Journal of Thoracic Oncology:
Supplementary Prognostic Variables for Pleural Mesothelioma: A Report from the IASLC Staging Committee
Pass, Harvey I. MD*; Giroux, Dorothy MS†; Kennedy, Catherine RM, RA‡; Ruffini, Enrico MD§; Cangir, Ayten K.‖; Rice, David¶; Asamura, Hisao#; Waller, David**; Edwards, John††; Weder, Walter‡‡; Hoffmann, Hans§§; van Meerbeeck, Jan P.‖‖; Rusch, Valerie W.¶¶; on behalf of the IASLC Staging Committee and Participating Institutions
*Department of Cardiothoracic Surgery, NYU Langone Medical Center, New York, New York; †Statistics Department, Cancer Research and Biostatistics, Seattle, Washington; ‡Department of Cardiothoracic Surgery, University of Sydney, Strathfield Private Hospital Campus, Sydney, Australia; §Department of Thoracic Surgery, University of Torino, Ospedale San Giovanni Battista, Torino, Italy; ‖Department of Thoracic Surgery, Ankara University Faculty of Medicine, Ankara, Turkey; ¶Department of Thoracic and Cardiovascular Surgery, M. D. Anderson Cancer Center, Houston, Texas; #Thoracic Surgery, National Cancer Center Hospital, Tokyo, Japan; **Department of Thoracic Surgery, Glenfield Hospital, Leicester, United Kingdom; ††Department of Cardiothoracic Surgery, Northern General Hospital, Sheffield, United Kingdom; ‡‡Thoracic Surgery, University Hospital Zurich, Zurich, Switzerland; §§Department of Thoracic Surgery, Thoraxklinik, University of Heidelberg, Heidelberg, Germany; ‖‖Thoracic Oncology, University Hospital, Ghent, Belgium; and ¶¶Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, New York.
Disclosure: The authors declare no conflict of interest.
Address for correspondence: Harvey I. Pass, MD, Department of Cardiothoracic Surgery, NYU Langone Medical Center, 530 First Avenue, 9V, New York, NY 10016. Email: email@example.com
This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivitives 3.0 License, where it is permissible to download and share the work provided it is properly cited. The work cannot be changed in any way or used commercially.
Introduction: The staging system for malignant pleural mesothelioma is controversial. To revise this system, the International Association for the Study of Lung Cancer Staging Committee developed an international database. This report analyzes prognostic variables in a surgical population, which are supplementary to previously published CORE variables (stage, histology, sex, age, and type of procedure).
Methods: Supplementary prognostic variables were studied in three scenarios: (1) all data available, that is, patient pathologically staged and other CORE variables available (2) only clinical staging available along with CORE variables, and (3) only age, sex, histology, and laboratory parameters are known. Survival was analyzed by Kaplan–Meier, prognostic factors by log rank and stepwise Cox regression modeling after elimination of nonsignificant variables. p value less than 0.05 was significant.
Results: A total of 2141 patients with best tumor, node, metastasis (TNM) stages (pathologic with/without clinical staging) had nonmissing age, sex, histology, and type of surgical procedure. Three prognostic models were defined. Scenario A (all parameters): best pathologic stage, histology, sex, age, type of surgery, adjuvant treatment, white blood cell count (WBC) (≥15.5 or not), and platelets (≥400 k or not) (n = 550). Scenario B (no surgical staging): clinical stage, histology, sex, age, type of surgery, adjuvant treatment, WBC, hemoglobin (<14.6 or not), and platelets (n = 627). Scenario C (limited data): histology, sex, age, WBC, hemoglobin, and platelets (n = 906).
Conclusion: Refinement of these models could define not only the appropriate patient preoperatively for best outcomes after cytoreductive surgery but also stratify surgically treated patients after clinical and pathologic staging who do or do not receive adjuvant therapy.
The role of surgery in patients with malignant pleural mesothelioma (MPM) would be less controversial if there was an accurate and minimally invasive method that could forecast outcomes for individuals who are surgical candidates. MPM patients tend to be older individuals who are frequently functionally impaired and may have difficulty with aggressive therapy; however, there is a cadre of MPM patients who, with favorable biology and a multimodal approach, benefit from intense therapy. Factors that predict to a poor overall survival or rapid time to progression could potentially help medical oncologists and surgeons select only those patients who should undergo potentially harmful cytoreductions with the present 4% operative mortality.1 The best-known clinical prognostic scoring systems for MPM have originated from the European Organisation for Research and Treatment of Cancer (EORTC) and the Cancer and Leukemia Group B,2,3 and use a combination of biological and clinical factors. Poor performance status (PS), nonepithelioid histology, male sex, low hemoglobin, high platelet count, high white blood cell count, and high lactate dehydrogenase were found to be poor prognostic indicators in mesothelioma, and subsequently validated. Such detailed analyses with sufficient numbers of patients for meaningful assessment have been lacking in the surgically treated population.
In collaboration with the International Mesothelioma Interest Group, the International Staging and Prognostic Factors Committee of the International Association for the Study of Lung Cancer (IASLC) formed a Mesothelioma Domain to improve the current staging system resulting in the first large, international MPM database, which includes more than 2000 staged patients with MPM diagnosed from 1995 to 2008 (see Supplementary Appendices, Supplementary Digital Content, http://links.lww.com/JTO/A583). As described by Rusch et al.,4 a set of covariates were identified as predictive of survival in a “CORE” model for this analysis, which included best staging information, age, sex, histology (epithelioid or not), and type of surgical procedure (palliative versus extrapleural pneumonectomy or pleurectomy decortication. This report summarizes an analysis of additional tumor or patient characteristics for their prognostic ability as mandated by the Prognostic Factors Subcommittee of the Mesothelioma Domain. Armed with the CORE model described above, the aim of this study was to analyze potential clinical and laboratory prognostic variables from a surgical and nonsurgical perspective by studying cohorts of patients from the registry with or without known pathologic staging (i.e., relying on clinical tumor, node, metastasis [TNM]) to develop prognostic models.
MATERIALS AND METHODS
From January 4, 1995, to August 18, 2009, a total of 3101 patients met the screening criteria for having been diagnosed with MPM after 1995 and were available for follow-up. Of these 3101 patients, 2316 were staged either by pathological findings (pTNM, n = 1976) or by clinical findings (cTNM, n = 1265). Of these 2316 cases with the best possible TNM staging, 2141 had complete data on age, sex histology, and type of surgical procedure, and are cases that form the “CORE” model of predictive factors.
Definitions for Supplementary Prognostic Variables
The CORE variable demographics for the 2141 subjects are detailed in Table 1. Additional potential prognostic clinical variables for MPM that were available in the database included the use of chemotherapy or radiotherapy at any time (adjuvant therapy), smoking history, history of asbestos exposure, history of weight loss (defined as greater than 5% versus lesser than 5% in the previous 6 months), Eastern Cooperative Oncology Group (ECOG) PS, chest pain, and dyspnea. Smokers included current and former smokers, and ECOG PS ranged from 0 to 3 in the full database but was limited to 0 to 1 in the 2141 patients included in the analysis. For this surgical cohort of patients, 72.2% of patients having either a potentially curative (extrapleural pneumonectomy, pleurectomy decortication, or other) or a palliative surgical procedure (surgical exploration, pleurectomy, or pleurodesis) received adjuvant therapy. Laboratory parameters that were also analyzed included, hemoglobin, white blood cell count, and platelet count. Table 2 documents the number of subjects with clinical and laboratory data for these variables. Missing data for the 2141 patients ranged from 9.7% (use of adjuvant therapy) to 84.4% (history of weight loss).
Survival was measured from date of pathologic diagnosis to the date of last contact (at which time they were censored) or death attributable to any cause. Median survival was estimated using the Kaplan–Meier regression method. Prognostic groups were assessed by Cox regression analysis of survival, using the SAS system for Windows version 9.2 (SAS Institute Inc., Cary, NC) PHREG method. Significance values from pair-wise comparisons reflect the Wald test; those from joint model effects (e.g., comparing the full model to the null model) reflect the likelihood ratio test. All covariates in regression analyses were modeled categorically using indicator variables, and the threshold for statistical significance was set at a p value of 0.05. Age was classified into three categories, with cutpoints at 50 and 65 years. Covariates that met the criteria for statistical significance by univariate analysis were further evaluated for inclusion in multivariable regression models, using a stepwise algorithm with backward selection.
CORE Model for Survival
Table 3 shows the hazard ratios and p values for comparisons based on a Cox regression model of the CORE survival model as of March 2013 for all 2141 patients without missing data. All comparisons shown in the table are significant, except stage II versus stage I and the oldest versus the middle age groups (not shown). This model can be further consolidated into two categories for age (≥50 years versus younger).
Cox Regression Models: Pathological Staging Included
Table 4 shows the results of the Cox regression models including each proposed covariate in a univariate model and each proposed covariate in addition to the covariates of best stage (pathological), histology, sex, age, and type of surgery (palliative versus EPP/pleurectomy/decortication). Only the covariates that were independently statistically significant in addition to the CORE model parameters were included in the stepwise Cox regression algorithm. These covariates included adjuvant therapy, asbestos exposure, weight loss, chest pain, hemoglobin, platelets, and white blood cell count (WBC) (Figures 1 and 2). Lack of adjuvant therapy, along with the presence of asbestos exposure, weight loss, and chest pain, as well as low hemoglobin, high platelet count, and high white blood count, was found to be associated with a worse prognosis independent of the CORE variables.
Stepwise Cox Regression modeling with backwards selection was performed on a number of models, all of which included combining the CORE model with combinations of the supplementary variables with and without laboratory data. In the initial model, all of the parameters were included, and, after this model fit, the covariate with the least significance for predicting outcomes was removed, and this was continued until all the remaining covariates in the model were significant at the 0.05 level. As seen in Table 5, a number of starting models were included, which varied in patient numbers from 268 to 1027.
Because the starting model must have all of the covariates, including the CORE variables, only 268 of the 2141 patients could be evaluated in this way (set 1 plus labs/set 4: only two North American data sets included weight loss). Table 6 reveals that the final model included best stage, histology, sex, type of surgery, adjuvant treatment, weight loss, and WBC.
The most robust model (but compromised because of the exclusion of cases with missing weight loss or asbestos exposure data) for 550 patients was set 5 (set 3 plus labs), which included an evaluation of best stage, histology, sex, age, type of surgery, adjuvant treatment, chest pain, WBC, hemoglobin, and platelets (Table 7).
Cox Regression Models: Clinical Staging
When clinical stage (available in 1265 patients) was substituted in the CORE variables instead of pathologic staging as the “best stage,” a final model of 627 patients was similar to that with pathologic staging with the exception that hemoglobin level was also an independent prognostic variable (Table 8).
Cox Modeling in the Absence of Staging: Presentation Model
To simulate the situation of a potential surgical patient presenting only with a diagnosis of mesothelioma before any staging procedure to evaluate the patient for surgery, the CORE model was adjusted to include only age, histology, and sex. In this case, the impact of adjuvant therapy, type of operation, or staging would be unknown. Of the 2749 individuals with CORE variables of histology, sex, and age, 906 individuals also had laboratory data. The univariate model (presentation model) added to the modified CORE model reveals that weight loss, chest pain, and the laboratory parameters were significant variables (Table 9).
The final model after stepwise backward regression (Table 10) reveals that histologic subtype of MPM, sex, age, platelet count, and white blood cell count was predictive of outcome.
The IASLC Mesothelioma Domain was the first international effort to improve on the staging of this orphan disease by establishing an international retrospective registry examining CORE variables associated with survival after either palliative or after potentially curative surgery. CORE variables that were associated in multivariate analyses to be prognostically important included best stage, age, sex, histology (epithelioid or not), and the type of surgical procedure (palliative versus EPP/progressive disease). The 2141 patients in the present registry represent the largest such collection of surgically treated patients with mesothelioma, in whom all of these CORE variables were recorded.4
When the registry was first developed, the registry designers were influenced by the Cancer and Leukemia Group B and the EORTC prognostic indices that were the first to attempt to define additional factors, which included PS, symptoms, and selected laboratory parameters. The EORTC analysis eventually included not only overall survival but also progression-free survival.5 The clinical factors chosen for the IASLC Mesothelioma Registry supplementary prognostic analyses included the use of chemotherapy at any time (adjuvant therapy), smoking history, history of asbestos exposure, history of weight loss, defined as greater than 5% versus less than 5% in the previous 6 months, ECOG PS, chest pain, and dyspnea. Laboratory parameters included hemoglobin level, platelet count, white blood cell count, and lactate dehydrogenase level before the attempted surgical procedures. The chemotherapy data were standardized neither for the regimen used nor for the timing of the therapy, that is, neoadjuvant or postoperative. In fact, whether the patients received preoperative or postoperative chemotherapy (or both) could not be ascertained from the data because it was collected, and this can be construed as a weakness of this registry. Moreover, 193 patients had radiation along with surgery without chemotherapy, 608 had chemotherapy along with surgery but no radiation, and 579 surgery patients had both chemotherapy and radiotherapy, and any subanalysis of these cohorts for supplemental prognostic factors did not have enough common elements to make insightful conclusions. The extent of missing data in this first registry is unfortunate but it is hoped that this problem will be minimized in the ongoing prospective registry. For the final analysis, only 252 of the 2141 (12%) individuals, representing data from four North American Institutions had information on all of these supplementary variables in addition to the CORE variables, and stepwise regression modeling revealed that adjuvant therapy use, smoking history, WBC level, and weight loss were prognostically relevant. Indeed, the parameters that were most problematic included smoking history, weight loss, and ECOG PS. Because this was a surgical series of patients, it can be safely assumed that the majority of patients were ECOG 0 or 1, and that PS may not stratify in the models because of its relative homogeneity. Other factors such as the important symptom of chest wall pain, as well as all of the laboratory parameters, were recorded in approximately 50% of the patients with CORE variables. As such, further analyses using as many patients as possible with the remainder of the supplementary variables and laboratory values (n = 550) revealed that adjuvant therapy, WBC count, and platelets were prognostic indicators. Obviously one must consider that such analyses are compromised by the missing data; however, the number of patients in these internationally based, but compromised, analyses compares favorably with all of the studies to date attempting to prognosticate MPM using clinical and laboratory data.
The goal of registries such as this one is to be able to find those prognostic factors that have high fidelity and require minimal cost/invasion of the patient, and that in some combinatorial model would potentially change the treatment algorithm for a mesothelioma patient. Because this is a surgical based registry, there are obvious advantages in developing such models, including the presence of complete pathologic data from the time of the cytoreduction. In real life, however, the decision to operate on a patient with mesothelioma relies on factors apart or potentially complementary to pathologic stage, which is a major portion of the CORE variables in this study. An analysis of the cohort of 906 patients, who had parameters that could be assessed noninvasively on presentation, validated the findings of many of the previous studies listed in Table 11. The phenotype for a poor prognosis was defined as males older than 50 years old who presented with nonepithelial histotype and elevated platelet and WBC counts.2,3,5–13
The future and use of the MPM registry will depend on prospective accumulation of international cases along with uniform standardization of important demographic variables. For the CORE variables, further subdivision of the type and extent of surgical cytoreduction will be accomplished by the incorporation of recently published guidelines for their definition.14 Supplementary prognostic fields must be expanded to include more precise quantification of radiographic parameters, such as tumor volume and standardization of positron emission tomography-computed tomography interpretation15–19 (Table 12). Numerous studies have documented a relationship between post-treatment/postsurgical MPM survival and elevated standard uptake values (SUV); however, validation of a specific threshold standardized uptake value or standardization of SUV quantitation is lacking.
Finally, a number of tissue-based and blood-based genomic, epigenetic, and proteomic markers have been published either as single entities or as part of a profile for the prognostication of MPM.20–24 The majority of these have not been validated either in independent cohorts or in blinded analyses. The challenge for the registry is whether such markers can be added as fields. At the least, however, if the prospective registry is maintained, and participating institutions have ongoing tissue and blood procurement protocols for archiving of samples, the registry will represent a valuable coordinating entity for such validations.
The Mesothelioma Applied Research Foundation and the International Association for the Study of Lung Cancer provided funding to support the International Mesothelioma Staging Project. The sponsors had no input in the committee’s analysis of the data or in the committee’s suggestions for revisions to the staging system.
1. Rusch V, Baldini EH, Bueno R, et al. The role of surgical cytoreduction in the treatment of malignant pleural mesothelioma: meeting summary of the International Mesothelioma Interest Group Congress, September 11–14, 2012, Boston, Mass. J Thorac Cardiovasc Surg. 2013;145:909–10
2. Fennell DA, Parmar A, Shamash J, et al. Statistical validation of the EORTC prognostic model for malignant pleural mesothelioma based on three consecutive phase II trials. J Clin Oncol. 2005;23:184–9
3. Edwards JG, Abrams KR, Leverment JN, Spyt TJ, Waller DA, O’Byrne KJ. Prognostic factors for malignant mesothelioma in 142 patients: validation of CALGB and EORTC prognostic scoring systems. Thorax. 2000;55:731–735
4. Rusch VW, Giroux D, Kennedy C, et al.IASLC Staging Committee. Initial analysis of the International Association for the Study of Lung Cancer mesothelioma database. J Thorac Oncol. 2012;7:1631–1639
5. Francart J, Vaes E, Henrard S, et al. A prognostic index for progression-free survival in malignant mesothelioma with application to the design of phase II trials: a combined analysis of 10 EORTC trials. Eur J Cancer. 2009;45:2304–2311
6. Baud M, Strano S, Dechartres A, et al. Outcome and prognostic factors of pleural mesothelioma after surgical diagnosis and/or pleurodesis. J Thorac Cardiovasc Surg. 2013;145:1305–1311
7. Nojiri S, Gemba K, Aoe K, et al. Survival and prognostic factors in malignant pleural mesothelioma: a retrospective study of 314 patients in the west part of Japan. Jpn J Clin Oncol. 2011;41:32–39
8. Tanrikulu AC, Abakay A, Kaplan MA, et al. A clinical, radiographic and laboratory evaluation of prognostic factors in 363 patients with malignant pleural mesothelioma. Respiration. 2010;80:480–487
9. Richards WG, Godleski JJ, Yeap BY, et al. Proposed adjustments to pathologic staging of epithelial malignant pleural mesothelioma based on analysis of 354 cases. Cancer. 2010;116:1510–1517
10. Yan TD, Boyer M, Tin MM, et al. Prognostic features of long-term survivors after surgical management of malignant pleural mesothelioma. Ann Thorac Surg. 2009;87:1552–1556
11. Flores RM, Zakowski M, Venkatraman E, et al. Prognostic factors in the treatment of malignant pleural mesothelioma at a large tertiary referral center. J Thorac Oncol. 2007;2:957–965
12. Neumann V, Rütten A, Scharmach M, Müller KM, Fischer M. Factors influencing long-term survival in mesothelioma patients—results of the German mesothelioma register. Int Arch Occup Environ Health. 2004;77:191–199
13. Herndon JE, Green MR, Chahinian AP, Corson JM, Suzuki Y, Vogelzang NJ. Factors predictive of survival among 337 patients with mesothelioma treated between 1984 and 1994 by the Cancer and Leukemia Group B. Chest. 1998;113:723–731
14. Rice D, Rusch V, Pass H, et al.International Association for the Study of Lung Cancer International Staging Committee and the International Mesothelioma Interest Group. Recommendations for uniform definitions of surgical techniques for malignant pleural mesothelioma: a consensus report of the International Association for the Study of Lung Cancer International Staging Committee and the International Mesothelioma Interest Group. J Thorac Oncol. 2011;6:1304–1312
15. Abakay A, Komek H, Abakay O, et al. Relationship between 18 FDG PET-CT findings and the survival of 177 patients with malignant pleural mesothelioma. Eur Rev Med Pharmacol Sci. 2013;17:1233–1241
16. Gill RR, Richards WG, Yeap BY, et al. Epithelial malignant pleural mesothelioma after extrapleural pneumonectomy: stratification of survival with CT-derived tumor volume. AJR Am J Roentgenol. 2012;198:359–363
17. Sharif S, Zahid I, Routledge T, Scarci M. Does positron emission tomography offer prognostic information in malignant pleural mesothelioma? Interact Cardiovasc Thorac Surg. 2011;12:806–811
18. Lee HY, Hyun SH, Lee KS, et al. Volume-based parameter of 18
F-FDG PET/CT in malignant pleural mesothelioma: prediction of therapeutic response and prognostic implications. Ann Surg Oncol. 2010;17:2787–2794
19. Nowak AK, Francis RJ, Phillips MJ, et al. A novel prognostic model for malignant mesothelioma incorporating quantitative FDG-PET imaging with clinical parameters. Clin Cancer Res. 2010;16:2409–2417
20. Gordon GJ, Dong L, Yeap BY, et al. Four-gene expression ratio test for survival in patients undergoing surgery for mesothelioma. J Natl Cancer Inst. 2009;101:678–686
21. Pass HI, Goparaju C, Ivanov S, et al. hsa-miR-29c* is linked to the prognosis of malignant pleural mesothelioma. Cancer Res. 2010;70:1916–1924
22. Kao SC, Vardy J, Chatfield M, et al. Validation of prognostic factors in malignant pleural mesothelioma: a retrospective analysis of data from patients seeking compensation from the New South Wales Dust Diseases Board. Clin Lung Cancer. 2013;14:70–77
23. Kao SC, Klebe S, Henderson DW, et al. Low calretinin expression and high neutrophil-to-lymphocyte ratio are poor prognostic factors in patients with malignant mesothelioma undergoing extrapleural pneumonectomy. J Thorac Oncol. 2011;6:1923–1929
24. Kao SC, Pavlakis N, Harvie R, et al. High blood neutrophil-to-lymphocyte ratio is an indicator of poor prognosis in malignant mesothelioma patients undergoing systemic therapy. Clin Cancer Res. 2010;16:5805–5813
25. Pass HI. Biomarkers and prognostic factors for mesothelioma. Ann Cardiothorac Surg. 2012;1:449–456
Mesothelioma; Surgery; Prognosis; Registry; Staging
Supplemental Digital Content
Copyright © 2014 by the European Lung Cancer Conference and the International Association for the Study of Lung Cancer.
Highlight selected keywords in the article text.