Cancer staging systems are intended to assist clinicians to categorize patients diagnosed with a particular malignancy in terms of their life expectancy and potential response to specific therapeutic strategies. Generally speaking, patients categorized as early stage have anatomically localized malignancy associated with a longer life expectancy and better response particularly to local therapies directed at the primary tumor mass. Late stage patients are presumed to have metastatic disease indirectly disseminated to distant anatomical sites associated with shorter life expectancy and requiring systemic therapeutic approaches. A staging system is a set of criteria that defines such categories for a specific malignancy. Stage categories are defined in anatomical terms: How large is the tumor and what anatomical structures have become involved with (i.e., contain invasive proliferations of) tumor cells?

Tumor stage may be used as an eligibility or stratification factor for clinical trials, as a component of algorithms for determining prognosis, and in predicting the efficacy of particular therapeutic strategies for individual patients. Thus, stage serves as an important element of risk–benefit discussions between cancer patients and clinicians. In the context of an aggressive and rapidly fatal malignancy such as malignant pleural mesothelioma (MPM), these issues are critical to clinical, patient, and caregiver decision-making in relation to balancing expected quantity and quality of life. The practical value of any staging system is measured by its ability to separate patients into categories associated with differing expectations in terms of symptom relief, side effects, freedom from disease progression, and/or survival duration in relation to available treatment strategies.

Staging categorizes the physical extent of tumor growth relative to a patient’s normal anatomy. Categorical labels typically ranging from stage I to stage IV reflect the progressive nature of the underlying biologic process. Corresponding qualitative descriptions of “local” (stage I), “regional” (stages II–III), and “distant” (stage IV) disease refer, respectively, to a tumor that remains confined to the tissue or organ within which it initially arose, one that extends beyond the initiating tissue or organ either by direct growth to immediately surrounding tissues or by microscopic dissemination via lymphatic vessels to draining lymph nodes, and one that is disseminated via the systemic circulation to establish metastases in remote anatomic locations. For some systems, stage categories are defined directly by lists of criteria. However, the recognition that direct extension of the primary mass and metastasis via lymphatic and systemic routes represent separate and mutually independent classification parameters that may be combined to optimally define stage led to the development and preferential use of the more nuanced tumor-node-metastasis (TNM) classification systems.

Definitive determination of stage requires microscopic evaluation of representative tissue samples from the defining anatomic structures. Accurate staging is therefore only possible when all primary tumors have been surgically resected with an adequate margin, relevant regional lymph nodes and any suspected metastases have been biopsied, and the specimens have been subjected to complete gross and microscopic pathologic examination. The resulting classification is referred to as pathologic stage. Prior to treatment, or when surgical resection is not undertaken, tumor stage may be estimated based on physical examination and noninvasive imaging (clinical stage), alone or in combination with endoscopic or surgical biopsy or exploration with tissue sampling (surgical stage).

It has been challenging to establish a practical staging system for MPM owing to the rarity of the tumor, its anatomical complexity, its histologic heterogeneity (epithelioid, sarcomatoid, biphasic), and the limited availability of effective therapy. All of the staging systems that have been proposed are derived from and primarily applicable to the subset of patients undergoing surgery. However, for the minority of patients with MPM who undergo resection of their tumors, definitive staging based on pathologic assessment of resected MPM specimens is at best modestly correlated with patient outcome. Close apposition of multiple vital structures to pleural surfaces does not permit wide surgical margins to be taken around resected tumor, and this leads to probable underestimation of involvement of adjacent structures in determining pathologic stage. Furthermore, there is poor correspondence between preoperative clinical stage and final pathologic stage. In part, this results from the unique morphology and growth pattern of the primary tumor that render radiographic assessment challenging. The accuracy of clinical staging for the majority of MPM patients who are not treated surgically cannot be directly evaluated, but is likely to be similarly low unless there is unambiguous evidence of metastatic disease.

Staging, therefore, has not been as useful for MPM as it has been for some other malignancies to determine prognosis or inform treatment decisions, particularly whether or not surgical resection should be contemplated. For example, consider non–small-cell lung cancer (NSCLC), for which surgical resection is the standard of care primary therapy for clinical stages I to II, which can be determined with high accuracy by a combination of imaging and mediastinoscopy. A malignant lung tumor typically grows as an expanding, roughly spherical mass that is usually surrounded by lung parenchyma. Its diameter is easily measured radiographically, is predictive of recurrence and prognosis, and thus is a direct determinant of T classification. Lymphatic metastasis occurs in a progressive and predictable pattern based on the location of the primary lesion. By contrast, MPM typically spreads in a nodular pattern over the pleural surfaces. It often encases the ipsilateral lung and may expand to considerable bulk while remaining encapsulated within the pleural envelope, and thus be classified by definition as early disease. On the other hand, the tumor may transgress the pleural boundary in one or more areas to variably involve lung parenchyma, chest wall, pericardium, diaphragm, and/or mediastinal structures, affecting stage classification in ways that are not readily appreciated radiographically. Regional lymph node involvement with tumor is also not accurately assessed using current imaging modalities because nodal size does not correlate well with the presence of metastasis, and because some relevant nodes, particularly hilar and internal thoracic stations, are located in areas that are commonly involved with or immediately adjacent to primary tumor.^1,2

The establishment of a standard and accurate staging system also has been hampered by the relative rarity and overall poor prognosis of MPM. Only a fraction of patients undergo surgical resection. Consequently, relatively few cases have been available to establish and pathologically validate classification criteria. Surgical series large enough to inform staging have generally been retrospective cohorts from single institutions. Differing degrees of resection, failing to distinguish histologic subtypes, variable application of nonsurgical therapies, and treatment-related morbidity and mortality – each of which may influence outcome independent of stage – have hampered attempts to elucidate more subtle influences of staging criteria on patient prognosis. These factors have led to the proposal over the past several decades of a number of independent staging systems that differ in the significance attributed to specific classification criteria (for review, see Ref. 3).

The earliest MPM staging system proposed by Butchart et al.⁴ and later modifications proposed by Mattson⁵ and Sugarbaker et al.^6,7 define each stage directly based on specified anatomic criteria and their association with outcome in series of surgically treated patients. TNM staging systems for solid tumors consider tumor size and/or patterns of local invasion (T classification), involvement of regional lymph nodes (N classification), and remote or systemic disease (M classification) in determining stage. Early TNM classification criteria for MPM were proposed by Chahinian.⁸ The Cancer Staging Manuals, periodically co-published by the International Union Against Cancer (UICC)⁹ and The American Joint Committee on Cancer (AJCC),¹⁰ first included TNM criteria for MPM in their 4th editions. Modified TNM classification (Table 116-1) and stage grouping (Table 116-2) criteria proposed by the International Mesothelioma Interest Group (IMIG)¹¹ were adopted by the AJCC and UICC and have since remained unchanged through the current (7th) edition.

Size Criteria

T classification of malignancies that typically progress as enlarging, approximately spherical masses (e.g., non–small-cell lung carcinoma) is based primarily on tumor size, with upstaging when invasion of specific structures is present. By contrast, tumor size is not easily established for MPM, either clinically or pathologically, owing to the tumor’s irregular and highly variable morphology. The difficulty inherent in documenting the size of individual MPM tumors is reflected in the requirement for MPM-specific modification of RECIST criteria for longitudinal assessment of response to therapy.¹² Tumor size is therefore not currently considered when establishing MPM stage. Staging of some other malignancies, particularly those that arise within luminal mucosa (e.g., esophageal, gastric, colorectal carcinoma) or epithelium-lined hollow organs (e.g., bladder cancer) and therefore tend to have irregular rather than spherical morphology, similarly does not include tumor size criteria. For many such tumors, direct invasion into adjacent regular, concentric layers of muscle and vasculature provides a reliable basis for pathologic and endoscopic T classification.

Patterns of Local Invasion

No simple solution exists for MPM, where it is also challenging to accurately determine the degree and pattern of tumor extension into adjacent structures. The pleurae are bordered by the chest wall, diaphragm, pericardium, mediastinal organs, and lung. Commonly, the tumor circumferentially fills the chest cavity creating the potential for concurrent invasion into multiple structures.

According to AJCC/UICC staging criteria, T1 MPM tumors are confined within the ipsilateral pleural surfaces. Boutin et al.¹³ described early MPM as arising on the parietal pleura, with subsequent progression to the visceral pleura accompanied by a worsening of prognosis among Buchart stage I patients undergoing thoracoscopy. This observation was the basis for T1 subclassification in the IMIG TNM system. Some authors have questioned whether this distinction is clinically meaningful, because so few T1a tumors are diagnosed.^14,15 Tumors extending into interlobar fissures or involving lung parenchyma or diaphragm muscle are classified T2. Extension of tumor to involve endothoracic fascia or mediastinal adipose tissue, into but not through the pericardium, or chest wall soft tissue at a single focus, constitute T3. Further direct extension of tumor to involve chest wall soft tissue diffusely or at more than one focus, brachial plexus, bone (rib or spine), mediastinal organs, contralateral pleura, or through diaphragm or pericardium, is classified T4.

Even under the best of circumstances, accurate pathologic determination of MPM stage is challenging. The structures that surround the pleura cannot be resected to unambiguously evaluate all margins. Nevertheless, with extensive and systematic sampling, extrapleural pneumonectomy (EPP) specimens may be rigorously staged.¹⁶ This process commonly requires microscopic examination of 20 to 30 sections per specimen. For patients undergoing less complete operative procedures, complete pathologic staging may not be possible as a result of the inability to microscopically evaluate retained tissues.

Lymph Node Metastasis

Regional lymph node staging of MPM uses the same classification system as employed for staging lung cancer.¹⁷ Ipsilateral hilar and intraparenchymal lymph nodes are classified N1. Subcarinal and ipsilateral mediastinal or internal mammary nodes are classified N2. Contralateral mediastinal or internal mammary, or any scalene or supraclavicular nodes, are classified N3.

Lymph node metastasis has long been recognized as an indicator of poor prognosis in MPM.⁶ Studies of patients treated with EPP who had complete pathologic analysis have found the presence of N2 metastasis to be among⁷ (or the only¹⁸) significant prognostic factors identified in multivariate analysis. The relationship of N1 and N2 lymph node involvement to prognosis and stage, however, is more complex for MPM than for NSCLC. In the case of lung cancer, these designations are indicative of an orderly peripheral-to-central pulmonary lymphatic drainage pattern that usually results in a predictable metastatic progression through intraparenchymal, hilar, and ultimately, mediastinal nodal stations. By contrast, direct lymphatic drainage from the diaphragmatic pleura to the mediastinal nodal chain¹⁹ allows MPM to metastasize directly to N2 lymph nodes without first affecting N1 stations. Studies consistently report that approximately 40% of patients with N2 metastases do not demonstrate concurrent N1 disease.^18,20,21 Metastasis to only N1 lymph nodes may alternatively arise by direct pleural spread to the hilum or secondary to tumor invasion of lymphatics within lung parenchyma, thence following the pattern characteristic of lung cancer.²²

Distant Metastasis

MPM has historically been characterized as having a primarily local growth pattern, with only rare distant metastasis.²³ Other studies have suggested that distant disease may be more common than generally appreciated,²⁴ particularly in the context of surgical removal of the primary, although not all studies of surgical patients have concurred.²⁵ The distinction between direct and hematogenous metastasis is sometimes problematic because MPM has a tendency to grow through anatomical barriers such as the diaphragm by direct extension to involve “distant” organs such as the liver. Although there is definitive evidence of metastasis to CNS, bone, kidney, adrenal glands, lung, and pleura,²⁶ the frequency and impact on the natural history is less clear.

Validation and Performance

Staging systems must be periodically reevaluated as therapeutic strategies evolve and new standards of care are established. Such validation involves assessing the ability of staging criteria to stratify outcome in cohorts of patients receiving similar therapy. During more than a decade since publication of the IMIG¹¹ and revised BWH⁷ criteria, multiple therapeutic innovations have been developed and investigated. These range from intravenous chemotherapeutic regimens to intraoperative adjuncts such as hyperthermic intracavitary chemotherapy (HIOC) and photodynamic therapy (PDT) to adjuvant high-dose hemithoracic and intensity-modulated radiation therapy (IMRT). Each staging system has been used in published reports of series of patients treated on- and off-protocol using these strategies. None of the studies has provided validation of either system in terms of stage by stage stratification of outcome. For patients treated without surgery, a much needed standard of care chemotherapy regimen, combination cisplatin-pemetrexed, was established with the publication of a landmark multi-institutional phase III trial.²⁷ Although the stage distribution of patients on trial was reported in a table to demonstrate balance between arms, tumor stage was not among eligibility criteria and patient survival per stage was not reported. A European trial of similar design using another chemically related compound, raltitrexed, which also supported the platinum-antifolate strategy, noted that UICC stage was only barely significant (p = 0.0466) in a multivariate prognostic model.²⁸ Thus the two largest and most influential multi-institutional phase III trials in the MPM treatment literature neither include stage among eligibility criteria nor validate the prognostic utility of current clinical staging.

Richards et al.¹⁵ reported poor stage distribution and survival stratification by pathologic stage among 354 patients with epithelial MPM using either TNM or Brigham criteria. Most patients were classified to stage III by both systems. TNM criteria identified fewer early stage cases, as observed in other studies.²⁰ The TNM system does consistently identify a significant proportion of patients with stage IV disease with poor prognosis, arguing that current T4 classification criteria are appropriate.

Efforts to Improve Malignant Pleural Mesothelioma Staging

The International Association for the Study of Lung Cancer (IASLC) is conducting a prospective international cohort study of mesothelioma staging to inform potential future adjustment of TNM classification criteria. This effort builds on the IASLC staging committee’s data-driven restructuring of NSCLC TNM staging, which was adopted in the AJCC and UICC 7th edition staging manuals. The MPM project was undertaken in two phases, initially focusing on pooling retrospective databases already in existence, while initiating prospective data collection. Initial analysis of the retrospective database that comprised institutional series of surgically managed patients has been published.²⁹

The IASLC study confirmed in a large international dataset observations that had been made based on single-institution reports^15,30: (1) patients with MPM treated with surgery-based therapy experience a broad range of survival from a few months to more than 10 years; (2) the prognostic value of staging depends on the therapy being applied (in this case the extent of macroscopic surgical resection); (3) pathologic TNM stage based on current criteria does not stratify patient survival in a clinically useful way; (4) clinical TNM stage is poorly predictive of pathologic stage; and (5) survival following surgery-based multimodality therapy is more profoundly influenced by tumor histology than by stage.

It is widely recognized that the prognosis of patients with epithelial MPM is more favorable than that of patients with biphasic or sarcomatoid histology tumors. The effect of histology on patient survival is one of considerable magnitude and represents the dominant (and often only) prognostic factor in most published multivariate analyses. Thus in mixed histology cohorts it is difficult to determine the relative influence of factors that are more subtly related to prognosis. The need to stratify by cell type when assessing a survival endpoint in MPM has been recognized,^1,15 but is rarely addressed in practice. Cohort sizes of most MPM studies are small, requiring that all patients be analyzed together to maximize statistical power. This fact, combined with a justifiable desire for parsimony, has resulted in the development of one-size-fits-all staging systems for MPM.