Prognostic Factors in Thymoma

The issue of determining the factors that affect clinical behavior in thymoma has long been a subject of debate. Currently, controversy still exists regarding which are the most reliable prognostic features of these tumors. Traditionally, estimation of prognosis in most tumor systems has been dependent on the assessment of a variety of clinical parameters as well as grading and staging of the lesions. Moreover, as chemotherapy protocols and new refinements in treatment modalities become available, a greater demand has been created by surgeons and oncologists for identifying additional parameters, such as hormonal status, expression of specific tumor-associated gene products, determination of proliferation index, DNA content, ploidy, cytogenetic abnormalities, expression of oncogenes or detection of selected mutations by molecular genetic techniques that will permit stratification of patients into more sophisticated therapeutic schemes.

Because of the rarity of thymoma and the difficulties for initiating prospective studies, little progress has been made in the evaluation of some of these newer modalities in these tumors. A large number of retrospective clinicopathologic studies reported in the literature have addressed the possible role of several clinical parameters in the clinical behavior of these lesions. For the most part, the majority of such publications have been hindered by the limited number of cases studied and by the retrospective nature of the analysis. Most studies, however, have found a lack of any statistical correlation between age, sex, clinical symptoms or mode of presentation.^(1-13) Size of the tumor was found in some studies to directly correlate with prognosis by univariate analysis,^(4,12) but this has not been substantiated by others. The association with myasthenia gravis and other autoimmune disorders has also been a topic of intensive study in these tumors. Although initially it was believed that thymoma patients with myasthenia gravis (MG) were associated with a worse prognosis,^(1,14,15) most recent studies have shown that this is not the case.^(2-13) The poorer survival reported in the older literature is most likely related to myasthenia-related perioperative tumor deaths rather than to the effects of the tumor itself. In fact, most recent studies have demonstrated that mortality, recurrence rate, and overall survival are not affected by the disease or show a somewhat better prognosis than non-myasthenic patients.^(5,6,8) This has been attributed to earlier diagnosis of the tumor and improved treatment for the disease in myasthenic patients.

The role of flow cytometry for the assessment of prognosis in thymomas has also been reported in several studies but remains inconclusive ^(16,17). Similarly, proliferative activity of tumor cells as determined by immunohistochemical staining of proliferating cell nuclear antigens (PCNA) or proliferation-associated nuclear antigen (Ki-67) have also been studied in thymoma, although no clinically applicable results have been obtained so far.⁽¹⁸⁾ Oncogene alterations have also been recently addressed in these tumors. One study found high expression of bcl-2 in invasive thymomas and thymic carcinomas, and lack of correlation between p53 expression with the histological subtypes and aggressive behavior.⁽¹⁹⁾ A subsequent study, however, showed an inverse relationship between bcl-2 and p53 expression, with p53 being more often expressed in invasive tumors, and bcl-2 being more often found in spindle cell thymoma.⁽²⁰⁾ Because of the limited number of cases studied so far, additional studies will be needed to determine the role of genetic alterations for the prognosis of thymoma.

The two parameters that the majority of studies appear to regard as being the most important for the assessment of prognosis in thymoma are staging and histology. Controversy, however, still exists regarding the relative value of the two, with opinions being divided between those who believe that staging is the only significant predictor of prognosis for these tumors, and those who argue that histology also represents an independent prognostic factor, with conflicting results often being reported in the literature. Assessment of these parameters has been further complicated by the lack of uniformity in the histologic interpretation and classification of these tumors among the various studies published so far. With the recent publication of the WHO schema for the classification of thymic tumors, a new and important tool has been provided that is able to facilitate comparison among the different studies published in the literature.⁽²¹⁾

The present review presents a retrospective meta-analysis of some of the most significant large studies of thymoma reported in the literature between 1973-1996, in an attempt to identify general trends shared by the various studies that could shed some light on the clinical behavior of these tumors (Table I).

Table I: Meta-analysis of Clinicopathologic Studies of Thymoma - 1973-1996:

Bernatz et al, Surg Clin N A, Vol.53, 1973 (Mayo Clinic, Rochester, MN, USA) Masaoka et al, Cancer, Vol.48, 1981 (Nagoya University, Japan) Verley and Hollman, Cancer, Vol.55, 1985 (Natl. Inst. Med. Research, France) Lewis et al, Cancer, Vol.60, 1987 (Mayo Clinic, Rochester, MN, USA) Nakahara et al, J Thor C-V Surg, Vol.95, 1988 (Osaka University, Japan) Pescarmona et al, Ann Thorac Surg, Vol.50, 1990 (University of Rome, Italy) Maggi et al, Ann Thorac Surg, Vol.51, 1991 (University of Torino, Italy) Wilkins et al, Ann Thorac Surg, Vol.51, 1991 (Mass. Gen. Hosp., Boston, USA) Park et al, Cancer, Vol.73, 1994 (M.D. Anderson Cancer Center, Texas, USA) Pan et al, Hum Pathol, Vol.25, 1994 (National Medical College, Taiwan) Quintanilla et al, Cancer, Vol.74, 1994 (Mass. Gen. Hosp., Boston, USA) Blumberg et al, Ann Thorac Surg, Vol.60, 1995 (Memorial Sloan-Kettering, USA) Regnard et al, J Thorac C-V Surg, Vol.112, 1996 (Lannelongue Hosp., France)

Care was taken to select only those studies in which similar parameters were addressed, and for which similar information was provided that would be amenable for comparison with one another. Although as many as 4 different histologic classifications were used in the various studies, the WHO schema permits easy adaptation and comparison for equivalence among the various studies (Table II).

Table II: Comparison of various histologic classifications with the WHO schema:

WHO schema	Traditional (Barnatz)	Muller-Hermelink	Verley & Hollman	Masaoka
Type A	Spindle cell	Medullary	Type I (spindle)	Spindle cell
Type AB	------	Mixed	------	------
Type B1	Predominantly lymphocytic	Predominantly cortical (organoid)	Type II (lymphocyte-rich)	Lymphocytic
Type B2	Mixed lymphoepithelial	Cortical	------	Lymphoepithelial
Type B3	Predominantly epithelial	Well-differentiated thymic carcinoma	Type III (differentiated epithelial thymoma)	Polygonal/clear cell
Type C	Thymic carcinoma		Type IV (undifferentiated epithelial thymoma)

Unlike most other tumor systems for which assessment of prognosis generally correlates with both the histology and the staging of the lesion, thymomas have posed difficulties for utilizing this approach. It was traditionally claimed that histology was a poor predictor of prognosis for these tumors, and that prognosis could only be reliably assessed based on the staging of the lesions.^(1-4,15) In more recent years, proponents of the histogenetic classification of thymoma introduced by Marino and Muller-Hermelink have claimed that histologic classification represents an independent prognostic factor for the evaluation of these tumors.^(6,11,22,23) The truth most likely lies somewhere in-between, and it will be postulated here that rather than relying on one or the other, the best approach to the evaluation of these tumors should incorporate a combined assessment of several rather than a single parameter.^(24-26)

The present meta-analysis comprised a total of 2050 patients spanning a period of 23 years (1973-1996). Overall 5, 10 and 15-year survival for the whole group was 75, 61 and 47%, respectively. Survival for non-invasive (encapsulated) tumors at 5, 10 and 15-years was 83, 75 and 50%, respectively, compared with a survival for invasive tumors of 49, 30 and 12%, respectively (Fig. 1).

Survival was then analyzed in relation to Masaoka staging and histologic types. Survival at 5, 10 and 15 years by Masaoka stage progressively decreased with higher stages (Fig. 2). Note, however, that the survival curves for stages I and II (that is, tumors that are essentially confined within the mediastinum) are very close and do not show a statistically significant difference, whereas a sharp drop-off occurs between these stages and stages III-IV (frankly invasive and metastatic tumors).

Analysis of histologic types (Table II), including WHO type A (spindle/medullary thymomas), WHO type B1 and 2 (lymphocytic, mixed lymphoepithelial, cortical, organoid), and type B3 (epithelial-rich, polygonal, squamoid, atypical thymoma, well-differentiated thymic carcinoma) showed the following survival curves (Fig. 3):

Although statistically significant differences could not be observed between these various histologic subtypes, predominantly epithelial tumors (excluding spindle cell thymoma) showed a slightly more aggressive behavior. Interestingly, in two of the studies in this group,^(1,3) when the authors compared the survival curves for the different histologic types according to stage it was noted that all tumors tended to behave very similar by stage irrespective of histology.

One additional parameter that was found to be of significance in several of the studies was the extent of the surgical resection.^{(2,5,7,12,13)} This feature was analyzed in 5 of the 13 studies, and showed that survival directly correlated with the status of resectability of the lesion at the time of initial diagnosis (Fig. 4).

Interestingly, in one of these studies,⁽⁵⁾ when survival curves for the different histologic types were plotted based on the extent of the surgery it was noted that no significant differences could be observed for the different histologic types in patients who underwent complete surgical excision, whereas there was a sharp drop in survival for patients undergoing subtotal resection or biopsy only, irrespective of the histology.

Some of the conclusions that we're able to gain from this meta-analysis are that in thymoma, at least three parameters appear to play a significant role in determining prognosis for these tumors: histology, staging, and extent of surgical excision. The results of most studies to date clearly support the fact that thymoma of predominantly round/polygonal (non-spindle cell) histology appear to behave in a biologically more aggressive fashion. Thymomas of predominantly spindle cell type appear to observe the most indolent behavior, and lymphocyte-rich and lymphoepithelial thymomas seem to fall somewhere in-between.

Care should be taken, however, not to mistake this for an indication that all spindle cell thymomas are universally benign and that all epithelial-rich thymomas are universally aggressive. Studies have clearly indicated that when these various histologic subtypes are separated according to stage, they tend to observe a similar biologic behavior. In recent years, some studies applying the Kirschner and Muller-Hermelink classification have made the claim that all "medullary" (i.e., spindle cell) and mixed thymomas are universally benign and have no capacity for aggressive behavior. The literature and the experience of others do not support this. Numerous examples are on record of spindle cell thymomas that recurred, metastasized, and otherwise demonstrated aggressive behavior.^(27-29) For this reason, we regard all thymomas as potentially aggressive malignant neoplasms and feel that proper evaluation of these tumors requires careful examination of the resected surgical specimen for proper staging.⁽²⁵⁾

The fact that when adjusted for clinical stage, the various histologic categories of thymoma tend to display a similar clinical outcome, and the fact that when adjusted for the extent of surgery (i.e., complete vs. incomplete excision) they will also show similar survival curves, is clearly indicating that the two latter parameters play a more important role in the evaluation of these tumors than histology alone. Staging and extent of resectability thus become two critical factors for the assessment of prognosis in these tumors.

Staging in thymoma has also been the source of controversy in the literature. The most widely accepted staging system for thymoma was the one proposed by Masaoka et al in 1981.⁽²⁾ A recent study by Koga et al,⁽³⁰⁾ however, introduced a modification to the Masaoka staging system whereby Masaoka's stage II-2 (microscopic invasion into the capsule) was included under stage I, based on the fact that survival curves revealed that the most important factor associated with the prognosis of thymoma was invasiveness of the tumor into surrounding organs and structures. The authors thus proposed that the staging system for thymomas should be simplified and divided into non-invasive (or circumscribed or confined to within the thymus) and frankly invasive.

Another important modification to the staging of thymoma was introduced by Regnard et al⁽¹³⁾ in a clinicopathologic study of one of the largest series of thymoma published to date. These authors remarked on the importance of the extent of surgical resection of the tumor for prognosis, and proposed that this factor also be incorporated into the clinicopathologic staging of these tumors (Table III).

Table III: Modified Staging of Thymoma – Regnard et al, 1996⁽¹³⁾

Stage Ia:	Complete resection of completely encapsulated thymoma with no microscopic capsular invasion.
Stage Ib:	Complete resection of completely encapsulated thymoma with peritumoral adhesions to surrounding structures.
Stage IIa:	Complete excision of thymoma with gross invasion into surrounding fat or mediastinal pleura.
Stage IIb:	Complete excision of thymoma with microscopic invasion into the capsule.
Stage IIIa:	Complete excision of thymoma with gross invasion of neighboring organs (pericardium, great vessels, lung).
Stage IIIb:	Incomplete excision or biopsy of thymoma with gross invasion of neighboring organs.
Stage IVa:	Complete resection of metastatic thymoma.
Stage IVb:	Incomplete excision of metastatic thymoma.

Although we do not necessarily endorse the above staging scheme in its present form, we do believe that the staging of thymoma requires urgent revision and incorporation of the information regarding extent of resection of the tumor. Likewise, a system of prognostication in thymoma needs to also incorporate in some form the information provided by the histology and staging of the lesions.

The issue of histologic classification of thymoma also remains controversial. The recently introduced WHO Schema for the classification of thymic epithelial neoplasms has come to fill an important void for the study of these tumors by permitting adequate comparisons among the various existing classifications and nomenclatures.⁽²¹⁾ By the admission of its own authors, however, this schema was neither meant to replace existing terminology nor was it proposed as a "new" classification.⁽²¹⁾ Moreover, the combined use of letters and numbers for the designation of pathologic entities is a quite unorthodox method of making diagnoses that can lead to much confusion for pathologists and clinicians alike.

On the same year that the WHO schema was proposed, we introduced a novel proposal for the classification of thymic epithelial neoplasms based on morphologic grades of differentiation.⁽²⁵⁾ The purpose of this proposed classification was not only to provide a conceptual functional and morphologic basis for the histological classification of these tumors, but also to considerably simplify the issue of morphologic reporting and classification of these lesions. Essentially, the proposed schema divides thymic epithelial neoplasms into 3 categories depending on the degree of preservation or absence of the organotypical features of thymic differentiation: 1) well-differentiated thymic epithelial neoplasms, designated by convention "thymoma"; 2) moderately-differentiated thymic epithelial neoplasms, designated as "atypical thymoma"; and 3) poorly-differentiated thymic epithelial neoplasms, designated by convention "thymic carcinoma".

The most significant change introduced by this classification was essentially that of eliminating the need for further subcategorization of the various histologic subtypes of well-differentiated thymic epithelial neoplasms (Table IV). The results of the current meta-analysis of the old literature, as well as several more recently published studies^{(11,13,31,32)} have corroborated this notion. In fact, the most recent study by Chalabreysse et al,⁽³²⁾ applying the new WHO schema, concluded from their multivariate analysis that only the histologic classification between types A and B as a group, as opposed to type C, remained statistically significant.

Table IV: Classification of Thymic Epithelial Neoplasms According to Grades of Differentiation,
Suster & Moran, 1999.⁽²¹⁾

Tumor type	Grading	Equivalent terms
Thymoma	Well-differentiated	-Thymoma type A, AB, B1, B2 (WHO Schema) -Cortical, medullary, mixed, organoid thymoma (Kirchner and Muller-Hermelink) -Lymphocyte-rich, mixed, spindle cell thymoma (Barnatz)
Atypical thymoma	Moderately-differentiated	-Thymoma type B3 (WHO Schema) -Well-differentiated thymic carcinoma (Kirchner and Muller-Hermelink) -Epithelial-rich thymoma (Barnatz)
Thymic carcinoma	Poorly-differentiated	-Type C thymoma (WHO Schema) -"Other" thymic carcinoma (Kirchner and Muller-Hermelink) -No equivalent (Barnatz)

We believe that there is definitely merit in separating the WHO type B3 (atypical thymoma in our classification; well-differentiated thymic carcinoma in the Kirchner/Muller-Hermelink classification) from the other types of thymoma because they are more often associated with invasion and present at more advanced stages at the time of diagnosis, and because biologically, they also represent a more advanced stage of tumor progression than better-differentiated thymic epithelial neoplasms.^(24-27) However, for the purpose of clinical evaluation this must be tempered with the other two important prognostic parameters, staging and status of resectability.

Based on these observations, we recently proposed a modified approach for the assessment of prognosis in thymic epithelial neoplasms (excluding thymic carcinoma) that takes into account all of the above factors⁽³³⁾ (Table V). We believe this combined approach should facilitate more reliable assessment of prognosis and clinical behavior in these tumors. Likewise, the adoption of a more simplified histologic classification system should allow pathologists in common practice to provide clinicians and surgeons with a simple and easily reproducible means of assigning patients to the various prognostic categories for the selection of treatment.

Table V: Prognostic Categories for Thymic Epithelial Neoplasms (excluding thymic carcinoma)⁽³³⁾

Favorable Prognostic Categories	Unfavorable Prognostic Categories
Group I: Encapsulated or minimally invasive thymoma Completely excised Histologic types A, AB, B1, B2 (thymoma)	Group IV: Widely invasive thymoma or with implants Incomplete excision All histologic types
Group II: Encapsulated or minimally invasive thymoma Completely excised Histologic type B3 (atypical thymoma)	Group V: Widely invasive thymoma w/wo metastases Unresectable/biopsy only All histologic types
Group III: Widely invasive thymoma or thymoma with implants Completely excised All histologic types	Group VI: Widely invasive thymoma with distant metastasis Unresectable/biopsy only All histologic types

References

1. Barnatz PE, Khonsari S, Harrison EG Jr, Taylor WF. Thymoma: factors influencing prognosis. Surg Clin N A 53:885-892, 1973.
2. Masaoka A, Monden Y, Nakahara K, Tanioka T. Follow-up study of thymomas with special reference to their clinical stages. Cancer 48:2485-2492, 1981.
3. Verley JM, Hollman KK. Thymoma. A comparative clinical study of clinical stages, histologic features, and survival in 200 cases. Cancer 55:1074-1086, 1985.
4. Lewis JE, Wick MR, Bernatz PE, Taylor WF. Thymoma. A clinicopathologic review. Cancer 60:2727-2743, 1987.
5. Nakahara A, Ohno K, Hashimoto J et la. Thymoma: results with complete resection and adjuvant postoperative irradiation in 141 consecutive patients. J Thorac Cardiovasc Surg 95:1041-1047, 1988.
6. Pescarmona E, Rendina EA, Venuta F et al. Analysis of prognostic factors and clinicopathologic staging of thymoma. Ann Thorac Surg 50:534-538, 1990.
7. Maggi G, Casadio C, Cavallo A et al. Thymoma: results of 241 operated cases. Ann Thorac Surg 51:152-156, 1991.
8. Wilkins EW, Grillo HC, Scannell JG, et al. Role of staging in prognosis and management of thymoma. Ann Thorac Surg 51:888-892, 1991.
9. Park HS, Shin DM, Lee JS et al. Thymoma. A retrospective study of 87 cases. Cancer 73:2491-2494, 1994.
10. Pan C-C, Wu H-P, Yang C-F, Chen YK, Chiang H. The clinicopathologic correlation of epithelial subtyping in thymoma: a study of 112 consecutive cases. Hum Pathol 25:893-899, 1994.
11. Quintanilla-Martinez L, Wilkins EW, Choi N et al. Thymoma. Histologic subclassification is an independent prognostic factor. Cancer 74:606-617, 1994.
12. Blumberg D, Port JL, Weksler B et al. Thymoma: a multivariate analysis of factors predicting survival. Ann Thorac Surg 60:908-914, 1995.
13. Regnard J-F, Magdeleinat P, Drower C et al. Prognostic factors and long-term results after thymoma resection: a series of 307 patients. J Thorac Cardiovasc Surg 112:376-384, 1996.
14. Lattes R. Thymoma and other tumors of the thymus. An analysis of 107 cases. Cancer 15:1224-1260, 1962.
15. Salyer WR, Eggleston JC. Thymoma. A clinical and pathological study of 65 cases. Cancer 37:229-249, 1976.
16. Davies SE, MaCartney JC, Camplejohn RS et al. DNA flow cytometry of thymomas. Histopathology 15:77-83, 1989.
17. Asamura H, Nakajima T, Mukai K, Noguchi M, Shimosato Y. Degree of malignancy of thymic epithelial tumors in terms of nuclear DNA content and nuclear area. An analysis of 39 cases. Am J Pathol 133:615-622, 1988.
18. Yang W, Efird JT, Quintanilla-Martinez L, Choi L, Harris NL. Cell kinetic study of thymic epithelial tumors using PCNA (PC10) and Ki-67 (MIB-1) antibodies. Hum Pathol 27:70-76, 1996.
19. Chen F-F, Yan J-J, Jin Y-T, Su I-J. Detection of bcl-2 and p53 in thymoma: expression of bcl-2 as a reliable marker of tumor aggressiveness. Hum Pathol 27:1089-1092, 1996.
20. Stefanaki K, Rontogianni D, Kouvidou CH, et al. Expression of p53, mdm2, p21/waf1 and bcl-2 proteins in thymomas. Histopathol 30:549-555, 1997.
21. Rosai J. Histological typing of tumours of the thymus; in: World Health Organization International Histological Classification of Tumors, 2^nd. Ed, Springer, Berlin, 1999, pp.9-14.
22. Marino M, Muller-Hermelink HK. Thymoma and thymic carcinoma. Relation of thymoma epithelial cells to the cortical and medullary differentiation of the thymus. Virch Arch Pathol Anat 407:119-149, 1985.
23. Kirchner T, Muller-Hermelink HK. New approaches to the diagnosis of thymic epithelial tumors. Prog Surg Pathol 10:167-189, 1989.
24. Suster S, Moran CA. Primary thymic epithelial neoplasms. Current concepts and controversies. Anat Pathol 1997, Vol. 2, Fechner RE, Rosen PP, (eds), ASCP Press, Chicago, IL, 1997, pp.1-19.
25. Suster S, Moran CA. Thymoma, atypical thymoma, and thymic carcinoma. A novel conceptual approach to the classification of neoplasms of thymic epithelium. Am J Clin Pathol 111:826-833, 1999.
26. Suster S, Moran CA. Primary thymic epithelial neoplasms. Spectrum of differentiation and histological features. Semin Diagn Pathol 16:2-17, 1999.
27. Suster S, Moran CA. Thymoma classification. The ride of the Valkyries? Am J Clin Pathol 112:308-310, 1999.
28. Shimosato Y, Mukai K. Tumors of the mediastinum. In: Atlas of Tumor Pathology, Fascicle 21, third series, Armed Forces Institute of Pathology, Washington, D.C., 1997, pp.40-119.
29. Suster S, Moran CA. Spindle cell thymic carcinoma. Clinicopathologic and immunohistochemical study of a distinctive variant of primary thymic epithelial neoplasm. Am J Surg Pathol 23:691-700, 1999.
30. Koga K, Matsuno Y, Noguchi M et al. A review of 79 thymomas: modification of staging system and reappraisal of conventional division into invasive and non-invasive thymoma. Pathol Internat 44:359-367, 1994.
31. Pan CC, Chen WYK, Chiang H. Spindle cell and mixed spindle/lymphocytic thymomas: an integrated clinicopathologic and immunohistochemical study of 81 cases. Am J Surg Pathol 25:111-120, 2001.
32. Chalabreysse L, Roy P, Cordier J-F et al: Correlation of the WHO schema for the classification of thymic epithelial neoplasms with prognosis. A retrospective study of 90 tumors. Am J Surg Pathol 26:1605-1611, 2002.
33. Suster S, Moran CA. The Mediastinum; in: Modern Surgical Pathology; Weidner N et al (Eds), Saunders, Co., Philadelphia, 2002; pp.439-504.