World Health Organization (WHO) histotype are the main factors influencing patient outcomes (4-9).

Thymomas have a unique tendency to spread locally to the nearby pleural and pericardial serosal surfaces. As such, the time-honored surgical approach involves minimal tumor manipulation, often referred to as the “no-touch” technique. Open surgical techniques (OTs) have historically been considered the gold standard to accomplish this goal (6). However, minimally invasive surgery (MIS) approaches, including video-assisted thoracic surgery (VATS) and robotic-assisted thoracic surgery (RATS), have become increasingly popular over the past two decades (10-13). Proponents of MIS reference several peri-operative advantages, such as reduced blood loss, shorter length of hospital stay, and decreased postoperative pain scores (10,11,13-19). However, there remains concern that MIS techniques could be associated with greater tumor manipulation resulting in inadvertent tumor seeding (9,10,16,20-30). Comparable overall survival (OS) following OT and MIS approaches has been reported, which is not surprising given the indolent nature of thymomas. However, information about the recurrence rate after MIS continue to be limited by relatively short follow-up (16,26,31-34). In this study, we analyzed our large institutional database to characterize outcomes and identify risk factors associated with recurrent disease in patients undergoing surgery for early-stage thymoma, with specific attention to whether the surgical approach utilized during resection influences oncologic outcomes. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2024-2227/rc).

Methods

Patient recruitment and data collection

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The institutional review board of Indiana University (No. IRB00000219) approved this study and individual consent was waived. This retrospective cohort study utilized an institutional database [Research Electronic Data Capture (REDCap)] of patients evaluated at Indiana University Melvin and Bren Simon Comprehensive Cancer Center (IUSCCC) for thymic epithelial neoplasms. The database was queried to identify a cohort of patients who underwent surgery for Masaoka-Koga stage I, II, or III thymic epithelial tumors between 2005 and 2021. Demographic, clinical, and pathologic information, initially entered into the database by a team of seven researchers, was subsequently verified and supplemented, as needed, through review of electronic medical records and clinical charts by two independent researchers to ensure completeness and accuracy. An initial total of 271 patients were identified who underwent surgery at IUSCCC or an outside institution (OSH) and subsequently presented to IUSCCC for care. Final inclusion criteria were surgery performed with curative intent and pathologically proven thymoma. Patients were excluded if they had <6 months of follow-up after surgery and there was no event during that period, missing records, reoperation for resection of the primary tumor, macroscopically positive margins (R2), thymic carcinoma, or any patient receiving neo-adjuvant therapy. Of the initial cohort, 178 patients met criteria for study inclusion.

Definitions and study endpoints

Open techniques encompassed median sternotomy, thoracotomy, or clamshell approaches, and MIS encompassed VATS or RATS approaches. All VATS and RATS approaches were lateral, with none through a subxiphoid approach. MIS cases with conversion to open were included within the MIS group, consistent with an intent-to-treat approach. At IUSCCC, all surgeons utilize both MIS and OT approaches, however, the specific approach selected was surgeon dependent. Over time, an MIS approach has been increasingly utilized to remove any smaller and well-contained anterior mediastinal mass suspected to represent an early stage thymoma without biopsy, consistent with current National Comprehensive Cancer Network (NCCN) guidelines (35).

The reported pathologic stages and cellular classifications were based on the Masaoka-Koga (36) and WHO systems (37), respectively. If the histologic type was mixed, the preponderant histology was used for categorization; if relative proportions of the types were not noted in the pathology report, the worst histology was used. Tumor size was recorded from the pathologic report and was analyzed as a continuous variable. Incomplete resection was defined as microscopically positive margins (R1), while complete resection was defined as microscopically negative margins (R0).

Postoperative surveillance during follow-up at the IUSCCC included chest computed tomography (CT) every 4–6 months for the first 2 years, then every 6 months until year 5, and annually thereafter until 10 years when follow-up was performed every 2 years. Surveillance schedules while patients were followed at an OSH varied. Local recurrence was defined as disease within the anterior mediastinum or the thymus surgical bed. Regional recurrence was defined as disease involving the pleura or pericardium. Distant recurrence was defined as disease involving any extrathoracic or intrapulmonary site. The date of recurrence was defined as the date of initial radiologic evidence of recurrence on CT scan and was classified according to the aforementioned definitions and laterality, as appropriate. Disease-free survival (DFS) was calculated from the date of surgery to the date of thymoma recurrence, death, or last follow-up. OS was calculated from the date of surgery to either the date of death or the last follow-up date.

Statistical analysis

Categorical variables were expressed as number (%) and continuous variables were expressed as median [interquartile range (IQR)] as the variables did not follow a normal distribution. Continuous variables were compared using the Wilcoxon Rank-Sum test and categorical variables were compared with either the Chi-square test or Fisher exact test. Survival was plotted using the Kaplan-Meier method and compared by surgical approach via a Kolmogorov-Smirnov-type test. P value for the latter test was computed using 1,000 bootstrap samples. To control for confounding in the survival analyses, inverse probability of treatment weighting (IPTW) was implemented. Propensity scores were calculated using a logistic regression model with the variables Charleston-Comorbidity Index (CCI), biopsy prior to surgery, location of surgery, tumor stage, and tumor size. Multivariable Cox regression analysis was performed to determine independent predictors of DFS and OS. Variables included in the models were selected based on their identification as potential risk factors in previous studies and/or their potential influence on the selection of surgical approach. The variables included were assessed for multicollinearity by calculating a condition index for the model and variance inflation factors for each variable, which demonstrated no significant multicollinearity. Sensitivity analysis was conducted through exploratory analyses that adjusted for additional variables, including age, sex, and body mass index (BMI), to confirm the robustness of significant predictors. Statistical analyses were performed in RStudio version 2023.09.1+494 software (R Foundation for Statistical Computing, Vienna, Austria). Figure 1 was created with BioRender.com. P values of <0.05 were considered statistically significant.

Figure 1.

Flow of patient selection. MIS, minimally invasive surgery; OT, open surgical technique; RATS, robotic-assisted thoracic surgery; VATS, video-assisted thoracic surgery.

Results

Flow of patient selection

The flow of the 178 patients selected for review and patients excluded is illustrated in Figure 1.

Patient demographics and clinical characteristics

The cohort’s clinical characteristics and demographics are listed in Table 1. Female patients were slightly overrepresented, and the median age at diagnosis was 56.2 (IQR, 45.4–65.4) years. Forty-seven (26%) patients had myasthenia gravis. Most patients were expected to have >50% 10-year survival based on their CCI. Sixty-six (37%) patients had a biopsy prior to surgery. Imaging-guided FNA/core biopsies were more frequent than thoracoscopic biopsies. WHO types B1 and B2 were the most common histology. The median tumor size was 6.1 (IQR, 4.0–8.3) cm. The distribution of stages was I 39% (n=70), IIA 31% (n=56), IIB 13% (n=23), and III 16% (n=29). Surgical approaches included MIS: VATS 13% (n=24), RATS 13% (n=23), and conversion to open after initial VATS approach 1% (n=2), and OT: sternotomy 57% (n=102), anterolateral thoracotomy 13% (n=24), and clamshell 2% (n=3). There were 32 (18%) cases of incomplete resection, and 39 (22%) patients received adjuvant therapy with radiation therapy being most common. The median follow-up was 64.0 (IQR, 26.3–112.8) months. At the time of last follow-up, 30 patients had died from all causes and 48 patients were alive with disease.

Table 1. Cohort clinicopathologic characteristics and bivariate analysis stratified by surgical approach.

Characteristics	Overall (n=178)	OT (n=129)	MIS (n=49)	P value
Age (years)	56.2 (45.4–65.4)	54.6 (44.0–64.7)	60.3 (48.1–65.7)	0.12
Sex				0.11
Male	79 [44]	52 [40]	27 [55]
Female	99 [56]	77 [60]	22 [45]
BMI				0.52
Non-obese	93 [52]	65 [50]	28 [57]
Obese	85 [48]	64 [50]	21 [43]
CCI				0.06
>50% est. 10-year survival	129 [72]	99 [77]	30 [61]
≤50% est. 10-year survival	49 [28]	30 [23]	19 [39]
Smoking status				0.82
Never	110 [62]	81 [63]	29 [59]
Former	51 [29]	35 [27]	16 [33]
Current	17 [10]	13 [10]	4 [8]
Biopsy prior to surgery				0.18
None	112 [63]	76 [59]	36 [73]
FNA/core	45 [25]	37 [29]	8 [16]
Surgical	21 [12]	16 [12]	5 [10]
Myasthenia gravis	47 [26]	34 [26]	13 [27]	>0.99
Location of surgery				0.03*
IUSCCC	94 [53]	75 [58]	19 [39]
OSH (referred)	84 [47]	54 [42]	30 [61]
Masaoka-Koga stage				0.13
I	70 [39]	51 [40]	19 [39]
IIA	56 [31]	40 [31]	16 [33]
IIB	23 [13]	13 [10]	10 [20]
III	29 [16]	25 [19]	4 [8]
WHO classification				0.77
A/AB/micronodular	68 [38]	48 [37]	20 [41]
B1/B2	73 [41]	55 [43]	18 [37]
B3	37 [21]	26 [20]	11 [22]
Resection margins				0.76
Complete (R0)	146 [82]	107 [83]	39 [80]
Incomplete (R1)	32 [18]	22 [17]	10 [20]
pT size (cm)	6.1 (4.0–8.3)	6.4 (4.0–9.0)	5.0 (4.0–7.5)	0.003**
Adjuvant therapy				0.76
No	139 [78]	102 [79]	37 [76]
Yes	39 [22]	27 [21]	12 [24]
XRT	29 [16]	18 [14]	11 [22]
Chemotx	2 [1]	1 [1]	1 [2]
Both	8 [4]	8 [6]	–
Follow-up (months)	64.0 (26.3–112.8)	65.0 (23.0–116.0)	57.0 (34.0–100.0)	0.59
Surgery era				0.48
2005–2010	40 [22]	32 [25]	8 [16]
2011–2015	74 [42]	52 [40]	22 [45]
2016–2021	64 [36]	45 [35]	19 [39]

Values are presented as n [%] or median (IQR). *, P<0.05; **, P<0.01. BMI, body mass index; CCI, Charleston Comorbidity Index; Chemotx, chemotherapy; est., estimated; FNA, fine needle aspiration; IQR, interquartile range; IUSCCC, Indiana University Simon Comprehensive Cancer Center; MIS, minimally invasive surgery; OSH, outside hospital; OT, open surgical technique; pT, pathologic tumor; WHO, World Health Organization; XRT, radiation.

The characteristics of the cohort stratified by surgical approach, are also detailed in Table 1. The clinicopathologic characteristics of patients stratified by site of initial surgery and the status of patients at time of IUSCCC referral for the OSH group are given in Tables S1,S2, respectively. Open surgery was the method more commonly employed at IUSCCC and in those referred from OSH, though a greater proportion of the MIS group consisted of OSH patients (P=0.03). There was a trend toward stage III tumors being more frequently managed with OT, whereas other stages were treated with both methods equally (P=0.13). Additionally, tumors treated with OT were significantly larger than those treated with MIS (P=0.003). A similar proportion of patients in each group underwent adjuvant therapy. BMI, smoking status, myasthenia gravis, histology, resection margins, surgery era, and follow-up duration were also similar for both groups. Over the three surgical eras cited in Table 1, there was no significant difference in rates of recurrence for MIS patients (P=0.93).

Patterns of recurrence

The patterns of recurrence stratified by surgical approach are demonstrated in Table 2. Overall, 50 patients (28%) developed recurrent disease. Of these, 48 (96%) patients developed locoregional recurrence, with 35 (70%), 18 (36%), and 1 (2%) developing recurrence in the pleura, mediastinum, and pericardium respectively. Only 3 (6%) patients recurred in distant locations.

Table 2. Comparison of recurrence patterns between OT and MIS approaches.

Characteristics	OT (n=31)	MIS (n=19)	P value
Recurrence locations
Locoregional	30 [97]	18 [95]	>0.99
Pleura (ipsilateral to approach)	19 [61]/2 [67]†	16 [84]/15‡ [94]	0.12
Mediastinum	12 [39]	6 [32]	0.76
Pericardial	1 [3]	0	>0.99
Distant	2 [6]	1 [5]	>0.99
Number of recurrence nodules	2.0 (1.0–3.5)	3.0 (1.0–5.0)	0.07
Time to recurrence (months)	37.3 (22.5–61.3)	27.8 (17.9–37.6)	0.24

Values are presented as n [%] or median (IQR), as indicated. Percent is calculated as the proportion of patients with recurrence who had recurrence in this location, as some patients had recurrence detected in more than one location at the time of the initial recurrence event, percent does not equate to 100. ^†, percentage with recurrence ipsilateral to the surgical approach for the OT cohort was calculated based on number of patients who underwent lateral thoracotomy approach (n=3); ^‡, 15 patients developed ipsilateral pleural recurrence, 14 of whom underwent unilateral approach with unilateral recurrence ipsilateral to the site and 1 of whom underwent bilateral approach and subsequently developed bilateral pleural recurrence. IQR, interquartile range; MIS, minimally invasive surgery; OT, open surgical technique.

There was a non-significant trend (P=0.12) toward pleural recurrence being more common in the MIS group than the OT group. The incidence of mediastinal, pericardial, and distant recurrences was nearly identical between the two groups. Importantly, recurrent disease was ipsilateral to the surgical approach in all 14 patients who underwent a unilateral MIS and subsequently developed pleural recurrence. Of the two MIS patients who underwent a bilateral approach with subsequent pleural recurrence, bilateral recurrence was seen in one, while the other developed pleural recurrence ipsilateral to the side through which the tumor was removed. In the open anterolateral thoracotomy patients, the trend of pleural recurrence ipsilateral to the surgical approach was seen in two of the three developing pleural recurrence, whereas the other developed bilateral pleural recurrence. Stratifying by MIS approach, there were 9 recurrences after RATS and 10 recurrences after VATS (P>0.99).

The median time to recurrence was 33.2 (IQR, 21.0–55.1) months and there was no significant difference in the time of recurrence by surgical approach (P=0.24). The median number of nodules at first recurrence for the entire cohort was 3.0 (IQR, 1.0–4.0). The MIS group had a non-significant trend toward a higher number of nodules at the time of the initial recurrence event (P=0.07).

Cox regression analysis of predictors of DFS and OS

Tables 3,4 show the results of the multivariable Cox regression analysis for DFS (Table 3) and OS (Table 4). Significant risk factors for DFS were image-guided FNA/core (P=0.002) and thoracoscopic biopsy (P=0.02) prior to surgery, MIS surgical approach (P=0.005), stage III disease (P=0.007), and incomplete resection (P=0.006). The Cox regression for OS demonstrated that a significant factor negatively influencing OS was a higher Charleston Comorbidity Index, reflected as the associated survival probability (P=0.01), whereas stage IIA disease was a protective factor (P=0.041). Factors that had non-significant trends for DFS included tumor size (P=0.09) and stage IIB disease (P=0.17). Univariate analysis is provided in Table S3. Sensitivity analyses were performed with additional granular demographic (Table S4A) and with a subset cohort of only patients with complete (R0) resection (Table S4B) confirming the significant predictors.

Table 3. Cox regression multivariable analysis of predictors of disease-free survival (concordance =0.761).

Variables	HR (95% CI)	P value
CCI
>50% est. 10-year survival	Reference
≤50% est. 10-year survival	1.449 (0.77–2.71)	0.25
Biopsy prior to surgery
None	Reference
FNA/core	2.738 (1.45–5.18)	0.002**
Surgical	2.646 (1.19–5.89)	0.02*
Location of surgery
IUSCCC	Reference
OSH	1.382 (0.71–2.69)	0.34
Surgical approach
OT	Reference
MIS	2.481 (1.31–4.71)	0.005**
Masaoka-Koga stage
I	Reference
IIA	0.753 (0.37–1.54)	0.44
IIB	1.788 (0.78–4.12)	0.17
III	2.951 (1.35–6.47)	0.007**
WHO histology type
A/AB/micronodular	Reference
B1/B2	1.359 (0.73–2.54)	0.34
B3	1.244 (0.62–2.50)	0.54
pT size	1.085 (0.99–1.19)	0.09
Final resection margins
Complete (R0)	Reference
Incomplete (R1)	2.569 (1.29–4.71)	0.006**

*, P<0.05; **, P<0.01. CCI, Charleston Comorbidity Index; CI, confidence interval; est., estimated; FNA, fine needle aspiration; HR, hazard ratio; IUSCCC, Indiana University Melvin and Bren Simon Comprehensive Cancer Center; MIS, minimally invasive surgery; OSH, outside institution; OT, open surgical technique; pT, pathologic tumor; WHO, World Health Organization.

Table 4. Cox regression multivariable analysis of predictors of overall survival (concordance =0.747).

Variables	HR (95% CI)	P value
CCI
>50% est. 10-year survival	Reference
≤50% est. 10-year survival	3.265 (1.32–8.06)	0.01*
Biopsy prior to surgery
None	Reference
FNA/core	1.583 (0.64–3.89)	0.32
Surgical	0.646 (0.14–2.88)	0.57
Location of surgery
IUSCCC	Reference
OSH	0.742 (0.32–2.38)	0.79
Surgical approach
OT	Reference
MIS	0.742 (0.27–2.02)	0.56
Masaoka-Koga stage
I	Reference
IIA	0.287 (0.09–0.95)	0.041*
IIB	1.268 (0.38–4.20)	0.70
III	1.007 (0.31–3.27)	>0.99
WHO histology type
A/AB/micronodular	Reference
B1/B2	1.765 (0.69–4.50)	0.24
B3	1.136 (0.28–3.83)	0.84
pT size	1.052 (0.94–1.18)	0.39
Final resection margins
Complete (R0)	Reference
Incomplete (R1)	1.504 (0.50–4.49)	0.47

*, P<0.05. CCI, Charleston Comorbidity Index; CI, confidence interval; est., estimated; FNA, fine needle aspiration; HR, hazard ratio; IUSCCC, Indiana University Melvin and Bren Simon Comprehensive Cancer Center; MIS, minimally invasive surgery; OSH, outside institution; OT, open surgical technique; pT, pathologic tumor; WHO, World Health Organization.

Survival outcomes

Based on the unweighted analysis, 5-year DFS rates of 67.6% and 58.5% and 10-year DFS rates of 54.3% and 51.2% were found in the OT and MIS groups, respectively (P=0.52; Figure 2A). With IPTW weighting included in the model, 5-year DFS rates were 68.9% and 52.6% and 10-year DFS rates were 55.1% and 48.9% for OT and MIS groups, respectively (P=0.35; Figure 2B).

Figure 2.

DFS and OS curves stratified by surgical approach. (A) DFS unweighted curve. (B) DFS weighted curve. (C) OS unweighted curve. (D) OS weighted curve. DFS, disease-free survival; MIS, minimally invasive surgery; OS, overall survival; OT, open surgical technique.

Based on the unweighted analysis, 5-year OS rates were 91.8% and 95.8% and 10-year OS were 83.2% and 85.6% for the OT and MIS groups, respectively (P=0.68; Figure 2C). With IPTW weighting included in the model, 5-year OS rates were 91.9% and 97.4% and 10-year OS rates were 82.1% and 88.3% for the OT and MIS groups, respectively (P=0.49; Figure 2D). Survival curves for DFS and OS stratified by tumor stage are provided in Figure S1.

Discussion

Early-stage thymoma is a highly curable neoplasm when treated with complete surgical resection. However, if pleural dissemination occurs at the at the time of resection, it can result in incurable disease associated with significant morbidity. Thymoma is typically an indolent tumor and when comparing different surgical approaches, it may be more important to consider the risk of recurrence than the risk of long-term mortality or operative morbidity. Recurrences occurring up to 180 months (15 years) after surgery in our study underscore the indolent nature of these tumors and the need for long-term follow-up. While MIS approach was an independent predictor of recurrent disease, it was not associated with inferior OS on Cox regression analysis, nor was there a statistically significant difference between the approaches by Kaplan-Meier analysis. We speculate that the typically indolent biological behavior of thymomas could contribute to these findings.

Relatively few tumor types (e.g., ovarian cancer, mesothelioma, low grade appendiceal mucinous neoplasm) have as great a propensity for metastatic serosal seeding as is seen with thymomas. Evidence of inferior outcomes after MIS resection of gynecologic neoplasms may suggest caution when resecting any tumor prone to serosal seeding via MIS approaches (38-44). Arguably the most compelling argument that MIS approaches for thymoma may have a greater propensity for pleural dissemination as compared to OT in our series comes from the observation of recurrence patterns by surgical approach. We found a strong association between the site of surgical approach and the site of recurrence (i.e., recurrences occurred ipsilateral to the surgical approach in every case of unilateral VATS/RATS), as well as a trend toward a greater number of recurrent nodules at the time of initial recurrence in the MIS group. There was a similar trend for ipsilateral pleural recurrence for patients who underwent open thoracotomy, however, these cases were less common. These results align closely with a study reported by Regnard et al. (45), who found that all intrathoracic recurrences occurred either near the original tumor site or in the same hemithorax where surgery was performed, and underscore the need to minimize tumor manipulation during surgery, regardless of approach, to avoid local dissemination.

A summary of current literature exploring oncologic outcomes (11,17,28,34,46-49) between the two approaches is outlined in Table 5. Kimura et al. (28) demonstrated an increased risk, albeit non-significant, of recurrence after VATS in tumors greater than 5 cm, with three recurrences in their VATS group and none after open surgery. Similarly, Liu et al. (17) observed a non-significant trend toward a higher recurrence rate after VATS, however, this study was also limited by a low number of recurrence events. Other reports on oncologic outcomes after MIS are similarly limited (32,46,47,50-53). Hwang et al. (11) recently published a large propensity-matched series investigating oncologic outcomes following OT and MIS approaches for thymoma. While no significant differences in the outcomes between approaches was identified, trends toward inferior DFS after MIS compared to OT. More recently, specific data comparing RATS to OT has been reported by Campbell et al. (49) and Niedermaier et al. (34). Campbell identified one recurrence event in the RATS group and none in the OT group and Niedermaier detected one recurrence event in each group. However, both studies were limited due to short follow-up of RATS cases and low sample size. In aggregate, the current body of literature suggests a slight trend toward increased risk of locoregional recurrence after MIS compared to OT.

Table 5. Summary of the current literature exploring oncologic outcomes between OT and MIS approaches.

Study	Study design	Approach (MIS/OT)	N	Stage (I/II/III/IV)	Histology (A, AB/B1, B2, B3/TC)	Tumor size (cm)	Follow-up (months)	Recurrence, n (%)	5-year DFS (%)‡	5-year OS (%)‡	Regression analysis for DFS of MIS (OT = ref)	Regression analysis for OS of MIS (OT = ref)
Agatsuma et al. 2017 (48)	PSM	VATS; MST	140; 140	81/59/0/0; 76/64/0/0	89/51/0; 85/55/0	4.4±2.8***; 5.7±4.0†	44.4 (0–190.8)***; 62.4 (0–216.0)§	3 (2.1); 4 (2.9)	93.9; 95.0	97.9; 97.1	–	–
Chao et al. 2015 (46)	PSM	VATS; MST	48; 48	17/31/0/0; 17/31/0/0	20/28/0; 15/33/0	5.8±1.8; 5.7±1.7†	66; 95§	2 (4.2); 3 (6.3)	–	–	–	–
Hwang et al. 2022 (11)	PSM	VATS, RATS; MST, LT	378; 378	202/176/0/0; 199/179/0/0	138/240/0; 132/246/0	4.4±1.9; 4.6±1.9†	76.7±1.2; 154.0±2.2†	11 (2.9); 13 (3.4)	93.0; 97.5;	98.1; 94.8;	1.92; 0.84–4.41; P=0.12	0.88; 0.42–1.84; P=0.74
Kimura et al. 2013 (28)	Cohort	VATS; MST	45; 29	41/4/0/0**; 17/12/0/0	21/24/0; 12/17/0	4.8±2.1**; 6.5±2.5†	53.7±24.5; 49.6±23.5†	3 (6.7); 0	–	–	–	–
Liu et al. 2014 (17)	Cohort	VATS; MST	76; 44	57/19/0/0; 37/7/0/0	43/28/0b; 23/18/0b	4.6±1.9**; 6.1±2.9†	61.9±52.0; 69.71±68.4†	2 (2.6); 1 (2.3)	96.9; 92.2	100; 96.8	–	–
Odaka et al. 2017 (47)	Cohort	VATS; MST	88; 52	38/45/5/0***; 17/15/18/2	37/45/6; 19/25/8	4.5 (1.4–9.7)**; 6.0 (1.5–13.0)§	54; 99§	8 (9.1); 8 (15.4)c	88.0; 84.2	–	–	–
Niedermaier et al. 2025 (34)	Cohort	RATS; MST, LT	35; 29	17/15/3/0; 14/12/3/0	22/13/0; 19/10/0	6.8 (5.7–9.0)**; 9.9 (7.5–13.2)¶	28 (20–45); 69 (37–95)¶	1(2.9); 1 (3.4)	–	–	–	–
Campbell et al. 2025 (49)	Cohort	RATS; MST, HC	25; 15	16/9/0/0; 7/7/1/0	17/8/0; 10//5/0	≥5a	58.8 (27.6–104.4); 115.2 (56.4–25.4)¶	1 (4.0); 0	95.0; 100	100; 100	–	–
Davis (IUSCCC) 2025	Cohort	VATS, RATS; MST, LT	49; 129	19/26/4; 51/53/25	20/29/0; 48/81/0	5.0 (4.0–7.5)**; 6.4 (4.0–9.0)¶	57.0 (34.0–100.0); 65.0 (23.0–116.0)¶	19 (38.8); 31 (24.0)	52.6; 68.9	97.4; 91.9	2.48; 1.31–4.71; P=0.005**	0.74; 0.27–2.02; P=0.56

^†, mean ± SD; ^‡, DFS/OS reported as weighted average calculated via stage-specific survival rates presented in respective papers; ^§, median (range) or median alone; ^¶, median (IQR); ^a, information on tumor size was not provided, but was greater than or equal to 5 cm in both groups; ^b, histology information for 5 patients in the MIS group and for 3 in the OT group were reported as missing; ^c, recurrences in the open group occurred only in stage III/IV patients, if stage I/II patients are isolated recurrence rates are MIS: 4 (4.8), open: 0 (0). **, P<0.01; ***, P<0.001. DFS, disease-free survival; HC, hemi-clamshell; IQR, interquartile range; IUSCCC, Indiana University Melvin and Bren Simon Comprehensive Cancer Center; LT, lateral thoracotomy; MIS, minimally invasive surgery; MST, median sternotomy; OS, overall survival; OT, open surgical technique; PSM, propensity score matched; RATS, robotic-assisted thoracic surgery; SD, standard deviation; TC, thymic carcinoma; VATS, video-assisted thoracic surgery.

Given the perioperative benefits of MIS techniques, MIS may be a preferable approach in many cases particularly with smaller tumors that are well contained within the anterior mediastinum soft tissues, wherein the risk of intraoperative dissemination is low. Careful patient selection for an MIS approach would seem prudent. If at any point during surgery there is concern for potentially violating the no-touch principle, conversion to OT should be considered. Along these same lines, there should be a low threshold to utilize OT from the outset for larger and more invasive tumors. On the other hand, perhaps a neoadjuvant approach could be considered to reduce tumor size and therefore decrease local recurrence rates. Further study would be needed to validate a neoadjuvant strategy for early stage thymoma. The role of adjuvant therapy in cases in which there is a concern for tumor dissemination similarly requires further study.

Other significant predictors of recurrence, including incomplete resection, and higher-stage disease, are consistent with prior literature (4-9). Our study also aligns with other reports that preoperative biopsy may increase the risk of locoregional tumor recurrence, consistent with current NCCN guidelines (9,35). While obtaining a biopsy prior to surgery may be associated with other factors contributing to recurrence risk, such as more invasive, larger tumors, biopsy was found to be an independent predictor of recurrence on multivariable analysis in the present study. Our study provides additional support to the notion that pre-surgical biopsies should be avoided in presumed early-stage thymomas in a low surgical risk patient. We speculate the risk of dissemination during pre-surgical biopsy may attributed to disruption of the tumor capsule and the ability of thymomas to survive and thrive on serosal surfaces, similar to inadvertent dissemination at the time of surgery. Given the current low evidence level of this NCCN guideline, we believe the results of our study meaningfully support this recommendation. Ongoing studies of radiomic signatures and artificial intelligence for predicting underlying histology and biologic behavior of solid tumors will be crucial for informing clinicians treating thymic tumors. These technologies may replace pre-surgical biopsies, as well as aide in guiding surgical approach (54).

Limitations

This study has several limitations. Primarily, the retrospective nature carries an inherent risk of selection bias. The IUSCCC is a center of excellence for thymic epithelial neoplasms and as such, a large proportion of patients with recurrent and stage IV thymoma are referred from OSH for evaluation. As a result, the present study had a higher recurrence and lower DFS rates than previous reports following surgery for early-stage thymoma. While we believe that patient selection for an MIS approach at IUSCCC was well balanced, for patients referred following surgery at OSH, we do not know the specific surgeon’s indication to utilize the MIS approach, nor their level of experience with MIS. All MIS patients underwent a lateral approach. A subxiphoid approach requires further study. Our study is also limited in discerning whether outcomes could differ between VATS and RATS approaches, which similarly requires further study. To minimize the impact of bias, multivariable models and IPTW were used to account for confounders in our study.

Conclusions

Patients with early-stage thymoma have an excellent OS with surgical therapy, either MIS or OT, which is impacted by stage, histology, and resection margin. However, our multivariable model suggests that DFS may be inferior in patients with pre-surgical biopsy and in select patients undergoing MIS approaches, converting a highly curable to a poorly or incurable situation. This insight is important for preoperative and surgical management in patients with anterior mediastinal masses which are suspicious for thymoma. During surgery, adherence to the time-honored “no-touch” technique is critical. This concern would seem to be particularly applicable to larger or more invasive tumors where MIS approaches should be utilized with caution at centers of surgical expertise. Larger series and prospective clinical trials with sufficient follow-up are needed to verify these findings.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the institutional review board of Indiana University (No. IRB00000219) and individual consent for this retrospective analysis was waived.

Data Sharing Statement

Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2024-2227/dss

jtd-17-08-5720-dss.pdf ^{(141.2KB, pdf)}

DOI: 10.21037/jtd-2024-2227