Ann Thorac Surg 2005;80:1994-2000
© 2005 The Society of Thoracic Surgeons
Original article: General thoracic
Proposal of a Novel System for the Staging of Thymic Epithelial Tumors
Amedeo Vittorio Bedini, MD
a
,
*
,
Stefano Michele Andreani, MD, FRCS (Ed)
a
,
Luca Tavecchio, MD
a
,
Alessandra Fabbri, MD
b
,
Roberto Giardini, MD
b
,
Tiziana Camerini, PhD
c
,
Rosaria Bufalino, MS
c
,
Alberto Morabito, PhD
d
,
Juan Rosai, MD
a
a Thoracic Surgery Unit, University of Milan, S. Paolo Hospital, Milan, Italy
b Pathology Department, University of Milan, S. Paolo Hospital, Milan, Italy
c Scientific Board Department, National Cancer Institute, University of Milan, S. Paolo Hospital, Milan, Italy
d Department of Medicine, Surgery, and Dentistry, Institute of Statistics in Medicine, University of Milan, S. Paolo Hospital, Milan, Italy
Accepted for publication July 6, 2005.
* Address correspondence to Dr Bedini, Istituto Nazionale Tumori, Via Venezian 1, 20133 Milan, Italy (Email: amedeovittorio.bedini{at}istitutotumori.mi.it).
 |
Abstract
|
|---|
BACKGROUND: We designed and assessed a new TNM staging system (herein called the INT [Istituto Nazionale Tumori] system) for thymic epithelial tumors in order to overcome the perceived drawbacks of Masaoka's system, which represents the current standard.
METHODS: In all, 123 cases were evaluated. The histologic types according to the World Health Organization (WHO) classification were as follows: subtype A: 5 cases; AB: 40; B1: 16; B2: 29; B3: 16; and C: 17 cases. There were 45 Masaoka's stage I, 33 stage II, 26 stage III, and 19 stage IV cases. A total of 11 INT definitions were grouped into three stages: locally restricted disease (75 cases), which included Masaoka's stage I and selected stage II cases (no pleural invasion); locally advanced disease (37 cases), which included Masaoka's stage III cases plus those staged II owing to pleural invasion and those staged IV owing to intrathoracic nodal or limited pleuropericardial involvement; and systemic disease (11 cases), which included the remaining Masaoka's stage IV cases.
RESULTS: Completeness of resection, WHO types, and both staging systems were significant prognostic factors (p < 0.0001) on univariate analysis. The 95-month progression-free survival rates according to Masaoka's system were stage I: 100%; II: 93.6%; III: 46.3%; and IV: 23.2%. The INT system corresponding figures were as follows: locally restricted disease: 98.6%; locally advanced disease: 46.9%; and systemic disease: 11.7%. The INT system was the prognostic factor with the greatest impact (p = 0.0218) on multivariate analysis (Masaoka's system: p = 0.2012; completeness of resection: p = 0.6855; histology: p = 0.9386).
CONCLUSIONS: The INT system allows finer disease descriptions than Masaoka's system, resulting in a stage grouping with higher prognostic distinctiveness.
|
Introduction
|
|---|
Clinical studies on thymomas are difficult to carry out, their rarity having so far prevented their prospective evaluation. Among the most significant papers addressing the topic [1–13], only one includes a relevant number of cases, achieved through the contribution of more than 100 institutions [11]. Few of the published reports consist of more than 150 patients, even though some of them have spanned 30 years or more [4–7]. A relevant obstacle to the interpretation of results of the various series is represented by the use of different pathologic classifications. The morphologic heterogeneity of these tumors has been the source of conflicting nomenclatures [14], some of them plagued with considerable intraobserver and interobserver variability [15]. The different selection criteria have led to variously populated, poorly comparable subsets ranging from 3 to 6 microscopic subtypes [14, 16–19]. The World Health Organization (WHO) proposed in 1999 a novel pathologic classification of thymic epithelial tumors [20]), with the aim to supersede the previous nomenclatures. The better differentiated tumors were subdivided into five subsets designated as A, AB, B1, B2, and B3, whereas thymic carcinomas were placed in subset C. Neuroendocrine, germ cell, lymphoid, and stromal tumors were excluded from this category. This scheme, which reflects increasing invasiveness from A to C types, was basically retained by the latest WHO publication [21]. A consensus has built upon this system [6, 12, 13], which is likely to become the standard in the field. This should provide a valuable basis for new clinical studies focused on staging systems and their prognostic significance, which have been made object of debate [1–4, 22–27].
The Masaoka proposal [28] for the staging of thymomas (Table 1) has been widely adopted. Although translated into more descriptive TNM categories (Table 2) by Yamakawa, Masaoka, and coworkers [29], the initial system persisted and is still largely in use. However, it has been suggested that an updating is desirable. The main criticisms are the following: (1) the system is not well suited for the staging of carcinomas, now included in the WHO classification of thymic epithelial tumors [1, 22]; (2) the system does not provide an appreciable prognostic separation between stages I and II [3, 4, 6, 7, 11]; and (3) tumors invading the mediastinal pleura are at higher risk of recurrence than the other stage II tumors [23, 26]. Modifications of Masaoka's system have therefore been proposed to correct its perceived faults and to improve the prognostic distinctiveness of the stages [1–4, 22–27].
We designed a new staging system, aimed to incorporate the most relevant suggestions made by several authors as well as our original observations. We call our proposal the INT system (INT being the acronym of Istituto Nazionale Tumori [National Cancer Institute]).
|
Patients and Methods
|
|---|
The definitions that translated into a TNM system of Masaoka's stages [29] were the basis to build up the INT classification. A comprehensive review of the literature was performed in order to select those papers in which a discussion of the staging system was included [1–4, 22–27]. We considered a modification of the original categories if there was evidence or suggestion of distinctive prognostic differences among some subsets of disease extents. Moreover, standard resectability, analogy and common-sense criteria guided our choices on the new definitions, meant to create an ordinal scale reflecting increasing invasiveness. We specifically avoided definitions assessable only by surgical or pathological evaluation. These were drafted before data collection (Table 3). No definition of the original TNM system was modified if the aforementioned criteria were not fulfilled. The main modifications we introduced are the following: (1) the evidence/suggestion criterion prompted us to upgrade tumors invading the mediastinal pleura from the T2 to the T3 category [23, 26]; (2) therefore, tumors infiltrating neighboring organs were upgraded from the T3 to the T4 category (common-sense criterion); (3) pleural-pericardial implants were included in the T4 category only if they were limited to the anterior aspect of the pericardium (resectability criterion), a type of spread that can be encompassed by the margins of the standard procedure of maximal thymectomy [30], whereas pleural-pericardial spread located beyond these boundaries placed the tumor into a M1a category (analogy and common-sense criteria); the "a" suffix was adopted following the evidence/suggestion criterion [26]; (4) the N3 category described only involvement of prescalene or supraclavicular nodes; more distant nodal spread was upgraded to the M1b category, like all other extrathoracic manifestations (analogy and common-sense criteria); and (5) presence of residual tumor was defined by the use of the "R" category, which fulfills the need to classify surgical results (analogy and evidence/suggestion criteria) in that debulking procedures have been described as potentially curative [1, 6, 10, 27].
Patient Population
The clinical records of 149 patients who had been admitted to the thoracic surgery unit of our institute between 1972 and 1995 and had a diagnosis of thymic tumor were reviewed. Cases were selected in whom both histological diagnosis and surgical/pathological staging could be evaluated. Patients with cytologic diagnosis only, chemotherapy or radiotherapy administered before diagnosis and staging assessment, diagnosis obtained at autopsy, and specimens unavailable for pathologic review were excluded. Pathologic material was subjected to a new evaluation according to the WHO criteria and was examined independently by three pathologists (A.F., R.G., J.R.) who were masked to clinical data. These pathologists discussed the cases with a discrepant interpretation in order to obtain a consensus in classification. The 123 cases that remained constitute the object of the present study. The main demographic data are shown in Table 4. Forty-seven patients (38%) were symptomatic. Myasthenia gravis (36 cases), chest pain (13 cases), cough (12 cases), dyspnea (8 cases), and superior vena cava syndrome (6 cases) were the most frequent complaints. The roentgenologic identification of a mediastinal mass was an incidental finding in one third of the patients. Seven patients were diagnosed in the course of staging for synchronous primaries or of restaging for previous tumors. Pathology results, surgical reports, and clinical data were used for stage assessment according to both the systems. The study followed the Helsinki declaration rules and was approved by our Institutional Review Board on April 1997 (study 1997/9c/2). No written informed consent was required of the patients still alive, who were, however, contacted individually to obtain additional outcome information in the case of incomplete follow-up records. Relatives, family doctors, or registry offices were interviewed to obtain data on outcome of dead patients. The minimum follow-up period was 6 years. No patient was excluded from survival analyses.
Statistical Methods
Deaths regardless of cause (thymoma-related and others) were considered events in overall survival computing. For the calculation of progression-free survival, only relapses due to thymic tumor were considered events. The date of surgery was time zero for survival computing. Survival was plotted by means of the Kaplan-Meier method. Differences in survival among subgroups were tested by log-rank test. Cox proportional hazards regression analysis model was used to select significant prognostic factors. Proportionality among the survival rates attributable to factors entering the Cox model was assessed by the plot of the log (log [survival function]) versus time in each subgroup identified by the covariates, which were Masaoka's system, INT system, histology, and completeness of resection. Histologies A and AB were considered as a single class in the Cox model, because subtype A had no events. Likelihood ratio test was used in order to calculate the additional contribution of each of the four prognostic factors. All statistical tests were two-sided at the conventional 5% significance level. The agreement between the Masaoka and INT staging systems was tested by resorting to kappa test. The program used was STATA software release 7.0 for Windows XP.
|
Results
|
|---|
Putative prognostic factors and related survival probabilities are shown in Table 4. The overall and progression-free survival curves are shown in Figure 1. Progression-free survival of Masaoka's stages are shown in Figure 2. The difference in outcome between stages I and II was not significant, as already pointed out by others [3, 4, 6, 7, 11]. Cumulative survival of these stages was 97.37%. Patients were grouped into three INT stages, designated as locally restricted disease, locally advanced disease, and systemic disease. Their progression-free survival curves are shown in Figure 3. Locally restricted disease included 45 T1 and 30 T2 patients. Locally advanced disease included 8 T3 N0 (3 presented with mediastinal pleura invasion), 23 T4 N0, and 6 N1-2 (with any T) cases. Systemic disease included 9 M1a and 2 M1b cases (any T, any N). Concordance in stage assignment between the two systems, after combination of Masaoka stages I and II, approached 91%. We performed additional comparative evaluations between the two systems after inclusion of Masaoka stages I and II into one set, because this combination seems to be one of the major changes that INT system achieves.
View larger version (15K):
[in this window]
[in a new window]
|
Fig 2. Progression-free survival curves according to Masaoka's stages:  2 was 87.97 (p <0.0001) after combination of stages I and II.
|
|
View larger version (18K):
[in this window]
[in a new window]
|
Fig 3. Progression-free survival curves according to Istituto Nazionale Tumori stages. (LAD = locally advanced disease; LRD = locally restricted disease; SD = systemic disease.)
|
|
The frequencies of Masaoka and INT stages according to histologic subtypes are shown in Table 5. Type C group included 8 squamous nonkeratinizing, 5 squamous keratinizing, 3 undifferentiated, and 2 sarcomatoid carcinomas. No patient with A or AB types had nodal spread or metastases. The involvement of N1-2 was assessed only in B3 and C types (overall 8 cases). The outcome of the 17 patients with thymic carcinomas was assessed separately. Two of them were staged as locally restricted disease and Masaoka stage II as well: none of them experienced tumor progression. Thirteen patients had locally advanced disease: 6 relapsed and 7 were disease free. This outcome was compared with that of the 24 patients with the same INT stage but without C histologies. The differences in their progression-free survival curves could be due to chance (p = 0.98). Tumor progression occurred in both patients with systemic disease. Nine of the initial 17 patients with C type tumors were stage III according to Masaoka's system; 4 relapsed within the third year of follow-up, and 5 were free of disease at 4.3 to 13 years. Four stage IV patients progressed at 1 to 23 months, and the remaining 2 were progression-free at 1.2 and 8.9 years.
Surgical results according to completeness of resection are shown in Table 4. Five patients with a R0 and all 6 patients with a R1-R2a resection received adjuvant radiotherapy. The R2b patients underwent combined treatment in 12 instances, chemotherapy only in 7, and radiotherapy only in 5. There was no information regarding treatment in 3 cases. Overall, there was a wide variability of treatment schedules, of which only incomplete information was available in several instances.
The multivariate analysis weighted the prognostic impact of the four covariates significantly associated to disease progression: Masaoka system, INT system, histotype, and completeness of resection. We excluded from this analysis the single case with microscopic residual disease (R1). The complete Cox model, in which the four covariates were included, showed that the factor with the greatest prognostic impact was the INT system. The additional contribution of each of the four prognostic factors to the other three covariates was then examined by calculating the likelihood ratio test of each covariate adjusted for the other three factors. The likelihood ratio test gave the following results:
2 = 7.65, p = 0.0218 for INT system,
2 = 4.63, p = 0.2012 for Masaoka's system,
2 = 2.27, p = 0.6855 for completeness of resection, and
2 = 1.26, p = 0.9386 for histology. Results after combination of Masaoka stages I and II were
2 = 10.95, p = 0.0042 for INT system,
2 = 2.39, p = 0.3021 for Masaoka's system,
2 = 2.24, p = 0.3257 for completeness of resection, and
2 = 1.08, p = 0.8979 for histology.
|
Comment
|
|---|
Masaoka's scheme [28] has been widely adopted and has proved to be quite durable owing to its prognostic relevance. Our proposal for a modification incorporated into it the suggestions of several authors [1–4, 22–27] who pointed out some shortcomings, had as its foundation the TNM definitions into which Masaoka's system was translated [29], and was also based on TNM criteria. The TNM system is used worldwide for the classification of tumors arising from practically all sites. Categories are required to be numerous enough to obtain accurate definitions of anatomic disease extents to fulfill the need of a flexible system that could undergo revisions over time regarding grouping criteria. However, the definitions of categories need not undergo modifications, provided that they have been initially detailed. We required a surgicopathologic assessment of stages—the only means that could offer accurate data—in our series. Patients had been collected over a long period: diagnostic techniques underwent considerable evolution in the meantime, potentially resulting in inhomogeneous clinical data.
The INT stage grouping, admittedly devised in an arbitrary fashion, divides the tumors into three stages and shows high agreement with Masaoka's system once stages I and II are combined. In fact, only 9% of the patients are involved by stage migrations following INT criteria. Although the outcome of these patients (eg, those with mediastinal pleura invasion, limited pericardial implants, or N1-2 spread) cannot be separately evaluated owing to their numerical inconsistency, their inclusion in the INT intermediate stage gives rise to a significantly higher distinctiveness of stage-related survival. Whereas progression-free survival probabilities of INT locally advanced disease and Masaoka stage III were superimposable, slight but consistent survival differences emerged when comparing the initial and advanced stages of the two systems. Thus, Masaoka stage II was associated with a poorer overall outcome than INT T2 N0 M0 cases, whereas Masaoka stage IV showed a better prognosis than INT systemic disease. Needless to say, there is room for further modification of stage grouping in INT system. For instance, some of the features included in INT intermediate stage (such as T3 versus T4 or N0 versus N1-2) could justify a subdivision of this stage into two. However, for the purposes of this initial effort, we thought it prudent to follow strictly the parsimony criterion in order to avoid numerical inconsistency of the groups.
The INT system, in contrast to Masaoka's scheme, seems fit to stage all the subtypes that the WHO nomenclature currently lists, including carcinomas, and present stages with distinct therapy options. Patients with locally restricted disease are candidates for surgical excision only, provided that a radical resection can be achieved [11, 26, 31]. The stage we defined as locally advanced disease is amenable to primary locoregional therapy within multimodality treatment schedules. Some cases in this category can benefit from surgery as the initial approach, with the realization that resections extended to lung and great vessels may be required. Adjuvant treatments are advisable. Tumors initially unsuitable to radical surgery should undergo induction therapy and primary locoregional treatment thereafter. Patients with systemic disease should be submitted to primary chemotherapy plus radiotherapy as palliative or adjuvant treatment. Major surgical procedures, such as thymectomy plus extrapleural pneumonectomy, should be undertaken only in very selected instances [10].
In summary, the definitions of the INT system introduce some modifications to the TNM system, which translated Masaoka's scheme. They incorporate what several authors and our own experience suggested in order to describe in detail anatomic extent and to allow easy reevaluation and refinements according to stage grouping. The INT stages appear of high prognostic distinctiveness, seem suitable for the staging of carcinomas, and include patients with specific treatment options. There is a plan to collect a larger multi-institutional series to perform further assessment of the INT system. We are hopeful that this proposal will be deemed worthy of independent evaluation by the scientific community.
|
References
|
|---|
- Blumberg D, Burt ME, Bains MS, et al. Thymic carcinomacurrent staging does not predict prognosis. J Thorac Cardiovasc Surg 1998;115:303-309.[Abstract/Free Full Text]
- Koga K, Matsuno Y, Noguchi M, et al. A review of 79 thymomasmodification of staging system and rappraisal of conventional division into invasive and non-invasive thymoma. Pathol Int 1994;44:359-367.[Medline]
- Regnard JF, Magdeleinat P, Dromer C, et al. Prognostic factors and long-term results after thymoma resectiona series of 307 patients. J Thorac Cardiovasc Surg 1996;112:376-384.[Abstract/Free Full Text]
- Begg CB, Cramer LD, Venkatraman ES, Rosai JR. Comparing tumor staging and grading systemsa case study and a review of the issues, using thymoma as a model. Statist Med 2000;19:1997-2014.
- Maggi G, Casadio C, Cavallo A, Cianci R, Molinatti M, Ruffini E. Thymomaresults of 241 operated cases. Ann Thorac Surg 1991;51:152-156.[Abstract]
- Okumura M, Otha M, Tateyama H, et al. The World Health Organization histologic classification system reflects the oncologic behavior of thymoma. A clinical study of 273 patients Cancer 2002;94:624-632.[Medline]
- Quintanilla-Martinez L, Wilkins Jr EA, Choi N, Efird J, Hug E, Harris NL. Thymoma. Histologic subclassification is an independent prognostic factor Cancer 1994;74:606-617.[Medline]
- Wilkins KB, Sheik E, Green R, et al. Clinical and pathological predictors of survival in patients with thymoma Ann Surg 1999;230:562-574.[Medline]
- Nakahara K, Ohno K, Hashimoto J, et al. Thymomaresults with complete resection and adjuvant postoperative irradiation in 141 consecutive patients. J Thorac Cardiovasc Surg 1988;95:1041-1047.[Abstract]
- Park HS, Shin DM, Lee JS, et al. Thymoma. A retrospective study of 87 cases Cancer 1994;73:2491-2498.[Medline]
- Kondo K, Monden Y. Therapy for thymic epithelial tumorsa clinical study of 1,320 patients from Japan. Ann Thorac Surg 2003;76878–5.
- Kondo K, Yoshizawa K, Tsuyuguchi M, et al. WHO histologic classification is a prognostic indicator in Thymoma Ann Thorac Surg 2004;77:1183-1188.[Abstract/Free Full Text]
- Strobel P, Bauer A, Puppe B, et al. Tumor recurrence and survival in patients treated for thymomas and thymic squamous cell carcinomasa retrospective analysis. J Clin Oncol 2004;22:1501-1509.[Abstract/Free Full Text]
- Shimosato I, Mukai K. Tumors of the mediastinumIn: Rosai J, Sobin LH, editors. Atlas of tumor pathology. Third series, fascicle 21. Bethesda, MD: Armed Forces Institute of Pathology; 1997. pp. 34-35.
- Dawson A, Ibrahim NBN, Gibbs AR. Observer variation in the histopathological classification of thymomacorrelation with prognosis. J Clin Pathol 1994;47:519-523.[Abstract/Free Full Text]
- Salyer WR, Eggleston JC. Thymomaa clinical and pathological study of 65 cases. Cancer 1976;37:229-249.[Medline]
- Rosai J, Levine GD. Tumor of the thymusIn: Firminger HI, editor. Atlas of tumor pathology. Second series, fascicle 13. Washington, DC: Armed Forces Institution of Pathology; 1976. pp. 141-153.
- Marino M, Muller-Hermelink HK. Thymoma and thymic carcinoma. Relation of thymoma epithelial cells to the cortical and medullary differentation of the thymus Virchows Arch A Pathol Anat Histopathol 1985;407:119-149.[Medline]
- Suster S, Moran CA. Thymoma, atypical thymoma, and thymic carcinoma. A novel conceptual approach to the classification of thymic epithelial neoplasms Am J Clin Pathol 1999;111:826-833.[Medline]
- World Health OrganizationRosai J, Sobin LH. Histological typing of tumours of the thymus. 2nd ed.. New York: Springer; 1999.
- Müller-Hermelink HK, Engel P, Kuo TT, et al. Tumours of the thymusintroduction. In: Travis WD, Brambilla E, Müller-Hermelink HK, Harris CC, editors. World Health Organization classification of tumours. Pathology and genetics. Tumours of the lung, pleura, thymus and heart. Lyon, France: IARC Press; 2004. pp. 148-151.
- Tsuchiya R, Koga K, Matsuno Y, Mukai K, Shimosato Y. Thymic carcinomaproposal for pathological TNM and staging. Pathol Int 1994;44:505-512.[Medline]
- Shimosato Y. Controversies surrounding the subclassification of thymoma Cancer 1994;74:542-544.[Medline]
- Verley JM, Hollmann KH. Thymoma. A comparative study of clinical stages, histologic features, and survival in 200 cases Cancer 1985;55:1074-1086.[Medline]
- Pescarmona E, Rendina EA, Venuta F, et al. Analysis of prognostic factors and clinicopathological staging of thymoma Ann Thorac Surg 1990;50:534-538.[Abstract]
- Haniuda M, Morimoto M, Nishimura H, Kobayashi O, Yamanda T, Iida F. Adjuvant radiotherapy after complete resection of thymoma Ann Thorac Surg 1992;54:311-315.[Abstract]
- Mornex F, Resbeut M, Richaud P, et al. Radiotherapy and chemotherapy for invasive thymomasa multicentric retrospective review of 90 cases. Int J Radiat Oncol Biol Phys 1995;3:651-659.
- Masaoka A, Monden Y, Nakahara K, et al. Follow-up study of thymomas with special reference to their clinical stages Cancer 1981;48:2385-2392.
- Yamakawa Y, Masaoka A, Hashimoto T, et al. A tentative tumor-node-metastasis classification of thymoma Cancer 1991;68:1984-1987.[Medline]
- Jaretzki III A, Wolf M. Maximal thymectomy for mhyastenia gravis J Thorac Cardiovasc Surg 1988;96:711-716.[Abstract]
- Hejna M, Haberl I, Raderer M. Nonsurgical management of malignant thymoma Cancer 1999;85:1871-1884.[Medline]
This article has been cited by other articles:
|
|
|
|
 
T. Utsumi and M. Okumura
A Staging System for Thymic Epithelial Tumors: More Discussion is Required.
Ann. Thorac. Surg.,
September 1, 2006;
82(3):
1170 - 1170.
[Full Text]
[PDF]
|
|
|
|
|
|
|
A. V. Bedini
Reply.
Ann. Thorac. Surg.,
September 1, 2006;
82(3):
1170 - 1171.
[Full Text]
[PDF]
|
|
|
|
|
|
|
K. Kondo
Invited commentary
Ann. Thorac. Surg.,
December 1, 2005;
80(6):
2000 - 2001.
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
