Thymoma and the increased risk of developing extrathymic malignancies: a multicentre study

Transcription

1 European Journal of Cardio-Thoracic Surgery 44 (2013) doi: /ejcts/ezs663 Advance Access publication 11 January 2013 ORIGINAL ARTICLE a b c d e f g Thymoma and the increased risk of developing extrathymic malignancies: a multicentre study Pier Luigi Filosso a, *, Claudia Galassi b, Enrico Ruffini a, Stefano Margaritora c,d, Luca Bertolaccini d, Caterina Casadio e, Marco Anile f and Federico Venuta f Department of Thoracic Surgery, University of Torino, Turin, Italy Center for Cancer Prevention in Piedmont, San Giovanni Battista Hospital, Turin, Italy Catholic University of Sacred Heart, Rome, Italy Centro Oncologico Fiorentino (C.F.O.), Florence, Italy Unit of Thoracic Surgery, Santa Croce e Carle Hospital, Cuneo, Italy Unit of Thoracic Surgery, Amedeo Avogadro University, Novara, Italy Department of Thoracic Surgery, Fondazione Eleonora Lorillard Spencer Cenci; Policlinico Umberto I, University of Rome Sapienza, Rome, Italy * Corresponding author. Department of Thoracic Surgery, University of Torino, San Giovanni Battista Hospital, Genova 3, Torino, Italy. Tel: ; fax: ; pierluigi.filosso@unito.it (P.L. Filosso). Received 19 July 2012; received in revised form 27 October 2012; accepted 11 November 2012 Abstract OBJECTIVES: Although thymoma is considered a relatively indolent neoplasia, patients affected by this disease are at high risk of developing second tumours (STs). The aim of this study is to assess the risk of developing STs after surgical thymoma resection. METHODS: A multicentre retrospective study of patients operated on for thymoma within five Italian Thoracic Surgery Institutions, between 2000 and 2011, was conducted. The overall STs number and incidence were calculated. The number of metachronous STs was compared with the expected cancer number (ECN) in an Italian population, and the standardized incidence ratio (SIR) and 95% confidence intervals were calculated. Potential variables of STs predictors were also evaluated. RESULTS: There were 302 patients; myasthenia gravis (MG) was observed in 166 (55%) and other autoimmune syndromes in 49 of them. In 118 patients (39.1%), the Masaoka thymoma stage was greater than II and in 194, the WHO histological type ranged from B1 to C. Fifty STs were observed (28 metachronous, 4 synchronous and 18 detected before thymoma). The observed metachronous STs number was significantly higher than ECN. An increased risk of STs development was observed in advanced stage thymomas and in those with histological high grade. On the contrary, MG seems to be a protective factor in STs development. CONCLUSIONS: Our study confirms the high risk of developing STs in patients with thymoma. Aggressive forms of thymoma are those in which this risk appears to be more evident. The central role of an intrinsic immune system alteration might be the key to interpret this phenomenon. Keywords: Thymoma Cancer Myasthenia gravis Surgery Cause of death THORACIC Thymoma and thymic carcinoma are relatively rare neoplasms arising from the epithelial cells of the thymus. In the adult population, they account for 30 50% of all the neoplasms of the anterior mediastinum [1, 2], with an overall incidence in the USA of 0.15 cases per person-year [3]. The thymus plays a central role in the immune system, because it is the site of maturation for T cells from bone marrow progenitors and screens them for autoreactivity. This screening is done through the autoimmune regulator system [4]. T lymphocytes with autoaggressive potential features are usually deleted into the thymic medulla; normally, only self-tolerant and immunosurveillant against cancer cells T lymphocytes are released systemically. Most thymic tumours present a thymopoietic activity Presented at the 20th European Conference on General Thoracic Surgery, Essen, Germany, June and generate CD4 + and CD8 + T cells [5] responsible for a number of paraneoplastic disorders; myasthenia gravis (MG) is the most common, occurring in 30 50% of the patients [6, 7]. Patients with thymoma show an increased risk of developing second primary extrathymic malignancies, with a frequency ranging between 8 and 28% [3, 8 13]. Most of the published studies include small series of patients; they report singleinstitution experiences or are based on National Cancer Registries; however, the association between thymic tumours and lymphoma, gastrointestinal, prostate and lung cancers is observed. Furthermore, patients with thymoma develop second malignancies more frequently than those undergoing thymectomy for non-thymomatous diseases [14, 15]; this risk seems not clearly correlated with thymoma stage or histology, presence of autoimmune disorders or administration of adjuvant therapy. The Author Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.

2 220 P.L. Filosso et al. / European Journal of Cardio-Thoracic Surgery We hereby report a multicentre study from five high-volume institutions for thymoma surgery; we have compared the results with the cancer incidence rates (sex, age and area of residency in Italy) reported by the Italian Cancer Registry Network (AIRTUM); the aim of this study is to assess the risk of developing an extrathymic cancer after surgical resection of thymoma. PATIENTS AND METHODS This is a multicentre retrospective study of patients operated for thymoma between 2000 and 2011 at five different institutions. The mean number of patients treated at each centre was The main end-point of our study was the assessment of the incidence of extrathymic second tumours (STs) (all sites) diagnosed after thymoma resection (metachronous STs), at the time of thymoma surgery (synchronous) or before the diagnosis of the thymic tumour. Data were collected at the time of diagnosis of thymoma or during the follow-up after surgery. Patients undergoing biopsy only and those with inoperable tumours were excluded from this study. Outcome data were obtained from hospital records, outpatient controls or telephone interviews. The standard preoperative assessment included complete radiological workup with total body computed tomography (CT). Magnetic resonance was also performed when required. Positron emission tomography scan was not routinely performed at any centre. Neurological consultation to rule out MG was performed on a case-by-case basis according to the clinical status. Complete sternotomy was the standard approach; lateral thoracotomy or other combined incisions have seldom been performed when required by the clinical presentation. Surgery was considered radical if complete tumour resection (R0) was achieved and incomplete in case of micro/macroscopic residuals (R1 R2). Histology was assessed according to the 2004 World Health Organization (WHO) classification [16]. All histological types were included in this study with the exception of primary neuroendocrine thymic tumours. Staging was determined according to the Masaoka-Koga classification system [17]. Preoperative chemotherapy (CT) was uniformly administered in case of anticipated tumour unresectability; postoperative radiotherapy (RT) and/or CT were administered to patients presenting with invasive thymoma (Stages II IV) according to the policy of each institution. The follow-up protocol was quite similar at all the Institutions, including CT scan every 6 months for the first 3 years and afterword on a yearly basis or on clinical demand. Data about extrathymic STs included tumour development date (synchronous, metachronous or before the thymoma diagnosis), site and treatment (surgery, CT and RT). The following variables were evaluated as potential STs predictors: presence of MG or other autoimmune disorders, thymoma stage and histology, administration of induction and adjuvant therapy. We dichotomized both WHO histology and Masaoka stages into two categories (WHO: 1, A+AB; 2, B1 to C; Masaoka stage: 1, I and II; 2, III and IV), as we did in a previous report [18]. Statistical analysis The incidence of metachronous STs (all sites) was compared with the expected number of cancers on the basis of cancer incidence rates (by sex, age and area of residence in Italy) reported by the Italian Cancer registry Network (AIRTUM) for the period Standardized incidence ratio (SIR) and 95% confidence intervals (95% CI) were then calculated on the whole population and then separately by sex, stage and histology of the thymoma. Cumulative incidence of metachronous ST was calculated considering the competing risk of death, using the stcompet Stata command. Competing-risks regression models according to the method of Fine and Gray, taking death into account as a competing event, were used to explore the role of variables significantly affecting the risk of metachronous STs. Hazard ratios (HR) for metachronous STs with 95% CI were calculated. We first performed univariable models, one for each of the possible predictors: male sex (reference: females); diagnosis of MG (reference: No); diagnosis of other autoimmune syndromes (reference: No); high stage of thymoma (Stage III IV with Stage I II as reference); high grade of thymoma (B1 to C with A to AB as reference); adjuvant chemotherapy/rt (reference: No); multivariable regression models were subsequently performed including sex, MG and thymoma stage as covariates, while excluding other autoimmune syndromes (due to colinearity with MG), grade of the thymoma and use of adjuvant chemotherapy/rt (due to colinearity with stage of thymoma) and age (not associated with STs) were performed. The association between risk of death and the incidence of ST was evaluated using a Cox regression model adjusted for sex, age and thymoma stage. The variable indicating the incidence of the ST was included in the model as a time-varying covariate. The statistical analysis was performed using the STATA 11.2 software (StataCorp LP, TX, USA). RESULTS Three hundred and two patients (49% males, mean age 54.4 years, range 8 87 years) undergoing surgery for thymoma were retrospectively included in the study. The characteristics of this group of patients are reported in Table 1. MG was observed in 166 patients (55%), while 49 (16.2%) had other autoimmune syndromes. One hundred and ninety-four patients (64.2%) had histological high-grade thymoma (WHO-type B1 to C), and in 118 (39.1%), the Masaoka stage was higher than II. Adjuvant chemotherapy/rt was administered in 161 patients (53.3%). After a median follow-up of 64.5 months (ranging from 0.2 to 138 months), 41 patients (13.6%) died. We observed 50 extrathymic STs: 28 metachronous, 4 synchronous and 18 detected before thymoma diagnosis; 8 patients had two STs. Table 2 reports STs according to timing at diagnosis and site. Lung cancer (#2) was the most frequent tumour among the synchronous STs. Digestive (#8), lung (#5), breast (#3), thyroid (#3) and kidney (#3) were the most common metachronous STs. The number of metachronous STs (#28) was significantly higher than the expected number of cancers (14.41, SIR 1.94, 95% CI ) (Table 3). In the group of metachronous STs, we observed no gender difference; the incidence of these tumours is slightly higher in patients with high-grade thymoma and in those at advanced stage (Table 4).

3 P.L. Filosso et al. / European Journal of Cardio-Thoracic Surgery 221 Table 1: Patients characteristics Patients with resected thymoma, n = 302 Age in years, mean (range) 54.4 (8 87) Males, n (%) 149 (49.0) Myastenia gravis, n (range) 166 (55.0) Other autoimmune syndromes, n (%) a 49 (16.2) Unknown 34 (11.3) WHO histology, n (%) A 14 (4.7) AB 53 (17.6) B1 39 (12.6) B2 97 (32.1) B3 71 (23.5) C 26 (8.6) Not available 2 (0.7) Masaoka stage, n (%) I 65 (21.5) II 118 (39.1) III 5 (31.5) IV 23 (7.6) Not available 1 (0.3) Adjuvant chemo/radiotherapy, n (%) 161 (53.3) Unknown 9 (3.0) Follow-up duration in months, 64.5 (64.6) median (IQR) Second tumours, n (%) Metachronous 28 (9.3) Synchronous 4 (1.3)) Diagnosed before thymectomy 18 (6.0 Deaths, n (%) 41 (13.6) a LES: Lupus erythematosus systemicus, autoimmune thyroiditis, chronic gastritis, rheumatoid arthritis, hypogammaglobulinaemia, Sjogren s syndrome, pure red cell aplasia, autoimmune neuritis. The cumulative incidence of metachronous STs at up to 60 months of follow-up is reported in Fig. 1. The probability of occurrence of STs at 6 months follow-up was about 8%. The results of the competing-risks analyses on potential predictors of metachronous ST are reported in Table 4. Although none of the examined variables was significantly associated with the incidence of STs, an increased risk of STs development was observed in patients with histological high-grade thymoma, and particularly among those with a high-stage thymoma. The presence of MG seems to have a protective effect on STs development, although these data again were not statistically significant. The incidence of STs significantly increases the patients subsequent risk of death (adj HR 3.43, 95% CI , P < 0.01). DISCUSSION Our study confirms that patients with thymoma show an increased risk of developing second malignancies when compared with the normal population. Extrathymic malignancies are heterogeneous (solid and nonsolid) and may present at any time with respect to the onset of thymoma: they may be synchronous or metachronous and may also be diagnosed years before the thymic tumour. In our series, the risk of developing ST is about 2-fold higher compared with the normal population; high-grade and advanced stage thymomas are clearly at increased risk of developing them. There are several potential explanations for this finding, including a possible effect of adjuvant therapy (especially RT) and the strict clinical/radiological follow-up that increases the chances of finding other tumours. THORACIC Table 2: Number of metachronous and synchronous STs and of tumours diagnosed before thymoma, by site Metachronous ST Synchronous ST T diagnosed before thymectomy n % n % n % Digestive Colon Gallbladder Pancreas Gastric Respiratory Lung Female breast Female reproductive Male reproductive Prostate Testis Urinary system Kidney Thyroid Lymphoma Head/neck Melanoma Cutaneous Kaposi (lung) All sites

4 222 P.L. Filosso et al. / European Journal of Cardio-Thoracic Surgery Table 3: Number of metachronous ST observed in the cohort of patients with resected thymoma (Obs), number of tumours expected (Exp) according to the incidence rates estimated by AIRTUM in the Italian population and SIR with 95% CI Obs Exp SIR 95% CI Whole population Males Females Thymoma Masaoka stage I II Thymoma Masaoka stage III IV Thymoma WHO grade A, AB Thymoma WHO grade B1 to C Data are presented for the whole population, and stratified by sex, stage and histology of thymoma. The values in bold indicate statistical significance. Table 4: Results of the competing-risks regression models according to the method of Fine and Gray HR 95% CI P Univariable models Male gender Age Myastenia gravis Other autoimmune syndromes Thymoma stage III IV Thymoma WHO grade B1 to C Adjuvant therapy Multivariable model Male gender Myastenia gravis Thymoma stage III IV Results are expressed as HR for metachronous ST, with 95% CI. We have recorded 50 extrathymic STs among 302 patients operated on during a relatively short period (from 2000 to 11); 28 STs were metachronous, 4 synchronous and 18 were diagnosed before thymectomy. The overall incidence was 16.5%. Evidence linking thymoma and extrathymic cancers is clearly supported by a literature review [3, 8 13], with an incidence ranging between 8 and 38%; however, most studies include small single-institution series. The relative rarity of thymoma (0.13 cases per personyear in the USA) [3] is one of the limiting factors in assessing the real incidence of STs. Soudjijan [19] reported a 21% incidence and compared it with patients with parathyroid adenoma, in whom the incidence was 8%. In 1994, Masaoka [20] reported a series of 392 patients operated for MG; 102 of them had thymoma with STs in 9 of them. More recently, Pan reported that the incidence of STs in a Taiwanese population of 192 patients with thymoma was 8% [9]. Using patients with nasopharyngeal carcinoma as control, he observed a statistically significant increased risk of STs in the thymoma population. Similar results were observed by Welsh [8] and Yen [11], even if they did Figure 1: Cumulative incidence of ST. not consider any control group. Engels [3, 3], Travis [10], Weksler [12], Gadalla [13] and de Jong [21] compared their data with those reported in the National Cancer Registries. A large number of patients (668 to 2170) were collected during a long period and a statistically significant risk of developing second cancers was observed also in these series. However, the correct patient clinical stratification is the limitation of such studies: thymomas were classified as benign or malignant without specifying what it means according to the Masaoka and WHO classifications. Moreover, as Engels pointed out [3], these registries usually lack cases, and some of them are uploaded with wrong diagnoses, especially at the centres with less experience in this field. This makes all the calculations potentially incorrect. Reports differed extremely from each other in the STs site. Engels [3, 20] reported an increased frequency of non-hodgkin s lymphoma (NHL), soft tissue sarcoma and digestive tumours. Welsch [8] observed an increased number of colorectal, lung, breast and thyroid cancers; furthermore, in his study, 14 patients had at least three STs. Weksler [12] identified lymphoma, prostate, breast and colorectal cancers as those more frequently diagnosed before thymectomy; lung, prostate, breast cancers and lymphoma are more commonly detected after the procedure. Gadalla [13] observed an increased risk for non-melanomatous skin cancers, NHL, endocrine and prostate cancers after thymectomy. In our series colon, lung, breast, kidney and thyroid were the most frequent metachronous primary extrathymic cancers, while lung cancer was the most common synchronously. Surprisingly, we did not find any increase in NHL, as previously reported by some authors. As mentioned before, the administration of postoperative RT might explain the increased risk of STs. Welsch [8] and Pan [9], even if their studies included a relatively small number of patients, concluded that this theory was not supported by their data. Despite these observations, the carcinogenicity risk of ionizing radiations is well known [22]: the standard adjuvant RT treatment after thymectomy consists in the administration of Gy to the mediastinum and may be significant for several important sites as bone marrow, oesophagus, lung, heart and stomach; however, previous reports did not find a significant excess of second cancers in many of these sites [22]. Our results do not confirm that adjuvant therapy may increase the risk of developing STs (HR 1.15; 95% CI ). The high number of patients with two or more primary extrathymic cancers reported confirms a sort of oncogenetic tendency in these patients; Friedmann [23] reported an occurrence

5 P.L. Filosso et al. / European Journal of Cardio-Thoracic Surgery 223 of a third neoplasm in >30% of patients with a clinical history of thymoma and haematologic cancers. An intrinsic immune abnormality of which thymoma may be a marker is suggested by the finding of an increased risk of extrathymic tumours before thymoma diagnosis along with a high incidence of autoimmune disorders, the most common of which is MG. As Welsh suggests [8] that extrathymic STs are true second cancers rather than cancers related to thymoma treatment. The awareness of this intrinsic risk should suggest an appropriate clinical surveillance with a lifetime follow-up, especially in case of aggressive thymic tumours [8, 12]. The differences in STs type and site may reflect a nonhomogeneous factor directly depending on the population characteristics. In other words, it is possible that different populations may have different risks of developing different cancers. This may explain why in Pan s study, based on the Taiwan population, the distribution of STs was different when compared with that of Gadalla, including a Swedish population, or our Italian study. The inhomogeneity of the USA population shown in the SEER-based study explains the differences between STs recorded in Engels, Travis and Weksler series and those coming from single-institution studies. Our report also supported that the association with MG seems to be protective against second cancers, although this finding did not reach statistical significance, probably due to the low power of the study. Owe [24], investigating the Norwegian Cause of Death Registry, observed that patients with MG show a lower cancer rate as cause of death when compared with the normal population. In fact, the STs rate in patients with thymoma was higher in those without MG, suggesting an intriguing MG immunological protective effect, possibly due to an antitumoral autoimmunity. An autoimmune system alteration associated with thymoma might be the key to interpret and explain the increased STs risk of subsequent cancers. This could also eventually explain the fascinating MG protective effect on STs development. A T-cell development alteration within the thymus is thought to be related to the progress of autoimmune disorders in patients with thymoma [6]. For this reason, we agree with other authors [8] on the potential role of a thymic intrinsic factor in the development of STs, supporting the theory that they are true second cancers rather than treatment-related tumours. In conclusion, although thymic tumours are considered relatively indolent, an increased risk of developing second cancers is confirmed, in particular for aggressive tumours. An intrinsic immune system abnormality certainly plays a crucial role, and the onset of thymoma might be considered a marker of it. We confirm the potential MG protective effect against STs development. Further studies are mandatory to assess the real immune system role in this process. ACKNOWLEDGEMENTS The Authors thank Andrea Evangelista, Anna Castiglione and Laura Bergamasco for their support in the statistical analysis. A particular gratitude goes to Alberto Oliaro and Giovannino Ciccone for their help in the manuscript project design and to Filippo Lococo for his valuable work on the database design. Finally, thanks to Alberto Sandri for his linguistic revision. Conflict of interest: none declared. REFERENCES [1] Detterbeck FC, Parsons AM. Thymic tumors. Ann Thorac Surg 2004;77: [2] Venuta F, Anile M, Diso D, Vitolo D, Rendina EA, De Giacomo T et al. Thymoma and thymic carcinoma. Eur J Cardiothorac Surg 2010;37: [3] Engels EA, Pfeiffer RM. Malignant thymoma in the United States: demographic patterns in incidence and associations with subsequent malignancies. Int J Cancer 2003;105: [4] Marx A, Hohenberger P, Hoffmann H, Pfannschmidt J, Schnabel P, Hofmann HS et al. The autoimmune regulator AIRE in thymoma biology. Autoimmune and beyond. J Thorac Oncol 2010;5:S [5] Strobel P, Helmereich M, Menioudakis G, Lewin SR, Rudiger T, Bauer A et al. Paraneoplastic myasthenia gravis correlates with generation of mature naïve CD4 + T cells in thymomas. Blood 2002;100: [6] Okumura M, Fujii Y, Shiono H, Inoue M, Minami M, Utsumi T et al. Immunological function of thymoma and pathogenesis of paraneoplastic myasthenia gravis. Gen Thorac Cardiovasc Surg 2008;56: [7] Mueller-Hermelink HK, Marx A. Thymoma. Curr Opin Oncol 2000;12: [8] Welsh JS, Wilkins KB, Green R, Bulkey G, Askin F, Diener-West M et al. Association between thymoma and second neoplasms. JAMA 2000;283: [9] Pan CC, Chen PC, Wang LS, Chi KH, Chiang H. Thymoma is associated with an increased risk of second malignancy. Cancer 2001;92: [10] Travis LB, Boice JD, Travis WD. Second primary cancers after thymoma. Int J Cancer 2003;107: [11] Yen YT, Lai WW, Wu MH, Lin MY, Chang JM, Hsu IL et al. Thymic neuroendocrine carcinoma and thymoma are both associated with increased risk of extrathymic malignancy: a 20-year review of a single institution. Ann Thorac Surg 2011;91: [12] Weksler B, Nason KS, Mackkey D, Gallagher A, Pennathur A. Thymomas and extrathymic cancers. Ann Thorac Surg 2012;93: [13] Gadalla SM, Rajan A, Pfeiffer R, Kristinsson SY, Bjorkolm M, Langden O et al. A population-based assessment of mortality and morbidity patterns among patients with thymoma. Int J Cancer 2011;128: [14] Levin N, Abramsky O, Lossos A, Karussis D, Siegal T, Argov Z et al. Extrathymic malignancies in patients with myasthenia gravis. J Neurol Sci 2005;237: [15] Owe JF, Cvancarova M, Romi F, Gilhus NE. Extrathymic malignancies in thymoma patients with and without myasthenia gravis. J Neurol Sci 2010;290:66 9. [16] Marx A, Strobel P, Zettl A, chan JKC, Harris NL, Kuo TT. Thymomas. In: Travis WD, Brambilla E, Muller-Hermelink HK, Harris CC (eds). World Health Organization Classification of Tumours. Pathology and Genetics of Tumours of the Lung, Thymus and Heart. Lyon: IARC Press, 2004, [17] Koga K, Matsuno Y, Noguchi M, Mukai K, Asamura H, Goya T et al. A review of 79 thymomas: modification of staging system and reappraisal of conventional division into invasive and non-invasive thymoma. Pathol Int 1994;44: [18] Ruffini E, Filosso PL, Mossetti C, Bruna MC, Novero D, Lista P et al. Thymoma: inter-relationships among world Health Organization histology, Masaoka staging and myasthenia gravis and their independent prognostic significance: a single-centre experience. Eur J Cardiothorac Surg 2011;40: [19] Souadjijan JV, Silverstein MN, Titus JL. Thymoma and cancer. Cancer 1968;22: [20] Masaoka A, Yamakawa Y, Niwa H, Fukai I, Tokudome S, Nakahara K et al. Thymectomy and malignancy. Eur J Cardiothorac Surg 1994;8: [21] de Jong WK, Blaauwgeers JLG, Schaapveld M, Timens W, Klinkenberg TJ, Groen HJM. Thymic epithelial tumours: a population-based study of the incidence, diagnostic procedures and therapy. Eur J Cancer 2008;44: [22] Boice JD Jr, Land CE, Preston DL. Ionizing radiation. In Shottenfeld D, Fraunemi JF Jr (eds). Cancer Epidemiology and Prevention, 2nd edn. New York: Oxford University Press, 1996, [23] Friedmann HD, Inman DA, Hutchison RE, Poiesz BJ. Concurrent invasive thymoma and T-cell lymphoblastic leukemia and lymphoma: a case report with necroscopy findings and literature review of thymoma THORACIC

6 224 P.L. Filosso et al. / European Journal of Cardio-Thoracic Surgery and associated hematologic neoplasms. Am J Clin Pathol 1994;101: [24] Owe JF, Daltveit AK, Gilhus NE. Does myasthenia gravis provide protection against cancer? Acta Neur Scand 2006;113:33 6. APPENDIX. CONFERENCE DISCUSSION Dr M. Okumura (Osaka, Japan): I enjoyed reviewing Dr Filosso s paper and studying the relationship between thymoma and extrathymic malignancies. This is an important subject, because thymoma has some immunological function, and resection of thymoma and thymus might affect the immunological surveillance of newly emerging cancer cells. I remember that Emeritus Professor Akira Masaoka was also interested in this subject and surveyed the patients in Osaka University and Nagoya City University in the days I was a postgraduate student at Osaka University. Dr Masaoka reviewed 390 myasthenia gravis patients, and he found that the presence of thymoma facilitates the occurrence of extrathymic malignancy but, at the same time, he suggested that thymectomy never enhances occurrence of malignancy. Dr Filosso s study also presents a high risk of developing extrathymic malignancies in thymoma patients, consistent with Dr Masaoka s study, but I have some things to discuss. Patients with advanced stages, with a histological high grade, and undergoing adjuvant radiotherapy or chemotherapy were shown to have higher risk of extrathymic malignancy. I imagine that these patient groups are overlapping each other. If so, the influence of adjuvant therapy on the second malignancy should be critically excluded. So how do you explain this? And what was the risk of a second malignancy in the patients who did not receive postoperative adjuvant therapy? The next question concerns the incidence of lung cancer which is relatively high. Is there a background of more smokers in this group compared with the general population? Before and after operation for thymoma, patients are supposed to experience imaging examinations or chest x-ray and CT more often than other populations. These regular examinations might have increased the chances of finding lung cancers. If these lung cancer patients are in the earlier stage, this might be explained by extensive examination of the chest, and this might be a reason for finding more lung cancers than malignancies of other organs. In connection with that, I think it is important to deny the possibility that pre- and postoperative systemic imaging screening helped early detection of malignancies in other organs also. So I would like to show the stages of other malignancies. (Slide) As the authors state, thymoma has a function of inducing CD4+ and CD8+ so-called double positive T cells. This implies that thymoma has a function to differentiate T cell precursors from mature T cells even in a neoplastic environment. The slide presents quantitative evaluation of immunological function of thymoma and thymic carcinoma survey in terms of the number of CD4+ and CD8+ double positive T cells. Either type A thymoma or thymic carcinoma hold little or no such function. Type B3 thymoma also has a very weak function. I would like to know whether the T cell differentiation function of thymoma has an impact on giving rise to a second malignancy. So is there any difference in second malignancy between type AB, B1, and B2 functioning tumours and type A, B3, and C less functioning tumours? (Slide) The possible explanation for pathogenesis of thymoma-associated paraneoplastic autoimmune disease is presented. In brief, thymoma has an immunological function very similar to thymic cortex. Unfortunately, thymoma neoplastic epithelial cells do not express HLA-DR molecules at an identical level to the normal thymus, resulting in the change of T cell affinity of positively selected T cells in thymoma. In addition, thymoma lacks medullary structure and expression of AIRE protein, as the authors point out. This results in insufficient negative selection. These situations might cause differentiation of self-reacting T cells and development of autoimmunity. In this paper, the authors suggested that myasthenia gravis is a protective factor. If so, my final question is how the authors relate T cell development or immunological function to cancer immunology in thymoma patients. Dr Filosso: Thank you for reminding us of Professor Masaoka, who represents one of the milestones in the history of thymoma surgery and was a great friend of my former chief, Professor Giuliano Maggi. The number of questions I have to answer is very impressive. Moving to the first one, as I stated in my manuscript, the higher risk of developing a second malignancy is in those patients who are in an advanced stage and with a high WHO grade tumour, and are usually those who are submitted to adjuvant treatment. So I agree with you when you say that they may overlap each other. However, the effect of the postoperative treatment, which is commonly radiotherapy, may not be excluded at all, especially when we observe the presence of patients with lung or thyroid cancer, which occurred in the field of radiation delivery. Another point to mention is the different sites of second malignancy in our series when compared to others previously published in the literature. I tried to find a possible explanation, and I agree with those authors who hypothesized the presence of a kind of nonhomogeneous factors. In other words, this means that a different population may have a different risk of developing a different second cancer, and this may explain how our series is different in terms of second tumour sites compared to the series of Pan, Welsh and Weksler, and to Masaoka s series. Patients in advanced stage generally receive a more accurate follow-up. Our study is a retrospective multicentric one, and may suffer from some bias. One may be the type of follow-up, because trying to be as homogeneous as possible, both the type of radiological procedures and their timing may vary among the single institutions. Anyway, as I pointed out in the manuscript, a stricter follow-up of these patients may increase the chance of detecting second malignancies. One suggestion may be to use a different kind of population as a control group. In this paper we used the normal Italian population, the data collected by AIRTUM (the Italian National Cancer Registry), but if we want to demonstrate the role of follow-up in the detection of a second tumour, another control population might be used (e.g. patients with a non-autoimmune cancer, I mean breast or colon cancer, who are controlled in a similar aggressive way to our patients. As you pointed out Dr Okumura, we observed an excess of lung cancers. To be very honest with you, data about smoking were not collected and a clear answer to your question regarding possible heavy smokers in our population is not possible. Anyway, one may also speculate that thymoma patients are not such heavy smokers as those with lung or other smoking-related cancers (e.g. oesophageal or laryngeal). As Dr Okumura pointed out, early stage second tumours and cancers in so-called biological thymoma (B1 and B2) were observed also in our experience. Finally, the possible role of autoimmune disorder in the carcinogenetic process may be, of course, one cause of second tumour development, because interleukin 12 and interferon-alpha are two of the most biologically active cytokines involved in the carcinogenetic process, and the production of autoantibody against interferon-alpha or interleukin 12, as Dr Okumura stated in his article cited in my paper s references, may be one of the possible causes to explain the abatement in the normal anti-tumoral defence of the human body.