Thymoma is known as the most common tumor arising in the anterior-superior mediastinum^1,2). The difficulty that investigators face is diffenentiating between benign and malignant thymoma^1–5). Although some investigators suggest that histologic classifications have an influence on the survival of patients with thymoma, histologies of the two types have no distinguishable difference. Numerous papers about prognostic factors of thymoma according to clinical parameters or treatment modality have been reported. The prognostic factors in clinical parameter or treatment modality have been reported. The prognostic factors in clinical parameters are the degree of invasion (stage), age and associated autoimmune disease. It is known to be important whether complete resection is done or not^6–14).

There are several Investigators who argue that complete resection played an important role in the prognosis, so they advocate aggressive surgery for complete resection in invasive thyoma. whenever possible, even in stage IVA^6,7,15,16). However, the role of radiotherapy is also an important therapeutic modality for unresectable invasive thymoma⁶⁾ as well as for post-operative combination therapy. Besides, several drugs have been reported to have anti-tumor activity in the thymoma^17,18) and, recently, combination chemotherapy is shown to play an important role in the treatment of inoperable invasive thymoma^19–22).

However, there are only a few papers written on prognostic factors in Korea^23–25). This study attempted to evaluate the following items to analyse the prognostic factors influencing survival of patients with thymoma.

Between April 1981 and July 1994, 100 patients with histologically proven thymoma were treated at the Seoul National University Hospital. Thymic carcinomas were excluded from this study. Surgical biopsies were done in 85 patients and aspiration cytologies were done in 15 patients with invasive thymoma. These fifteen patients were previously determined as inoperable by their computed tomography results. The invasiveness of each tumor was evaluated according to the clinical staging procedures of Masaoka and associates. Stage I thymomas were completely encapsulated without capsular invasion. Stage II showed microscopic or macroscopic invasion into the capsule or macroscopic invasion of mediastinal fatty tissue or of mediastinal pleura. Stage III displayed macroscopic invasion of pericardim, greater vessels or lung. Stage IVA had pleural or pericardial dissemination. Stage IVB had lymphogenous, hematogenous metastasis, or both. The degree of invasion was determined by CT in the fifteen patients who had been diagnosed with aspiration cytology Stage III, 9, Stage IVA, 4, stage IVB, 2. Stage I thymoma patients were treated by thymectomy alone, Stage II-thymectomy with or without postoperative radiotherapy, stage III-themectomy and postoperative radiotherapy with or without chemotherapy, Stage IVA-radiotherapy and/or chemotherapy, stage IVB-chemotherapy with or without radiotherapy. CAP chemotherapy (cyclophosphamide 700 /m², doxorubicin 50mg/m², and cisplatin 50/mg²: intravenous infusion on day 1) was administered to the eligible patients. Treatment cycles were repeated every 21 days. Patients were evaluated for response after two cycles of therapy. For patients with responsive or stable disease, treatment was continued unless grade 4 toxicity or tumor progression was noted. When complete remission was achieved, the chemotherapy was finished after two more cycles. Survival rates were computed by the Kaplan-Meier method and statistical differences were tested by a log-rank method. Multivariate analysis of prognostic factors was done using the Weibull model. A probability value, less than 5%, was regarded as significant.

Neoplasms of the thymus differ in histogenesis. Epithelial, lymphoid, neuroendocrine, germ cell, and mesenchymal tumors have been descrived^2,31–34). According to the definiton given by Rosai and Levines,^2,5) thymomas are tumors arising from the thymic epithelial cell. The traditional histological grading based on the lymphocyte/epithelial cell ratio (predominantly lymphocytic, predominantly epithelial or mixed) has become insufficient because this classification has little prognostic value^{4,28,36–37)}. In 1985, Marino and Muller-Hermelink proposed a new classification system the different subset of thymic epithelial cells was suggested^38,39). Cortical thymomas were related morphologically and phenotypically to thymic cortical epithelial cells and medullary thymomas were related to medullary epithelial cells. Mixed thymomas were also discribed and further divided into three subgroups : the mixed common type, mixed with cortical predominance, mixed with eddullary predominance and sometimes arbitrary. In some series, the survival data showed signifiant differences in prognosis among three main types^40–42). The medullary thymoma is comparatively rare, benign tumor, arising late in life and usually not associated with MG⁴⁰⁾. Park et al. showed no statistical correlation of invasiveness among the three types. Three cases showed that the medullary thymoma were all non-invasive; these cases were, however, too small to compare statistical significance²⁴⁾.

The thymoma represents seventy-five percent of the primary tumors in the anterior mediastinum and it is one of the most frequently found neoplasms in this compartment¹⁾. The patients suffering from thymoma are usually between 40 and 60 years of age, and there is no sex predilection. The chief complaints in patients with thymoma were as follows : abnormal findings on the chest X-ray without symptoms, thoracic symptoms, symptoms of myasthenia gravis etc. We found similar clinical features.

The rapid increase in the number of the patients with thymomas seen after 1974 is likely due to the radical change in the treatment policy (removal of thymus) of myasthenia gravis. This treatment policy may have caused the increased discovery of smaller-sized thymomas that would not have been detected in other ways. In this study, 50% of the thymoma patients were in Stage I. The incidence of myasthenia gravis in patients with thymoma has ranged from 7 to 54% in various series, with a mean of approximately 35%⁴³⁾.

Thymomas are found in one tenth of the patients with myasthenia gravis^{7,11,43, 44)} A few cases of MG developing after operation are described. We observed five cases in which MG developed after treatment. The presence of residual thymic tissue locally or in extramediastinal sites, or the relapse of the tumor, have been suggested to explain re-explored because of newly diagnosed MG (two cases) or significant worsening of neurological symptoms (one case), had no evidence of disease at operation⁷⁾. In our series, it was speculated that myasthenic symptoms were associated with the presence of thymic tissue because 5 patients, who were newly diagnosed as having MG, had residual disease or relapse. Even though MG may have been the cause of death, the presence of MG was not a poor prognostic factor in thymoma patients in our study. This conflicted with the reports of the 1960s and early 1970s. Bernatz¹⁴⁾ ovserved that MG did not affect survival if the thymomas were invasive but survival was decreased if the tumors were non-invasive. However, some studies reported a higher survival rate in MG pateints^14,32,48,47. They explained that the earlier diagnosis of thymoma (permitted by the development of MG symptoms and the improvement in treatment of MG) might account for this finding. We calculated the difference of the survival rate in invasive and non-invasive group separately according to the presence or absence of MG, but found no survival difference.

About 40% of the patients had thoracic symptoms and the remainder of such tumors are usually detected on routine chest radiography¹¹⁾. Paraneoplastic phenomena, other than MG, occurred with thymoma in about 10% of the patients^11,48,49). These disorders largely consisted of autoimmune-related syndromes or hematologic abnormalities. Thymoma-associated systemic syndromes, other than MG, were ovserved in 2 cases. One case was pure red cell aplasia and the other was systemic lupus erythematosus. Pure red cell aplasia are found in about 5% of patients with thymomas, and thymectomy induced remissions of the disease in 38% of the patients. However, our case did not show remission, even after thymectomy¹⁾.

The classification of Masaoka and associates based on the extent of tumor invasion is valuable. Survival curves show that locally invasive thymomas have a less favorable outcome compared to encapsulated forms. When intrathoracic disseminations or distant metastasis are found, survival rates decrease significantly. Even though we found no survival difference between intrathoracic dissemination (Stage IVA) and metastasis (Stage IVB), staging was found to be the most important prognostic factor in our study. Wilkins⁵⁰⁾ argued that patients with pathologic Stage I thymoma, although they merit careful follow-up, need no adjuvant therapy and have an excellent prognosis. For stage II thymoma, it is generally agreed that the best treatment is an operation, followed by radiotherapy²¹⁾. We agree with this treatment strategy, even though the results of our study were too small to draw definitive conclusion. For stage III, radical operation followed by radiotherapy is considered the treatment of choice²¹⁾. Several investigators argue that complete resection play an important role in the prognosis and they advocate aggressive surgery (complete resection) in invasive thymoma, even in stage IVA^6,7) But we found no stastistical difference in the group of thymoma patients in stage III, whether complete resection was done or not. Instead, we found early death and high relapse rates (44%). Radiotherapy was recommended in unoperable cases, but even when patients had a complete remission, half of them relapsed. In Stage III, we were faced with the need of systemic therapy because the systemic rather than local relapse was more common. This need occurred even though patients had complete remission by local modality treatment (complete resection and post-operative RT or RT alone). At present, combination chemotherapy is used after incomplete resection or after incompleter resection and radiotherapy. The overall response rate to combination chemotherapy was about 50% and complete response was about 10–40%^21,22,51,52). Our study showed similar results. In this study, the patients, who had relapsed after local modality treatment, had systemic CAP chemotherapy. Radical resection was difficult in some cases due to locally invasive lesion. In these cases, some authors have suggested that neoadjuvant chemotherapy followed by surgery may be useful to reduce the size and degree of the infiltration of the lesion, thus increasing the probability of radical operation²¹⁾. This therapy may be helpful to pervent systemic relapses, but there is no Phase III randomized controlled study.