Introduction

Thymic epithelial neoplasms are classified as thymoma, thymic carcinoma (TC), and thymic neuroendocrine carcinoma (TNEC). The prevalences of TC and TNEC among all thymic neoplasms are around 15–20% and 2–3%, respectively [1, 2]. In the European Society of Thoracic Surgeons (ESTS) thymic database, in which 1122 patients were registered from 2007 to 2017, 28% of surgical cases had TC and 6.6% had TNEC [3]. In Japan, there were 2340 cases of resection of thymic epithelial neoplasms in 2016, including 314 TC (13%) and 40 TNEC (2%) [4]. In addition to limited data, TC and TNEC have a high rate of recurrence of 30–40% at 5 years, even after complete resection [5,6,7,8]; and the incidence of distant recurrence after surgery for TC and TNEC is almost as high as that of locoregional recurrence [8, 9], whereas this is not seen for recurrent thymoma.

Surgical resection for thymic epithelial neoplasms is the treatment of choice and it also has a significant impact on survival and recurrence [10,11,12]. In TC, complete resection, early pathological stage at the first surgery, and postoperative radiotherapy are significant prognostic factors for a lower rate of recurrence [5, 6, 11, 13,14,15]. The survival and recurrence rates of completely resected TNEC are similar to those for TC [6, 14]. A literature search suggested that little is known about the therapeutic strategy for postoperative recurrence of TC and TNEC. A recent report from the Mayo Clinic on long-term outcomes of treatment for recurrent thymic epithelial tumors proposed surgery for resectable and locoregional recurrent tumors, but a multimodal approach is often used including repeat surgery for recurrent TC and TNEC [9]. In a limited number of patients with recurrent TC, chemotherapy was associated with a prolonged progression-free interval [9].

To the best of our knowledge, few studies have so far examined the association of background factors and treatment on the outcomes of recurrent TC and TNEC. Therefore, the aim of this study is to examine these relationships in patients with a recurrence of these tumors.

Methods

Patients and data collection

This study was approved by the Kyoto University Hospital Institutional Review Board (IRB) (reference number: R1872) and the IRBs at Nishi-Kobe Medical Center, Fukuoka University Hospital, Kobe City Medical Centre General Hospital, Kitano Hospital, Shizuoka Municipal Hospital, Tenri Hospital, Shiga Medical Center for Adults, Nagara Medical Center, Takatsuki Red Cross Hospital, Osaka Red Cross Hospital, St. Luke’s International Hospital, Kyoto-Katsura Hospital, and Otsu Red Cross Hospital, all of which granted a waiver of consent for the study. A retrospective chart review was performed to identify patients who underwent thymectomy for TC and TNEC with a curative intent in the databases of the above twelve hospitals between May 1995 and December 2018. Data were collected regarding background (age and sex), intra- and perioperative data (surgical approach, extent of resection, completeness of resection, neoadjuvant therapy, and adjuvant therapy), pathological findings (histology, maximum diameter of the specimen, lymph node metastasis, and Masaoka stage), and follow-up (recurrence site, treatment for recurrence, and cause of death). In a case with an incomplete resection at initial resection, a complete disappearance of the disease was confirmed by the first postoperative surveillance chest computed tomography and a new discrete lesion by imaging modalities was regarded as recurrence. Surgery for recurrent disease was typically performed in patients with an intrathoracic, solitary, and resectable lesion based on the findings of CT with or without PET.

Statistical analysis

Categorical variables were compared by Fisher exact test, and continuous variables by Wilcoxon signed-rank test. Post-recurrence survival was calculated as the time from the date of recurrence until the date of death from any cause or the date on which the patient was last known to be alive. The diagnosis of recurrence was confirmed radiologically, including use of computed tomography (CT) with or without positron emission tomography (PET). Post-recurrence survival was analyzed using the Kaplan–Meier method with a log-rank test, and with a Cox proportional hazard model, with p < 0.05 considered to be significant. All analyses were performed using JMP® 14 software program (SAS Institute Inc., Cary, NC, USA).

Results

From our multi-institutional database, 60 recurrent cases (37 males, 23 females) were identified among 152 patients who underwent resection. The demographic data, tumor characteristics, and perioperative therapy for these cases are shown in Table 1. The median age was 63 (range 19–89) years. The median interval from the initial operation to recurrence was 17.5 (1.8–142) months. Complete resection at the initial operation was achieved in 45 patients (75%). Treatment for recurrence was performed in 42 patients (70%), including chemotherapy in 30 patients (50%), surgery in 16 (27%), and radiotherapy in 10 (17%); with 15 patients (25%) receiving more than one treatment. In the chemotherapy cases, the treatment was chemotherapy alone (n = 18) and in combination with surgery (n = 5), radiotherapy (n = 3), and surgery and radiotherapy (n = 2).

Table 1 Characteristics of the patients
Full size table

The median follow-up period from the initial operation was 44.2 (5.87–181) months, and this period after recurrence was 14.8 (0.1–153) months. The 5-year OS after recurrence was 23%. In the whole cohort, median post-recurrence survival was 14.8 months. According to a univariable analysis, advanced stage (HR 2.81, 95% CI 1.09–9.54), the interval between primary surgery and recurrence (HR 0.97, 95% CI 0.95–0.99), any treatment for recurrence (HR 0.27, 95% CI 0.13–0.58) and chemotherapy for recurrence (HR 0.46, 95% CI 0.22–0.95) were significant factors related to post-recurrence survival (Table 2). According to a multivariable analysis, the interval between primary surgery and recurrence (HR 0.98, 95% CI 0.95–1.01) and any treatment for recurrence (HR 0.23, 95% CI 0.04–1.16) tended to show a prolonged post-recurrence survival, although it was not statically significant. (Table 2). The survival curve comparing treatment and best supportive care (BSC) is shown in Fig. 1a. The 5-year post-recurrence survival rates after any treatment and BSC were 32% and 0%, respectively. In patients with any treatment after recurrence (n = 42), the 5-year post-recurrence survival rates after chemotherapy and non-chemotherapy were 33% and 28%, respectively (Fig. 1b). The patterns of recurrence in each treatment pattern are shown in Table 3.

Table 2 Univariable and multivariable analyses of the parameters with a potential influence on post-recurrence survival
Full size table
Fig. 1
figure 1

a Post-recurrence survival of all patients in the study (n = 60). b In thepatients with treatment after recurrence (n = 42)

Full size image
Table 3 Pattern of recurrence for each treatment
Full size table

Survival and treatment in patients with recurrence

A comparison of the demographic data, tumor characteristics, and perioperative therapy between BSC and any treatment is shown in Table 4. There was no significant difference between the two groups, but there were tendencies towards a higher age (p = 0.15), more advanced Masaoka stage (p = 0.19) at the initial operation, and a shorter interval between primary surgery and recurrence (p = 0.05) in patients who received BSC.

Table 4 Characteristics of the patients who received best supportive care (BSC) and any treatment
Full size table

Surgical resection for recurrence was performed in 16 patients (27%) (Table 5), and 11 of these patients received further therapy after this surgery, with chemotherapy only (n = 5), radiotherapy only (n = 3), and chemoradiotherapy (n = 3). Five patients received no additional therapy after surgery. Repeated surgical resection was performed in 2 patients (4 and 2 times, respectively). The survival curves comparing surgery and non-surgery for recurrence in patients with treatment after recurrence (n = 42) are shown in Fig. 2a. The 5-year post-recurrence survival rates in the surgery and non-surgery groups were 30% and 33%, respectively.

Table 5 Characteristics of the patients who underwent surgical resection for recurrence
Full size table
Fig. 2
figure 2

Post-recurrence survival in a subgroup analysis. a Survival curves comparing surgery and non-surgery for recurrence in patients with treatment after recurrence (n = 42). Non-surgical patients consisted of patients undergoing chemotherapy, radiotherapy, or both for recurrent thymic carcinoma and neuroendocrine carcinoma. b Survival curves comparing surgery and non-surgery for recurrence of thymic carcinoma (n = 42). Non-surgical patients consisted of patients undergoing chemotherapy, radiotherapy, or both or best supportive care for recurrent thymic carcinoma

Full size image

Chemotherapy for recurrence was performed in 30 patients (50%), with regimens of carboplatin and paclitaxel (n = 11), carboplatin and etoposide (n = 3), cisplatin and etoposide (n = 2), cisplatin, doxorubicin, vincristine, and cyclophosphamide (n = 2), and others (n = 8). Repeated chemotherapy was performed in 13 patients.

Radiotherapy for recurrence was performed in 10 patients (17%), including 2 treated with radiotherapy alone. One of these cases received radiotherapy for local recurrence and died of pneumonia at 5.7 months after recurrence. The other patient received palliative radiotherapy for multiple bone metastases and died of an unknown cause at 24.5 months after recurrence.

Subgroup information and analysis

Of the 60 patients, 42 were diagnosed with TC, 16 with TNEC, and 2 with TNEC combined with squamous cell carcinoma. According to a univariable analysis for TC cases (n = 42) (Table 6), chemotherapy for recurrence was significantly associated with improved post-recurrence survival after recurrence (HR 0.33, 95% CI 0.14–0.78), and surgery for recurrence (HR 0.44, 95% CI 0.10–1.28) tended to show a prolonged post-recurrence survival, although it was not statistically significant. According to a multivariable analysis, chemotherapy for recurrence was significantly associated with an improved post-recurrence survival (HR 0.32, 95% CI 0.13–0.76), and surgery for recurrence (HR 0.42, 95% CI 0.10–1.22) tended to show this association. In thymic carcinoma cases (n = 42), the survival curve comparing surgery and non-surgery is shown in Fig. 2b. The 5-year post-recurrence survival rates in the surgery and non-surgery groups were 27% and 19%, respectively. According to a univariable analysis for TNEC cases (n = 18) (Table 6), radiotherapy for recurrence was significantly associated with a worse post-recurrence survival after recurrence (HR 11.5, 95% CI 1.63–230).

Table 6 Univariable and multivariable analyses of the parameters with a potential influence on the overall survival
Full size table

Discussion

There is relatively little information on management of recurrent TC and TNEC after the initial resection of the primary lesion. Using the database of the Japanese Association for Research on the Thymus, Mizuno et al. found that more than 70% of patients with postoperative recurrent TC were managed non-surgically; however, detailed treatment information was scarce for all of the patients [16]. In the current study, we found that heterogenous treatment combinations are used for postoperative recurrent TC and TNEC, but with chemotherapy playing a major role in the multidisciplinary treatment.

Evidence-based guidelines for the management of recurrence are lacking, presumably due to the rarity of TC and the limited sample sizes of the published studies. Both NCCN [17] and ESMO (18) guidelines discuss the management of “advanced and recurrent” diseases, instead of making separate recommendations for advanced disease without prior treatment and recurrent disease after prior treatment. The Japanese Lung Cancer Society guidelines [19] recommend surgical resection in a multidisciplinary setting for resectable recurrent thymic epithelial tumors (including thymoma and other histologies). For TC alone, chemotherapy is recommended in patients with performance status 0–2 and recurrent disease, but information on the regimen, number of cycles, and no recurrence criteria have been established in the above guidelines.

In our literature search of chemotherapy for thymic carcinoma, we found several studies that included patients with advanced disease without treatment and with recurrent disease after treatment. In all identified studies, including three clinical trials [22, 24, 25] and two using real world data [21, 23], patients were treated with chemotherapy without radiotherapy or surgical resection. Lemma et al. [20], Furugen et al. [21], and Kim et al. [22] did not include cases with recurrent disease, Igawa et al. [23] had no cases with postoperative recurrence, and Hirai et al. [24] included 8 cases with postoperative recurrence, which accounted for 20% of all cases. The median progression-free survival (PFS) in the whole cohort was 7.5 (6.2–12.3) months. Inoue et al. [25] included 10 cases with postoperative recurrence, which accounted for 30% of all cases with TC. The median PFS was 7.6 months. These results cannot be compared to our study, given the heterogeneity of the treatment patterns in our patients. In addition, no previous study has so far examined the short- or long-term outcomes limited to patients with recurrent disease after prior treatment.

A suitable chemotherapy regimen can sometimes be selected based on the results of animal experiments, and there are a few thymoma-related animal models [26, 27], but none for TC or TNEC. Therefore, the selection of the optimal regimen has to be based on previous reports for related diseases. In a systematic review of 55 articles, Berghmans et al. evaluated the efficacy of different systemic therapies in thymoma and TC [28], and found that carboplatin plus amrubicin, carboplatin plus paclitaxel, and cisplatin plus docetaxel are favorable regimens [28, 29]. However, the efficacy of these regimens for recurrent TC and TNEC is uncertain.

Unlike surgical management of recurrent thymoma, the role of surgery in recurrent TC and TNEC is undetermined, and little information is available on surgical resection of these tumors [30, 31]. Okumura et al. [30] described a single patient who underwent resection for recurrent TC, who died 2 years later from TC. Yen et al. [31] reported 7 patients with surgical resection of recurrent TC, but the site, completeness of resection, and the number of lesions were not available, and all the patients underwent subsequent chemotherapy, which appeared to be associated with more favorable post-recurrence survival compared to radiotherapy plus chemotherapy [30]. In our series, 16 patients underwent surgical resection of recurrent TC, of whom 8 (50%) received subsequent chemotherapy with or without radiotherapy. In multivariable analysis, the significance of chemotherapy after surgery for recurrence was undetermined and this, therefore, requires a further study.

Resection of recurrent TC and TNEC, as well as the initial resection, should aim to achieve a complete resection; however, 6 patients of all surgical patients (n = 16) in our series ended up with only a debulking resection. One of the interpretations would be that we may unexpectedly encounter disseminated lesions that were not detected radiologically prior to surgery. In the current literature, there has been no study supporting debulking resection for recurrent TC and TNEC. Our findings suggested that debulking resection should be followed by chemotherapy, as was administered in 5 of 6 patients undergoing debulking resections.

A subgroup analysis of recurrent TNEC included 18 patients. The only significant factor associated with post-recurrence survival was radiotherapy, which was presumably performed in a more palliative setting. The small number of patients with recurrent TNEC did not allow us to identify any other significant factors and the management of recurrent TNEC is thus considered to be similar to those of recurrent TC.

There is also limited information on the utility of radiotherapy alone for recurrent TC. In a 2012 report from the Mayo Clinic [9], one of 9 patients with recurrent TC underwent radiotherapy alone for multiple bone metastasis, presumably for palliative purposes, without a response. Three of these 9 patients underwent radiotherapy for supraclavicular lymph node metastases (partial response); multiple bone lesions (stable disease), and liver and multiple mediastinal lymph node metastases (partial response). In our study, radiotherapy for recurrence tended to be associated with a poor prognosis, and only 2 of 10 patients received radiotherapy alone after recurrence. These findings suggest that radiotherapy as a monotherapy is essentially palliative in recurrent TC and TNEC.

There are several limitations associated this study, including the retrospective design; the relatively small sample size, which probably affected the power of the study; little information is available on the performance status and renal functions in our database; insufficient follow-up for thorough evaluation of the treatment outcome in each patient; and inconsistencies in decision-making about treatment. Patient selection and confounding factors presumably affected the results and thus were associated with some bias. Therefore, further and ideally prospective studies with a large sample size and more information are required to confirm our findings.

In conclusion, chemotherapy rather than surgery appears to be the mainstay treatment in managing patients with postoperative recurrent TC and TNEC. Long-term and careful follow-up after the initial resection is recommended. Multidisciplinary treatment including chemotherapy may, therefore, be most effective treatment modality for postoperative recurrent TC and TNEC.