Myasthenia gravis with thymic epithelial tumour: a retrospective analysis of a Japanese database

Transcription

1 European Journal of Cardio-Thoracic Surgery 49 (2016) doi: /ejcts/ezv380 Advance Access publication 3 November 2015 ORIGINAL ARTICLE Cite this article as: Nakajima J, Okumura M, Yano M, Date H, Onuki T, Haniuda M et al. Myasthenia gravis with thymic epithelial tumour: a retrospective analysis of a Japanese database. Eur J Cardiothorac Surg 2016;49: Myasthenia gravis with thymic epithelial tumour: a retrospective analysis of a Japanese database a Jun Nakajima a, *, Meinoshin Okumura b, Motoki Yano c, Hiroshi Date d, Takuya Onuki e, Masayuki Haniuda f and Yoshifumi Sano g, for the Japanese Association for Research of Thymus Department of Thoracic Surgery, The University of Tokyo Graduate School of Medicine, Tokyo, Japan b Department of General Thoracic Surgery, Osaka University, Osaka, Japan c Department of Oncology, Immunology and Surgery, Nagoya City University, Nagoya, Japan d Department of Thoracic Surgery, Kyoto University, Sakyo-ku Kyoto, Japan e Department of General Thoracic Surgery, Tsuchiura Kyodo General Hospital, Tsuchiura, Ibaraki, Japan f Division of Chest Surgery, Department of Surgery, Aichi Medical University, Nagakute, Aichi, Japan g Center of Chest Medicine and Surgery, Ehime University, Toon-city, Ehime, Japan * Corresponding author. Department of Thoracic Surgery, The University of Tokyo Graduate School of Medicine, Hongo, Bunkyo-ku, Tokyo , Japan. Tel: ; fax: ; nakajima-tho@h.u-tokyo.ac.jp ( J. Nakajima). Received 14 July 2015; received in revised form 21 September 2015; accepted 28 September 2015 Abstract OBJECTIVES: Myasthenia gravis (MG) has been reported to correlate with earlier stage and Type B thymomas by the World Health Organization classification. We analysed a large database of clinical characteristics of patients with MG and thymic epithelial tumours to elucidate whether the severity of MG affected postoperative survival of those with thymic epithelial neoplasms. METHODS: We conducted a multi-institutional study on the patients who had undergone surgical treatment for thymic epithelial tumours between 1991 and We examined Masaoka stage, pathological type, serum titre of antiacetylcholine receptor antigen, severity of MG and postoperative prognosis of the patients with or without MG. RESULTS: Of the 2835 registered patients at 32 institutes belonging to the Japanese Association for Research on the Thymus, 2638 were eligible for the study. MG was present in 598 patients (23%). Patients with MG had thymic epithelial tumours with significantly earlier stage (P = ) and significantly smaller tumours (P = 0.000) than those without. Type A, Type AB thymomas and thymic carcinomas were less frequently observed in patients with MG. Three of 304 patients (1%) with thymic carcinomas had MG preoperatively. Serum titres of antiacetylcholine receptor antibodies were positive in 98% of patients with MG, and 23% of those without. Patients with generalized MG had significantly higher titres of antiacetylcholine receptor antibodies than those with ocular MG (P = 0.000). The postoperative 30-day mortality rate was 0.3%. Postoperative 5- and 10-year survival rates of thymoma patients with MG and those without MG were 94 and 96, and 89 and 89%, respectively. We found no statistical difference in the postoperative survival rate or recurrence-free rate between the two groups. We found no significant statistical differences of these rates by MGFA classification or surgical approach. CONCLUSIONS: We conclude that earlier stage, smaller size or Type B thymomas are more frequently associated with MG, and MG may have no impact on the overall survival of patients with thymoma. We suggest that postoperative survival time of the patients with MG and thymic epithelial tumours may be mainly affected by the tumours, not by MG. Keywords: Myasthenia gravis Thymoma Thymic carcinoma Thymectomy Survival analysis INTRODUCTION The relationship between thymoma and myasthenia gravis (MG) is well known. No randomized, controlled trials exist that examine the effectiveness of thymectomy for patients with both MG and thymoma. Large cohort studies on patients with these diseases had been conducted that indicated MG correlated with earlier Masaoka stage and Type B thymomas by World Health Organization (WHO) classification [1, 2]. We analysed a nation-wide multi-institutional database on a large number of patients with thymic epithelial tumours who had undergone thymectomy to examine clinicopathological characteristics, serum titre of antiacetylcholine receptor antigen and postoperative prognosis of the patients with or without MG. We sought to elucidate whether the severity of MG affected postoperative survival of the patients with thymic epithelial neoplasms or not. The Author Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.

2 J. Nakajima et al. / European Journal of Cardio-Thoracic Surgery 1511 MATERIALS AND METHODS In 2011, the Japanese Association for Research on the Thymus asked the major centres in Japan to participate in a database project on thymic epithelial tumours. Thirty-two institutions responded to the request and joined the project. The collected information included patient characteristics, status of MG, type of surgery, pathological diagnosis, perioperative therapy, clinical outcomes and additional treatments for recurrence. For patients who underwent surgery between 1991 and 2010, the diagnosis of thymic epithelial tumour was reviewed and confirmed by pathologists in each centre according to the 2004 revision of the WHO classification [3]. The study protocol was approved by the institutional review boards of all participating hospitals, including the Ethics Committee of The University of Tokyo Graduate School of Medicine (Approval Number: 3778), and the need for patient consent was waived. We performed a retrospective observational study using the database, focusing on the WHO classification of thymoma, Masaoka s classification of thymoma [4], serum titre of antiacetylcholine receptor antibodies (ARABs), Myasthenia Gravis Foundation of America (MGFA) classification [5], association with other autoimmune diseases, overall survival time, thymic epithelial tumour recurrencefree interval, cause of death and other clinical factors. Recurrence-free survival time was defined as the time between surgery and the last follow-up or recurrence of the thymic epithelial tumour. We used the χ 2 test for categorical variables, and unpaired t-test for continuous variables. The survival curves were estimated according to the Kaplan Meier method, and the differences between groups were compared using a log-rank test. The Cox proportional hazard model was used for calculating hazard ratios of covariates, including the presence of MG and other various potential prognostic factors showing P-value less than 0.1. The data were considered significant when the P-value was less than The statistical analysis was performed with EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), a graphical user interface for R (R Foundation for Statistical Computing, Vienna, Austria) [6]. RESULTS Of the 2835 patients registered in the database, 2724 who had undergone complete or partial (no less than 80% of the tumour was resected) resection were reviewed. The following cases were excluded: 25 cases with no data on MG, 48 without Masaoka staging and 13 with no postoperative follow-up data. Thus, the total number of eligible patients was 2638 (Fig. 1). The number of patients who had been diagnosed as MG preoperatively was 598 (23%). Mean observation time was 5.61 years. Age at thymectomy was 56.7 ± 13.5 years [mean ± standard deviation (SD)]. Patients with MG were significantly younger than those without MG. The number of male patients was 1241 (47%). The proportion of male patients with MG was significantly lower than that without MG. A total of 968, 1003, 398, 149 and 120 patients had Masaoka stage I, II, III, IVa and IVb thymic epithelial tumours, respectively. The proportion of patients with MG with early Masaoka stage was greater than those without MG (P = ). Three hundred and four (12%) had thymic cancers, including 51 with neuroendocrine tumours. Thymoma was found in 2334 patients. WHO pathological classification was examined: 183, 649, 561, 602 and 302 patients had Type A, AB, B1, B2 and B3 thymomas. The distribution of WHO pathological type was significantly different between patients with MG and those without MG (P = 0.000). The percentages of Type A and AB thymomas were lower in patients with MG (Table 1). Three of 304 patients (1%) with thymic carcinomas had MG preoperatively. The pathological types of these tumours included one basaloid carcinoma, one undifferentiated and one not otherwise specified. The maximal diameter of the tumour of the patients without MG was 6.0 ± 2.7 cm (mean ± SD), and that of the patients with MG was 4.6 ± 2.2 cm. Patients with MG had smaller thymic epithelial tumours in this study (Table 1). Forty-one patients (2%) had red cell aplasia. Other autoimmune diseases including hypogammaglobulinaemia, rheumatoid arthritis, Sjogren syndrome and others were found in 85 patients. We observed no significant differences in the incidence of autoimmune Figure 1: Diagrammatic representation of the number of patients available for analysis. MG: myasthenia gravis. Table 1: Masaoka stage and World Health Organization pathological type of the thymic epithelial neoplasms Total MG MG+ P Number of patients Age at operation 56.7 ± ± ± (mean ± SD) Male/female 1241/ / / Masaoka stage Stage I 968 (37%) 732 (36%) 236 (40%) Stage II 1003 (38%) 769 (38%) 234 (39%) Stage III 398 (15%) 322 (16%) 76 (13%) Stage IVa 149 (6%) 111 (5%) 38 (6%) Stage IVb 120 (4%) 106 (5%) 14 (2%) WHO pathological type Type A 183 (7%) 158 (8%) 25 (4%) Type AB 649 (25%) 550 (27%) 99 (17%) Type B1 561 (21%) 421 (20%) 140 (23%) Type B2 602 (23%) 362 (18%) 240 (40%) Type B3 302 (11%) 221 (11%) 81 (13%) Others 37 (1%) 27 (1%) 10 (2%) Carcinoma 304 (12%) 301 (15%) 3 (1%) Size of the main tumour (cm) Mean ± SD 5.7 ± ± ± SD: standard deviation; MG: myasthenia gravis; WHO: World Health Organization. THORACIC

3 1512 J. Nakajima et al. / European Journal of Cardio-Thoracic Surgery disease between patients with MG (5%) and those without MG (4%) (Table 2). Serum titre of ARAB was examined preoperatively in 1737 patients. Of 1608 patients with thymoma, 813 (51%) had positive ARAB. On the other hand, 10 of 129 patients with thymic carcinomas had positive ARAB (8%). ARAB titre was positive in 559 of 568 patients with MG (98%). On the other hand, ARAB titre was positive in 264 of 1169 patients without MG (23%). In 1043 patients with thymoma but not MG, 257 had positive ARAB (25%) (Fig. 2). Masaoka staging or WHO pathological typing was not associated with ARAB titre in patients with thymoma. The MGFA score was assessed in 549 of the 598 patients with MG preoperatively. Of the 549 patients, 173, 272, 89, 6 and 9 were graded as Class I, II, III, IV and V, respectively. Serum titres of ARAB in patients with Class I MG (ocular type) and those with Class II V MG (generalized type) were 23.8 ± 46.2 pmol/ml (mean ± SD) and 62.9 ± 96.4 pmol/ml, respectively. Patients with generalized MG had significantly higher titres of ARAB than those with ocular MG (P = 0.000) (Fig. 3). Corticosteroids were administered to 163 of 499 patients with MG (33%) and 10 of 1421 patients without MG (1%) preoperatively. One hundred and eighty of 433 patients with MG (42%) and 43 of 1209 patients without MG (4%) were given corticosteroids preoperatively. It was found that 145 of 293 patients with MGFA stage 2 or higher stages (49%) had taken steroids; on the other hand, 27 of 113 patients with MGFA stage 1 (24%) had taken steroids. Patients with generalized MG were more frequently given steroids preoperatively (P = 0.000). Of the 595 patients with both thymoma and MG, thoracoscopic surgery was performed on 46 (8%). On the other hand, 191 of 1739 patients (11%) with thymoma without MG underwent thoracoscopic surgery. The rate of thoracoscopic surgery of the patients with both thymoma and MG was significantly lower than those with thymoma without MG (P = 0.023). Forty-three of 46 patients (93%) undergoing thoracoscopic surgery and 426 of 549 patients (78%) undergoing open surgery had Stage I or II thymoma with MG, respectively. Patients who underwent thoracoscopic surgery had a greater rate of incidence of Stage I or II thymoma than those who underwent open surgery (P = 0.011). Forty-four of 46 patients (96%) with both thymomas and MG underwent total thymectomy through thoracoscopy. On the other hand, 68 of 191 patients (36%) with thymomas without MG underwent total thymectomy through thoracoscopy. Total thymectomy through thoracoscopy was more frequently performed on thymoma patients with MG (P = 0.000). Table 2: Other coexisting autoimmune diseases Of the 2638 patients, 7 (0.3%) died within 30 days postoperatively. During postoperative follow-up, 41 patients with MG (7%) and 203 patients without MG (10%) had died. MG crisis was the Figure 2: Status of antiacetylcholine receptor antibody in patients with thymic epithelial neoplasms and myasthenia gravis. MG: myasthenia gravis; ARAB: serum antiacetylcholine receptor antibody. Figure 3: Serum titre of antiacetylcholine receptor antigen and MGFA classification. MGFA: Myasthenia Gravis Foundation of America. Total MG MG+ P value n = 2638 n = 2040 n = 598 Pure red cell aplasia 41 (2%) 34 (2%) 7 (1%) Hypogammaglobulinaemia 12 (0.4%) 10 (0.5%) 2 (0.3%) Rheumatoid arthritis 12 (0.5%) 9 (0.4%) 3 (0.5%) Sjogren syndrome 10 (0.4%) 1 (0.1%) 9 (2%) Thyroid diseases 8 (0.3%) 3 (0.2%) 5 (0.8%) Other diseases 33 (1.3%) 28 (1.4%) 5 (0.8%) Total 116 (4%) 85 (4%) 31 (5%) 0.29 MG: myasthenia gravis.

4 J. Nakajima et al. / European Journal of Cardio-Thoracic Surgery 1513 cause of death in 5 patients with MG and 2 patients without MG who had developed MG postoperatively. The deaths of 10 patients with MG and 113 without MG were caused by thymic tumour recurrence. Eighty of the 113 patients without MG had thymic carcinoma. Malignant neoplasm not associated with thymic tumours was the secondary cause of death (Table 3). Postoperative 5-year and 10-year overall survival rates of all patients were 92 and 85%, respectively. Five-year postoperative survival rates of patients with thymoma and those with thymic carcinoma were 95 and 68%, respectively. Postoperative survival prognosis was significantly worse in patients with thymic carcinomas (P = 0.000). We thus excluded patients with carcinomas to examine whether MG affected the survival prognosis of patients with thymoma. Among patients with both thymoma and myasthenia gravis, 5 of 46 patients (11%) who underwent thoracoscopic surgery and 92 of 549 patients (17%) who underwent open surgery suffered Table 3: Early and late deaths in patients with and without thymic carcinoma Cause of death Total MG+ MG n = 2638 n = 598 n = 2040 Early and late deaths, including patients with thymic carcinoma MG crisis 7 (0.3%) 5 (0.8%) 2 (0.1%) Recurrence of tumour 123 (5%) 10 (2%) 113 (6%) Other malignancies 34 (1%) 9 (2%) 25 (1%) Other diseases/unknown 80 (3%) 17 (3%) 63 (3%) Total 244 (9%) 41 (7%) 203 (10%) n = 2334 n = 595 n = 1739 Early and late deaths excluding patients with thymic carcinoma MG crisis 7 (0.3%) 5 (0.8%) 2 (0.1%) Recurrence of tumour 43 (2%) 9 (2%) 34 (2%) Other malignancies 28 (1%) 9 (2%) 19 (1%) Other diseases/unknown 67 (3%) 17 (3%) 50 (3%) Total 145 (6%) 40 (7%) 105 (6%) MG: myasthenia gravis. from postoperative complications. Three of 46 patients (7%) and 36 of 549 patients (7%) had postoperative MG-related complications. The rate of postoperative complications was not significantly different between the two groups. Postoperative 5-year and 10-year survival rates of the thymoma patients with and without MG were 94 and 96, and 89 and 89%, respectively. We found no statistical difference in the survival rate between the two groups by the log-rank test. Postoperative 5-year and 10-year thymoma recurrence-free survival rates of patients with and without MG were 93 and 92, and 89 and 87%, respectively. We found no statistical difference in recurrence-free survival between the two groups by the log-rank test. We also analysed postoperative survival and recurrence-free rates stratified by MGFA classification. We found no significant statistical differences between MGFA classes (Fig. 4). Univariate analysis by the log-rank test showed that older age, R1/R2 resection, advanced Masaoka stage and WHO type B pathology negatively affected postoperative survival of the 2334 patients with thymoma. Multivariate analysis revealed that older age, R1/2 resection, advanced Masaoka stage and WHO type B pathology had negative impact on postoperative survival in the patients with thymoma. However, MG had no significant effect on postoperative overall survival (Table 4). Survival rates by surgical approaches of the patients with both thymomas and MG were examined. Five-year overall survival rates of the patients undergoing open surgery and thoracoscopy were 95 and 93%, respectively. Thymoma-free survival rates of the patients undergoing open surgery and thoracoscopy were 94 and 97%, respectively. No statistical differences of survival rates were observed between the two groups. Preoperative steroid intake did not affect postoperative complications or overall survival rate of the patients with both MG and thymoma: five-year overall survival rates of those patients with preoperative steroids and those without steroids were 93 and 96%, respectively. In summary, thymoma patients with MG underwent thoracoscopic surgery less frequently than those without MG. However, total thymectomy was intended for almost all patients with earlier stages of thymoma and MG. We found no difference in the rate of postoperative complications or postoperative survival between the patients undergoing thoracoscopy and those undergoing open surgery. THORACIC Figure 4: Postoperative overall survival curves (left) and recurrence-free survival curves (right) of patients with thymoma stratified by MGFA classification. Patients with myasthenia gravis versus those without myasthenia gravis. MG: myasthenia gravis; MGFA: Myasthenia Gravis Foundation of America.

5 1514 J. Nakajima et al. / European Journal of Cardio-Thoracic Surgery Table 4: Potential prognostic factors by univariate and multivariate analyses of the patients with thymoma (n = 2334) Risk factors Univariate analysis P value Myasthenia gravis 0.82 Other autoimmune diseases 0.70 Performance status 2/ Older age (>56 years) Male gender Tumour resection without 0.94 thymectomy Larger diameter of tumour 0.63 (>5.6 cm) R1/R2 resection Open surgery Masaoka stage III/IVa/IVb WHO type B1/B2/B Odds ratio 95% CI P value Multivariate analysis Myasthenia gravis Older age (>56 years) Male gender R1/2 resection Open surgery Masaoka stage III/IVa/IVb WHO type B1/B2/B CI: confidence interval; WHO: World Health Organization. DISCUSSION Relationship between thymic epithelial tumours and myasthenia gravis The role of thymoma in autoimmunity is still not clear. However, it is well known that thymomas are frequently associated with MG: 10 28% of patients with MG had thymoma [7 10]. A systemic review reported that the pooled estimate of the incidence of thymoma in patients with MG was 21% [11]. On the other hand, 15 50% of patients with thymoma have MG as a comorbid condition [1, 2, 12]. In our multi-institutional study, the incidence of MG in patients with thymic epithelial tumour was 22.7% and the incidence rate became 25.5% if patients with thymic cancer were excluded. Three patients had both thymic cancer and MG. They all had a high ARAB serum titre preoperatively. It has been reported that thymic cancer is occasionally accompanied by MG [13]. The majority of these thymic cancers were pathologically diagnosed as squamous cell carcinoma. These might have been borderline cancers that lie between overt cancer and invasive thymoma, but others were diagnosed as poorly differentiated carcinomas [14]. In our cases, a basaloid carcinoma in 1 patient and an undifferentiated carcinoma in 1 patient, which were clearly different from thymoma, were included. Relationship between World Health Organization pathological typing of thymoma and myasthenia gravis Patients with Type B thymomas as defined by the WHO pathological classification more frequently had MG than those with Type A or AB thymomas [15]. In fact, MG was more frequent in patients with Type B1 to B2 cortical thymomas [16]. This suggests that the number of CD4+CD8+ double-positive immature T lymphocytes infiltrating the thymoma might be related to the onset of MG. In this study, we also found significant differences in the rate of MG between patients with different WHO pathological types of thymoma. MG patients without thymoma who have a high serum titre of ARAB often have thymic abnormalities, with 60 70% with hyperplastic germinal centres in their thymus. On the other hand, patients with both MG and thymoma usually have an involuted thymus that does not appear to affect the onset of MG [17]. ARAB is produced in the lymph nodes and bone marrow, not within the thymoma [18]. Relationship between Masaoka staging and myasthenia gravis It is well known that WHO typing correlates with Masaoka staging: WHO type B3 thymomas frequently show an advanced Masaoka stage. In the present study, Type B1, B2 or B3 thymomas were more frequently associated with MG than A or AB thymomas. However, patients with MG had an earlier stage of thymoma [2]or a similar stage [19]. In this study, we found that patients with MG had smaller mediastinal epithelial tumours than those without MG. This implies that some patients were found to have a thymoma because of the onset of MG. Serum antiacetylcholine receptor antibody titre and thymic epithelial tumour Almost all patients with both thymoma and MG have positive serum ARAB. The serum ARAB level reflects the severity of MG [20]. In this study, thymoma patients with MGFA stage II or greater had a higher ARAB titre than those with MGFA stage I. On the other hand, 24% of thymoma patients without symptomatic MG had positive serum ARAB. Thus, a positive ARAB is not always associated with overt MG. Post-thymectomy MG is another issue: 1 28% of patients undergoing thymectomy had postoperative MG [1, 21]. Postthymectomy MG can occur in patients with or without recurrence of thymoma. Positive serum ARAB in the preoperative period is reported to be a risk factor for post-thymectomy MG. The serum ARAB titre is always high in patients with post-thymectomy MG, irrespective of the status of recurrence [22]. Postoperative survival and myasthenia gravis In this study, we found no significant differences in postoperative overall or recurrence-free survival between patients with MG and those without MG. Completeness of resection, Masaoka stage and WHO type of the thymoma are factors influencing the postoperative overall survival rate [23]. Older reports showed that MG negatively affected postoperative survival of patients with thymoma [24]. This was partly due to an increased surgical risk caused by perioperative exacerbation of MG [25]. We agree that MG does not affect postoperative survival [4, 11, 16, 20]. In this study, we made a further subset analysis on patients with both MG and thymoma. Patients with generalized MG preoperatively took steroids more frequently than those with ocular MG. Preoperative steroid intake did not affect postoperative

6 J. Nakajima et al. / European Journal of Cardio-Thoracic Surgery 1515 complications or overall survival rate of the patients with both MG and thymoma: Five-year overall survival rates of those patients with preoperative steroids and those without steroids were 93 and 96%, respectively (P =0.28). Cause of late death in patients with thymic epithelial tumours In our study, the majority of patients died from recurrence of thymic tumour. Excluding patients with thymic cancer, we found that extrathymic malignancy was the second leading cause of death. Papastestas et al. retrospectively studied the incidence of extrathymic neoplasms in patients with MG. Of 226 patients with both thymoma and MG, 19 had extrathymic neoplasms. Those patients who did not undergo thymectomy had a higher incidence of extrathymic neoplasms [7]. The limitation of the present study was the retrospective nature of the study that might result in biases. We also did not have data on the postoperative course of MG of patients in the database. In conclusion, we elucidated the characteristics of thymic epithelial tumours that the patients with MG had: Patients with MG were more likely to have thymomas with WHO type B pathologies, earlier stages and smaller sizes. We suggested that postoperative survival time of the patients with MG and thymic tumours might be mainly affected by the tumours, not by MG. ACKNOWLEDGEMENTS The Japanese Association for Research on the Thymus comprises the following Japanese institutes (representatives); Osaka University (Meinoshin Okumura), Chiba University (Ichiro Yoshino), Nagoya City University (Motoki Yano), National Cancer Center Hospital (Hisao Asamura), Shinsyu University (Kazuo Yoshida), Kyoto University (Hiroshi Date), National Cancer Center Hospital East (Kanji Nagai), Nagoya University (Kohei Yokoi), University of Tokyo ( Jun Nakajima), Okayama University (Shinichiro Miyoshi), Osaka Medical Center for Cancer and Cardiovascular Diseases (Masahiko Higashiyama), Juntendo University (Kenji Suzuki), Niigata University (Masanori Tsuchida), Nippon Medical School (Shuji Haraguchi), Seirei Mikatahara General Hospital (Hiroshi Niwa), Tokushima University (Kazuya Kondo), Tokyo Metropolitan Cancer and Infectious diseases Center Komagome Hospital (Hirotoshi Horio), Kinki Chuo Chest Medical Center (Akihide Matsumura), Kyushu University (Tatsuro Okamoto), Tokyo Medical University (Norihiko Ikeda), University of Occupational and Environmental Health (Fumihiro Tanaka), Kobe University (Yoshimasa Maniwa), Tokyo Women s Medical University (Takamasa Onuki), Kumamoto University (Makoto Suzuki), Nagasaki University (Takeshi Nagayasu), Fukuoka University (Akinori Iwasaki), Shikoku Cancer Center (Hiroshi Suehisa), Shizuoka Cancer Center (Yasuhisa Ohde), Hokkaido Cancer Center (Keishi Kondo), Tsuchiura Kyodo Hospital (Takuya Onuki), Aichi Medical University (Masayuki Haniuda) and Ehime University (Yoshifumi Sano). Conflict of interest: none declared. REFERENCES [1] Kondo K, Monden Y. Thymoma and myasthenia gravis: a clinical study of 1,089 patients from Japan. Ann Thorac Surg 2005;79: [2] Filosso PL, Venuta F, Oliaro A, Ruffini E, Rendina EA, Margaritora S et al. Thymoma and inter-relationships between clinical variables: a multicentre study in 537 patients. Eur J Cardiothorac Surg 2014;45: [3] Ströbel P, Marx A, Zettl A, Müller-Hermelink HK. Thymoma and thymic carcinoma: an update of the WHO Classification Surg Today 2005; 35: [4] Masaoka A, Monden Y, Nakahara K, Tanioka T. Follow-up study of thymomas with special reference to their clinical stages. Cancer 1981;48: [5] Jaretzki A III, Barohn RJ, Ernstoff RM, Kaminski HJ, Keesey JC, Penn AS et al. Myasthenia gravis: recommendations for clinical research standards. Task Force of the Medical Scientific Advisory Board of the myasthenia gravis Foundation of America. Neurology 2000;55: [6] Kanda Y. Free statistical software: EZR (Easy R) on R commander. (21 September 2015, date last accessed). [7] Papatestas AE, Genkins G, Kornfeld P, Eisenkraft JB, Fagerstrom RP, Pozner J et al. Effects of thymectomy in myasthenia gravis. Ann Surg 1987;206: [8] Monden Y, Uyama T, Nakahara K, Fujii Y, Hashimoto J, Ohno K et al. Clinical characteristics and prognosis of myasthenia gravis with other autoimmune diseases. Ann Thorac Surg 1986;41: [9] Drachman DB. Myasthenia gravis. N Engl J Med. 1994;330: [10] Phillips LH. The epidemiology of myasthenia gravis. Ann N Y Acad Sci 2003;998: [11] Mao ZF, Mo XA, Qin C, Lai YR, Hackett ML. Incidence of thymoma in myasthenia gravis: a systematic review. J Clin Neurol 2012;8: [12] Ruffini E, Detterbeck F, Van Raemdonck D, Rocco G, Thomas P, Weder W et al. Tumours of the thymus: a cohort study of prognostic factors from the European Society of Thoracic Surgeons database. Eur J Cardiothorac Surg 2014;46: [13] Liu HC, Hsu WH, Chen YJ, Chan YJ, Wu YC, Huang BS et al. Primary thymic carcinoma. Ann Thorac Surg 2002;73: [14] Filosso PL, Guerrera F, Rendina AE, Bora G, Ruffini E, Novero D et al. Outcome of surgically resected thymic carcinoma: a multicenter experience. Lung Cancer 2014;83: [15] Okumura M, Miyoshi S, Fujii Y, Takeuchi Y, Shiono H, Inoue M et al. Clinical and functional significance of WHO classification on human thymic epithelial neoplasms: a study of 146 consecutive tumors. Am J Surg Pathol 2001;25: [16] Sakamoto M, Murakawa T, Konoeda C, Inoue Y, Kitano K, Sano A et al. Survival after extended thymectomy for thymoma. Eur J Cardiothorac Surg 2012;41: [17] Müller-Hermelink HK, Marx A, Geuder K, Kirchner T. The pathological basis of thymoma-associated myasthenia gravis. Ann N Y Acad Sci 1993; 681: [18] Shiono H, Roxanis I, Zhang W, Sims GP, Meager A, Jacobson LW et al. Scenarios for autoimmunization of T and B cells in myasthenia gravis. Ann N Y Acad Sci 2003;998: [19] Yu L, Zhang XJ, Ma S, Jing Y, Li F, Krasna MJ. Different characteristics of thymomas with and without myasthenia gravis. Ann Surg Oncol 2012;19:94 8. [20] Lucchi M, Ricciardi R, Melfi F, Duranti L, Basolo F, Palmiero G et al. Association of thymoma and myasthenia gravis: oncological and neurological results of the surgical treatment. Eur J Cardiothorac Surg 2009;35: [21] Namba T, Brunner NG, Grob D. Myasthenia gravis in patients with thymoma, with particular reference to onset after thymectomy. Medicine (Baltimore) 1978;57: [22] Nakajima J, Murakawa T, Fukami T, Sano A, Takamoto S, Ohtsu H. Postthymectomy myasthenia gravis: relationship with thymoma and antiacetylcholine receptor antibody. Ann Thorac Surg 2008;86: [23] Okumura M, Ohta M, Tateyama H, Nakagawa K, Matsumura A, Maeda H et al. The World Health Organization histologic classification system reflects the oncologic behavior of thymoma: a clinical study of 273 patients. Cancer 2002;94: [24] Wilkins EW Jr, Edmunds LH Jr, Castleman B. Cases of thymoma at the Massachusetts General Hospital. J Thorac Cardiovasc Surg 1966;52: [25] Loach AB, Young AC, Spalding JM, Smith AC. Postoperative management after thymectomy. Br Med J 1975;1(5953): THORACIC