Postoperative radiation therapy after complete resection of thymoma has little impact on survival
Abstract
BACKGROUND:
Postoperative radiation therapy for thymoma is widely used, although the clinical benefits are controversial. Furthermore, to the authors' knowledge, the relation between postoperative radiation therapy and cell type classified by the World Health Organization (WHO) is not known.
METHODS:
The records of 324 patients (ages 17-83 years; mean, 51 years; 160 males and 164 females) who underwent complete resection of a thymoma between 1970 and 2005 were reviewed. Mediastinum postoperative radiation therapy was performed for 134 patients. Survival rates and patterns of recurrence were determined according to Masaoka stage and WHO cell type.
RESULTS:
The 10-year disease-specific survival rates for patients with and without postoperative radiation therapy were 92.8% and 94.4%, respectively (P = .22). Subset analyses after stratifying by Masaoka stage and WHO cell type demonstrated that the 10-year disease-specific survival rate for patients without postoperative radiation therapy with Masaoka stage I and II, as well as those with WHO cell types A, AB, or B1, was 100%, which was satisfactory. Furthermore, the rates for patients with Masaoka stage III/IV and those with WHO cell types B2/B3 with or without postoperative radiation therapy were not found to be significantly different. In 24 patients with disease recurrence, pleural dissemination was observed most often, followed by distant metastases; local disease recurrence without other recurrence occurred in 2.
CONCLUSIONS:
The authors concluded that surgical resection alone is sufficient for thymoma patients with Masaoka stage I and II, and those with WHO cell types A, AB, and B1. Furthermore, an optimal treatment strategy should be established for patients with Masaoka stage III/IV and WHO cell type B2/B3 thymomas. Cancer 2009. © 2009 American Cancer Society.
When classifying the advancement of thymomas, the Masaoka staging system1, 2 has been widely used because it is a good predictor of prognosis for those patients,3 in addition to its clarity in assigning patients to an appropriate stage. Conversely, thymomas are also histologically classified based on a system proposed in 1999 and revised in 2004 by the World Health Organization (WHO).4 In that classification, which is also considered to compose another prognostic factor independent of Masaoka stage,5-7 thymomas are divided into 5 types; A, AB, B1, B2, and B3.
Although to our knowledge no randomized controlled trial has been conducted to date to establish a standard treatment for a thymoma, empiric evidence has led surgical resection to become the mainstay therapy. However, thymomas are also sensitive to radiation therapy (RT), which is often used in nonresectable patients and after surgery as an adjuvant therapy. The impact of postoperative RT on survival after complete resection of a thymoma has been reported in several reports, although the results vary.8-15 All of those reports used Masaoka stage to stratify their patients; however, to our knowledge, no known correlations between the effects of postoperative RT and WHO cell type have been presented to date. We conducted the current retrospective study to determine whether certain patients with a thymoma could achieve prolonged survival by receiving postoperative RT after a complete resection, by stratifying the patients according to Masaoka stage and WHO cell type. The relation between RT status and pattern of disease recurrence was also examined to investigate the effects of RT on the clinical course of thymoma patients.
MATERIALS AND METHODS
We reviewed the records of 324 patients who underwent complete resection of a thymoma at Osaka University Hospital and its affiliated hospitals during the 36-year period between 1970 and 2005. Those with a final diagnosis of thymic carcinoma and those who underwent a biopsy alone were not included. The patients included 160 males and 164 females, who ranged in age from 17 to 83 years (mean, 51 years). Postoperative RT was initially recommended for all patients who underwent a complete resection; however, it was not performed in cases of patient refusal. In addition, the criteria for postoperative RT at Osaka University Hospital changed during the study period, that is, Masaoka stage I patients were first eliminated from this recommendation in 1985, followed by stage II patients in 1998. Conversely, postoperative RT was not always recommended to patients at other hospitals that participated in this study. As a result, 134 patients received postoperative RT, of whom 119 were treated at Osaka University Hospital.
Postoperative RT was administered using megavoltage technology. Cobalt-60 and 6-megavolt (MV) x-ray devices were used from 1970 through 1981; cobalt-60 and 10-MV x-ray devices were used from 1982 through 1993; and 4-MV, 6-MV, and 10-MV x-ray devices were used from 1993. In 1997, evaluation of dose distribution using computed tomographic images and 3-dimensional planning became available (n = 20). The typical volume treated included the entire tumor bed and part of the involved adjacent lung when there was parenchymal involvement or as delineated by surgical clips, with at least a 1.5-cm margin. For some patients, the entire mediastinum was included in the treatment field, whereas noninvolved supraclavicular fossa was never included. Treatment portals included opposing anterior-posterior fields with differential weighting (1:1, 3:2, 2:1, 3:1) or a single anterior field. For all patients, a total dose of 40 to 50 grays (Gy) with 2 Gy per fraction was intended, although the RT course could not be completed in 12 patients. We adopted in principle 40 Gy, which adhered fundamentally to the recommendation stated in a study performed at our former institute.9 Thus, 84% of the patients in the postoperative RT group received a dose of 40 Gy (median, 40 Gy; average, 39.3 Gy [range, 10-50 Gy]).
Patient characteristics in regard to Masaoka stages and other perioperative therapy with reference to the status of postoperative RT are summarized in Table 1. Actuarial disease-specific and overall survival rates for all patients and those in each stage were calculated after dividing them by the status of postoperative RT. As we previously reported, the survival rates of patients with a stage III thymoma were found to be correlated with the involved organs.3 Therefore, stage III patients were divided into 3 groups according to the involved organs: those with invasion to the great vessels (V[+]L[+ or −]P[+ or −]), invasion to the lung but not the great vessels (V[−]L[+]P[+ or −]), and invasion to the pericardium but not the great vessels or lungs (V[−]L[−]P[+]).
| Postoperative RT | Yes | No |
|---|---|---|
| Masaoka stage | ||
| I | 31 | 119 |
| II | 43 | 33 |
| III | 53 | 30 |
| IVA | 4 | 5 |
| IVB | 3 | 3 |
| Total | 134 | 190 |
| Preoperative therapy | 6 | 25 |
| RT | 2 | 9 |
| Chemotherapy | 4 | 10 |
| Steroids | 2 | 11 |
| Postoperative therapy other than RT | 17 | 18 |
| Chemotherapy | 8 | 5 |
| Steroids | 12 | 13 |
- RT indicates radiation therapy.
Pathologic examinations were performed using hematoxylin and eosin-stained sections derived from paraffin-embedded blocks. Histologic diagnosis with reference to the WHO classification system was performed to classify the patients according to cell type A, AB, B1, B2, or B3, which was identified in 290 patients. Actuarial survival rates were calculated, and patterns of disease recurrence were reviewed based on each cell type.
When considering the clinical course of a patient with a thymoma, which lasts for more than a decade in many, disease-specific survival rates may accurately reflect the clinical course of thymomas as compared with overall survival rates. Therefore, both disease-specific and overall survival rates were calculated in the current study. Patients who died from a disease other than the thymoma were regarded as censored at the time of death in treating disease-specific survival rates. Actuarial disease-specific and overall survival rates were calculated using the Kaplan-Meier method, and statistical differences between survival curves were examined with a log-rank test. The frequencies of distribution between groups were tested with a chi-square test. A P value of <.05 was considered significant. Statistical analyses were performed using the personal computer software package StatView 5 (SAS Institute, Cary, NC). The institutional review board of Osaka University Hospital approved the design of the study and consented to waive the need to obtain informed consent from the patients.
RESULTS
The actuarial 10-year and 20-year disease-specific survival rates for patients who received postoperative RT were 92.8% and 83.5%, respectively, whereas they were 94.4% and 94.4%, respectively, for those treated without RT (P = .2208). Conversely, the actuarial 10-year and 20-year overall survival rates for those who received postoperative RT were 80.7% and 56.5%, respectively, whereas they were 86.2% and 40.4%, respectively, for those treated without RT (P = .0640). The distribution of patients in regard to Masaoka stage was skewed in the present population, thus disease-specific and overall survival rates for the patients were calculated by stage (Table 2). The disease-specific survival rate for patients with stage I and II was >90% regardless of the status of postoperative RT. Furthermore, in all patients with stage III disease, there were no significant differences noted in disease-specific survival rate between those treated with and without postoperative RT (Table 2). Next, we compared stage III patients after dividing them according to the involved organs. The actuarial 10-year and 20-year disease-specific survival rates for patients in the V(+)L(+ or −)P(+ or −) group who underwent postoperative RT (n = 12) were 62.3% and 41.6%, respectively, whereas they were 77.1% and 77.1%, respectively, for those treated without RT (n = 13; P = .48). Similarly, those rates for the V(−)L(+)P(+ or −) group treated with postoperative RT (n = 27) were 93.8% and 54.8%, respectively. Although the follow-up period for those treated without postoperative RT in this group (n = 12) was not adequate to calculate the 10-year disease-specific survival rate, the disease-specific survival rate after 9 years was 80.0% (P = .2565). No patients in the V(−)L(−)P(+) group died during the follow-up period of between 1.8 to 20.6 years (median, 9.9 years), which included 14 patients treated with postoperative RT and 5 treated without it. In patients with stage IVA disease, all 5 patients without postoperative RT and 3 of 4 patients with postoperative RT survived >5 years, whereas the remaining 1 patient with postoperative RT died 1.4 years after surgery. With regard to those with stage IVB disease, all 6 died within 5 years.
| Stage | Postop RT | No. | 10DSR, % | 20DSR, % | P | 10OSR, % | 20OSR, % | P |
|---|---|---|---|---|---|---|---|---|
| I | + | 31 | 96.3 | 90.3 | 77.3 | 58.7 | ||
| − | 119 | 100 | 100 | NA | 92.3 | .0126 | ||
| II | + | 43 | 100 | 100 | 85.0 | 73.6 | ||
| − | 33 | 100 | NA | 96.7 | .2486 | |||
| III | + | 53 | 87.7 | 62.3 | 79.9 | 35.5 | ||
| − | 30 | 85.1 | 85.1 | .8010 | 69.1 | 46.1 | .4852 | |
| IV | + | 7 | 62.5 (5DSR) | 62.5 (5OSR) | ||||
| − | 8 | 66.7 (5DSR) | .6628 | 66.7 (5OSR) | .6628 |
- Postop RT indicates postoperative radiation therapy; 10DSR, 10-year disease-specific survival rate; 20DSR, 20-year disease-specific survival rate; 10OSR, 10-year overall survival rate; 20OSR, 20-year overall survival rate; +, received; −, not received; NA, not applicable; 5DSR, 5-year disease-specific survival rate; 5OSR, 5-year overall survival rate.
The overall survival rate for patients with stage I and II disease without postoperative RT was >90%, whereas it was smaller in those treated with postoperative RT compared with those treated without. Furthermore, in all patients with stage III disease, there were no significant differences in the overall survival rate noted between those treated with and without postoperative RT. With regard to patients with stage IV disease, all deaths that occurred were because of the tumor, thus, the overall survival rates were identical to the disease-specific survival rates, which were analyzed above (Table 2).
The distribution by WHO cell type according to Masaoka stage and status of postoperative RT for each category is summarized in Table 3. None of the patients with WHO cell types A (n = 17) or B1 (n = 54) died during the median follow-up period of 5.4 years and 8.3 years, respectively, regardless of the status of postoperative RT. Furthermore, no patient of disease recurrence occurred in cases with cell type A, whereas 1 patient with cell type B1 developed disease recurrence with pleural dissemination, which had been classified as stage III disease at the time of resection and the patient received postoperative RT. Among those with cell type AB (n = 92), 3 patients treated with postoperative RT died at 1.5 years, 3.8 years, and 16.7 years, respectively, after surgery, whereas no patient treated without postoperative RT died. Patterns of disease recurrence were found in 1 patient with pleural and pericardial dissemination, 1 with pleural dissemination and lymph node metastasis, and 1 with disease recurrence at the site of a needle biopsy. The actuarial 10-year and 20-year disease-specific survival rates for all patients with these 3 cell types who underwent postoperative RT (n = 50) were 95.8% and 91.5%, respectively, whereas the 15-year disease-specific survival rate for those treated without postoperative RT was 100% (n = 113).
| WHO Cell Type | Postop RT | Masaoka Stage | Total | 10DSR, % | 20DSR, % | 10OSR, % | 20OSR, % | Patten of Disease Recurrence | ||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| I | II | III | IVA | IVB | ||||||||
| A | + | 2 | 1 | 3 | 100 | 100 | 66.7 | 66.7 | None | |||
| − | 12 | 1 | 1 | 14 | 100 | — | 100 | — | None | |||
| AB | + | 9 | 12 | 5 | 2 | 28 | 92.4 | 84.0 | 88.9 | 66.6 | PD (2), local, unknown | |
| − | 47 | 15 | 1 | 1 | 64 | 100 | — | 85.1 | — | None | ||
| B1 | + | 6 | 10 | 3 | 19 | 100 | 100 | 88.8 | 66.0 | PD | ||
| − | 25 | 5 | 5 | 35 | 100 | — | 90.5 | — | None | |||
| B2 | + | 10 | 12 | 25 | 3 | 1 | 51 | 94.9 | 73.9 | 76.5 | 48.3 | PD (8), lung, local, brain, and pericard |
| − | 21 | 6 | 12 | 5 | 1 | 45 | 92.9 | 92.9 | 82.5 | 41.3 | Lung and PD (2) | |
| B3 | + | 7 | 4 | 1 | 12 | 88.9 | — | 67.9 | — | PD + local + liver, local + lung, and PD + lung | ||
| − | 4 | 3 | 3 | 1 | 11 | 65.6 | — | 65.6 | — | Lung | ||
| Others* | + | 0 | ||||||||||
| − | 6 | 1 | 1 | 8 | ||||||||
| Total | 142 | 72 | 61 | 9 | 6 | 290 | ||||||
- WHO indicates World Health Organization; Postop RT, postoperative radiation therapy; 10DSR, 10-year disease-specific survival rate; 20DSR, 20-year disease-specific survival rate; 10OSR, 10-year overall survival rate; 20OSR, 20-year overall survival rate; +, received; −, not received; PD, pleural dissemination, pericard, pericardial dissemination.
- * “Others” in WHO cell type includes cases that could not be assigned to type A, AB, B1, B2, or B3.
Of 96 patients with WHO cell type B2, 8 died, which included 6 with stage III disease and 2 with stage IVA disease. All but 1 of the patients who died received postoperative RT. There was no significant difference noted with regard to actuarial disease-specific survival rate between patients treated with and those treated without postoperative RT (P = .4368) (Table 3). Disease recurrence occurred in 16 patients, which included 1 patient with stage II disease, 11 patients with stage III disease, and 4 patients with stage IVA disease, with postoperative RT performed in 13 of those. The pattern of disease recurrence in most included pleural or pericardial dissemination and distant metastases. Local disease recurrence without the existence of other recurrent disease occurred in 1 patient with stage III disease who underwent postoperative RT.
Among patients with WHO cell type B3, 3 died from the thymoma, 2 of whom received postoperative RT. No significant difference in actuarial disease-specific survival rates were found between those treated with and without postoperative RT (P = .3501). Disease recurrence occurred in 4 patients, which included 2 with stage III, and 1 each with stage IVA and stage IVB disease, with postoperative RT performed in 3 of those. The patterns of disease recurrence included pleural dissemination, distant metastases in the liver and lung, and local disease recurrence. Local disease recurrence accompanied by distant metastasis occurred in 2 patients.
No group of patients divided by WHO cell type demonstrated a significant difference in overall survival rate according to status of postoperative RT (P = .7496, .9634, .1837, and .7229 for WHO cell types AB, B1, B2, and B3, respectively) (Table 3). Statistical analysis of overall survival rates for patients with WHO cell type A was not attempted because no events occurred in those patients treated without postoperative RT, and the number of patients treated with postoperative RT was too small.
DISCUSSION
In the current study, no significant improvement in survival was noted for patients who were treated with postoperative RT compared with those without, regardless of Masaoka stage, which is consistent with previously reported results.10-16 With regard to those with Masaoka stage I and II disease, the results of the current study can be explained in part by the finding that the long-term outcomes of our thymoma patients who did not undergo postoperative RT were satisfactory, which made it difficult to demonstrate that postoperative RT had a further benefit on survival, thus suggesting that complete resection alone was sufficient to achieve a good prognosis in this population. Although the survival rates of stage III and IV disease patients were reduced compared with those with stage I and II, the differences in survival according to status of postoperative RT among each stage were not found to be significant, which may indicate that postoperative RT after complete resection in patients with stage III and IV disease does not alter long-term survival.
The indication of postoperative RT for thymoma after complete resection has long been considered controversial, mainly because, to the best of our knowledge, no randomized controlled study has been conducted to date. For patients with stage II and III disease, Curran et al.8 and Nakahara et al.9 advocated postoperative RT. Curran et al.reviewed 19 patients with stage II disease and 7 with stage III disease, and their findings supported postoperative RT for stage II disease based on their finding that mediastinal recurrence occurred in 6 of 18 patients treated without RT within 5 years, whereas no recurrence was noted in 1 stage II case and 4 stage III patients treated with postoperative RT. The mediastinum was the most common site of disease recurrence in their report. In addition, reports on the outcome of patients with a thymoma have been accumulating,7, 11, 12, 14, 16, 17 which led us to speculate regarding the effects of postoperative RT on stage II thymomas. Those reports studied from 25 to 61 patients with stage II disease, and noted a low rate of disease recurrence (range, 2.0-9.8%), with a frequent pattern of disease recurrence being pleural dissemination, which is consistent with the findings in the current study. Although this issue would ideally be settled by a prospective randomized controlled study, we consider that the accumulated data are sufficient to judge that postoperative RT for stage II thymoma is not effective.
Conversely, Nakahara et al insisted that a patient with a stage III thymoma should undergo postoperative RT for the entire mediastinum after a complete resection.9 That report, produced by our former institute, was based on a 1-arm observation study that did not include patients who did not undergo postoperative RT. Later, after accumulating clinical data for an additional 20 years, we reported that the prognosis and pattern of disease recurrence in patients with a stage III thymoma depended on the involved organs, and that recurrence was most often noted as pleural dissemination.3, 18 Thus, we currently consider that establishment of a strategy against pleural dissemination, rather than mediastinal recurrence, is more important than postoperative RT for this group of patients, as described earlier. To the best of our knowledge, no beneficial effect of postoperative RT on the survival of individuals with stage III thymoma has been demonstrated to date in reports published from other institutions,11, 15 which is consistent with the results of the current study.
Long-term survival rates for patients with WHO cell types A, AB, and B1 were considered adequate without postoperative RT, making it difficult to demonstrate further improvement because of postoperative RT and suggesting that its role was not apparent in this population, which indicates that surgery alone is a sufficient treatment strategy for thymoma patients with WHO cell types A, AB, and B1, as well as those classified as having Masaoka stage I and II disease who undergo a complete resection. Thus, such patients may be eliminated as candidates for postoperative RT. Furthermore, of the 56 patients in the current study classified as having stage III, IVA, or IVB disease, and WHO cell types B2 or B3, 37 were Masaoka stage III and WHO cell type B2, and there was no significant difference in survival noted with regard to the status of postoperative RT among them (data not shown). The other combinations included no more than 8 patients, which was not considered to be a sufficient number with which to demonstrate the effect of RT on long-term survival.
Patterns of disease recurrence are another issue regarding postoperative therapy for patients with completely resected thymomas. The irradiation area in the patients in the current study was the mediastinum, which may function as a prophylactic against local recurrence of the tumor. The most frequent pattern of disease recurrence was pleural dissemination, followed by distant metastasis, mainly in the lungs. However, that phenomenon was common in patients who were treated both with and without postoperative RT and thus it does not appear that postoperative RT to the mediastinum prevents local recurrence. Other reports have also noted that the pleura is the most frequent site of recurrence,11, 15, 19 and that performance of postoperative RT does not affect the rate of pleural recurrence.11 Therefore, it would be reasonable to form a strategy against pleural dissemination and distant metastasis, rather than local control in the mediastinum, for thymomas that may require treatment in addition to complete resection. Some investigators have reported reductions in disease recurrence rates after entire hemithorax RT of approximately 15 Gy as adjuvant therapy after complete resection of a thymoma.20, 21 However, they also noted that radiation pneumonitis of grade 2 or greater occurred in 15% to 25% of those patients who underwent entire hemithorax RT, which led us to consider that this treatment should not be routine. In this context, systemic chemotherapy may be a good candidate modality. Although we could not find any reports in literature indicating that adjuvant chemotherapy after complete resection would ameliorate the outcome, there are studies describing favorable effects of systemic chemotherapy for patients with advanced thymomas. Various therapeutic regimens were used in those, including the quadruplet of doxorubicin, cisplatin, vincristine, and cyclophosphamide22 and triplets of cisplatin, doxorubicin, and cyclophosphamide23; cisplatin, doxorubicin, and cyclophosphamide24; cisplatin, epirubicin, and etoposide24, 25; and cisplatin, doxorubicin, and methylprednisolone,26 as well as others. Because to our knowledge there is no standard chemotherapeutic protocol for a thymoma, it is important to determine the most suitable regimen as well as establish a multimodality treatment strategy for patients with stage III/IV thymoma and WHO cell types B2/B3.
A limitation of the current study is its retrospective design. We clearly understand that it would be ideal for a sufficient number of patients to be recruited within a short period to form a prospective study; however, the relative rareness of the disease makes it quite difficult to do so. In addition, confirmation of the effects of treatments require a long period, because the clinical course of a thymoma is slow, making it difficult to perform prospective randomized clinical studies of affected individuals. Therefore, the distribution of patients in the current study in regard to RT status was not randomized. However, we are certain that the bias used for selecting patients who received postoperative RT was minimal, because we determined the criteria for postoperative RT based on Masaoka stage, not on a “case-by-case” basis. Conversely, the majority of the patients who received postoperative RT underwent surgery at Osaka University Hospital, which might cause a bias from the differences among the hospitals in this study. Nevertheless, all thoracic surgeons who participated in the current study were trained under a system established by Osaka University Hospital and its affiliated hospitals, and we therefore are confident that quality of the surgical procedures as well as perioperative care was maintained at a high level. Another important issue in the current study may be the long duration required to accumulate an adequate number of patient records because of the rareness of the disease, as such a long period may skew the outcome of a retrospective study. Over the study period of several decades, many aspects of thymoma treatment have changed, such as application of thoracoscopic surgery, new apparatuses for RT, new chemotherapeutic agents, and improvements in radiographic imaging technologies. Any of those variations in the patients in the current study might have skewed our results.
In conclusion, the effects of postoperative RT on survival and disease recurrence after complete resection in patients with thymoma were examined based on Masaoka stage and WHO cell type. A good candidate group of thymoma patients could not be identified after dividing by Masaoka stage and WHO cell type. However, patients with stage I and II thymoma, as well as those with WHO cell types A, AB, and B1, can be eliminated from the list of candidates for postoperative RT. Conversely, establishment of an optimal treatment strategy for patients with Masaoka stage III and IV disease, and WHO cell types B2 and B3, is needed to further improve their long-term outcome.
Conflict of Interest Disclosures
The authors made no disclosures.
