Skip Navigation LinksHome > July 2013 - Volume 8 - Issue 7 > Combination of Mesothelin and CEA Significantly Improves the...
Journal of Thoracic Oncology:
doi: 10.1097/JTO.0b013e31828f696b
Original Articles

Combination of Mesothelin and CEA Significantly Improves the Differentiation between Malignant Pleural Mesothelioma, Benign Asbestos Disease, and Lung Cancer

Muley, Thomas PhD*‖; Dienemann, Hendrik MD, PhD‡‖; Herth, Felix JF MD, PhD‡‖; Thomas, Michael MD, PhD§‖; Meister, Michael PhD†‖; Schneider, Joachim MD, PhD

Free Access
Article Outline
Collapse Box

Author Information

*Translational Research Unit, Departments of Surgery, Pneumology and Respiratory Medicine, §Oncology, Thoraxklinik am Universitätsklinikum Heidelberg, Heidelberg, Germany, Translational Lung Research Centre Heidelberg (TLRC-H), Member of the German Centre for Lung Research, Heidelberg, Germany; and Institut und Poliklinik für Arbeits und Sozialmedizin, Universitätsklinikum, Giessen, Germany.

Disclosure: The authors declare no conflict of interest.

Address for correspondence: Thomas Muley, PhD, Sektion Translationale Forschung, Thoraxklinik am Universitätsklinikum Heidelberg, Amalienstrasse 5, 69126, Heidelberg, Germany. E-mail:

Collapse Box


Introduction: Soluble mesothelin-related peptides (SMRP) have been reported as potential markers for the diagnosis of malignant pleural mesothelioma (MPM). We wondered, whether a combination with a carcinoembryonic antigen (CEA) test might improve the relatively low diagnostic yield of the SMRP test.

Methods: In a retrospective study, SMRP (mesothelin) and CEA serum concentrations were measured, using commercially available kits, in 93 previously untreated MPM patients, 75 patients with benign asbestos disease, and 139 patients suffering from lung cancer (LC).

Results: The differentiation between MPM, LC, and benign asbestos disease could be improved by applying the ratio mesothelin/CEA. Whereas CEA expression was found to be low in MPM, most LC patients had elevated CEA serum levels. The area under curve (AUC) of the receiver operator characteristics curve for mesothelin alone was found to be only 0.708. For mesothelin/CEA the AUC of the receiver operator characteristics curve increased to 0.978. The sensitivity was 93% (69%) at 95% (100%) specificity for the differentiation between MPM and LC. Comparison of MPM and benign asbestos disease showed that the AUC was 0.887 and the sensitivity 56% (47%) at 95% (100%) specificity. In contrast, the AUC for the mesothelin test alone was only 0.715, and for the CEA test alone it was 0.16. An average increment in sensitivity of 38% (range, 16%–63%) could be achieved by the quotient mesothelin/CEA compared with the sensitivity of mesothelin alone.

Conclusion: The diagnostic yield of the mesothelin test can be considerably improved when combined with a CEA test with regard to the differential diagnosis between MPM and LC and between MPM and benign asbestos disease.

The number of occupational compensated asbestos-induced diseases in Germany is still rising. Currently, approximately three quarters (74.6%) of all occupational cancers eligible for compensation are ascribed to asbestos fibers.1 Because of the latency period the incidence of malignant mesothelioma will continue to increase, with a peak expected between the years 2015 and 2020.

Effective treatment of asbestos-induced malignancies requires a confirmation of the diagnosis. Histopathological and cytological examinations are necessary for discrimination between either malignant pleural mesothelioma (MPM) and pleural metastasis from lung cancer (LC), or from benign asbestos disease, like rounded atelectasis, which may mimic a tumor mass. Cytology is able to identify MPM only in a minority of cases (30%–50%).2 Histological differentiation between adenocarcinoma (AC) of the lung and mesothelioma is often indistinguishable.2 In addition, the prognosis of these malignancies is essentially poor because of disease detection mostly at an advanced stage and the high resistance of MPM to radiotherapy and chemotherapy.

Tumor markers should help to establish the differential diagnosis between MPM, primary LC, and benign asbestos disease. Also, markers should assist in the decision making of the therapy process, that is, selecting patients for a specific therapeutic approach.

Measurement of the serum concentrations of soluble mesothelin-related protein (SMRP or mesothelin) has been proposed by various groups as a valuable tool for the diagnosis of MPM.3–5 However, differentiation among patients with MPM, benign asbestos disease, and LC was found to be rather poor with respect to specificity and sensitivity.6 It has been known for a long time that carcinoembryonic antigen (CEA) is down-regulated in MPM7,8 but is elevated in various other lung conditions, that is, LC.9 The aim of this retrospective study was to improve our previously published results6 of the mesothelin test with regard to diagnostic accuracy and the power to differentiate between patients with MPM, primary LC, and benign lung disease related to asbestos exposure, by combining these results with the results of an additional CEA measurement. In this respect, we applied a quotient of mesothelin and CEA serum levels.

Back to Top | Article Outline



In this retrospective study, 93 patients (87 men, 6 women; aged 62 ± 8.3 years) with cytologically or histologically confirmed but previously untreated MPM were examined. On the basis of World Health Organization criteria,10 the histological subtyping of the untreated patients yielded 66 epithelial, 15 biphasic, and 12 sarcomatoid MPMs. According to the Union for International Cancer Control (UICC) recommendations11 there were six stage I, 20 stage II, 36 stage III, and 31 stage IV MPMs. The group of 139 patients with newly diagnosed primary LC (122 men, 16 women; mean age, 65 ± 9.6 years) consisted of 25 small-cell carcinomas (65 ± 9.2 years), three large cell carcinomas (67 ± 4.4 years), 61 squamous cell carcinomas (66 ± 8.1 years), and 50 ACs (62 ± 11.6 years). The exclusion criteria were any cancer-specific therapy before venepuncture, pulmonary metastases of extrapulmonary tumors, sarcoma, and lymphoma. The tumor patients were compared with a control group of 75 patients (74 men, 1 woman) with benign asbestos-related disease (for criteria for diagnosis, see refs. 12,13) that is, 19 patients with pleural plaques (65 ± 4.9 years), 49 patients with hyalinosis complicata or rounded atelectasis (64 ± 8.1 years) (for diagnosis, see ref. 14), and seven patients with lung asbestosis (small irregular opacities at a profusion of greater than or equal to one of one, according to International Labor Organisation criteria) (60 ± 6.6 years).15 The serum samples were collected between 1998 and 2005.

Written consent was obtained from all participants for the use of their blood samples for biomarker testing. The study was approved by the local clinical review boards.

Back to Top | Article Outline

Blood serum samples were processed within 120 minutes after venepuncture. Sera were kept frozen at −80°C, until analysis.

SMRP (mesothelin) was quantified by the sandwich-type ELISA kit MesoMark (Fuji Rebio Diagnostic, Malvern, PA; distributed by CIS bio GmbH, Germany). CEA was measured with Advia Centaur immunoassay (Siemens Healthcare Diagnostics Inc., Tarrytown, NY). Both tests were performed according to the manufacturers’ instructions.

Back to Top | Article Outline
Statistical Analysis

Data were statistically analyzed with SPSS 18.0 for Windows (SPSS, Chicago, IL). Because the distribution found in all groups was not normal, the statistical comparisons among the groups were carried out by the nonparametric Mann–Whitney U test. A p value less than 0.05, resulting from a two-sided test, was considered as statistically significant. Results of mesothelin measurement in the various subsets of patients were expressed as median and range. In addition, box-plots were created. The box represented the 25% to 75% quartiles. The whiskers indicated the ± 1.5-fold interquartile range up to the last value included. Values outside the 1.5-fold interquartile range were signed as extreme values. The medians were given as horizontal lines. Receiver operating characteristics (ROC) curves were used to analyze test accuracy. Optimal cutoff values, which define the best diagnostic efficiency, were calculated according to Youden index (Yi = sensitivity + specificity − 1).16 The cutoff value that resulted in a maximum value of Yi was considered as the optimal cutoff value.

Back to Top | Article Outline


Table 1 summarizes the results of mesothelin and CEA concentration measurements in the various study groups. Mesothelin was found to be significantly increased in patients suffering from MPM in comparison with those with benign asbestos disease (p < 0.0001), or primary LC (p < 0.0001). The median values were found to be 1.4 nmol/L, 0.9 nmol/L, and 0.8 nmol/L, respectively. There was no significant difference between mesothelin concentrations in benign asbestos disease and LC (p = 0.7). The highest CEA levels were found in LC followed by benign asbestos disease and MPM. The mesothelin/CEA ratio was significantly different among MPM, LC, and benign asbestos disease (p < 0.0001 for each groupwise comparison).

Table 1
Table 1
Image Tools

In Figure 1, the mesothelin concentration is plotted against the CEA concentration in the various study groups. MPM and LC patients can be separated clearly by the use of both markers.

Figure 1
Figure 1
Image Tools

In Figure 2 the data distribution of CEA, mesothelin, and the quotient mesothelin/CEA are given as a box plot. There was much less overlap between boxes of the MPM and LC group for the mesothelin/CEA quotient compared with mesothelin alone or CEA.

Figure 2
Figure 2
Image Tools

The differing diagnostic accuracies among MPM, LC, and benign asbestos disease are shown as ROC curves in Figure 3. For the comparison of MPM and LC, the quotient mesothelin/CEA resulted in a nearly perfect ROC curve. The AUC was found to be 0.987. In contrast, mesothelin alone gave an AUC of only 0.708, and CEA resulted in an AUC of 0.011. The differentiation between MPM and benign asbestos disease was, similarly, improved by the use of the quotient with an AUC of 0.887 compared with an AUC of 0.715 for mesothelin alone and an AUC of 0.16 for CEA alone.

Figure 3
Figure 3
Image Tools

Table 2 summarizes the results of sensitivity and specificity rates for various comparisons. The sensitivity rate was 93% (69%) at 95% (100%) specificity for the differentiation between MPM and LC, and 56% (47%) at 95% (100%) specificity between MPM and benign asbestos disease patients. By using the quotient mesothelin/CEA, an average increment in sensitivity of 38% (range, 16%–63%) compared with mesothelin alone could be achieved. Similarly, after Youdon optimization the improvement of sensitivity with an increment of at least 34% for the marker combination mesothelin/CEA remains.

Table 2
Table 2
Image Tools
Back to Top | Article Outline


Our current data clearly show that a combination of mesothelin and CEA as a quotient significantly facilitates the differential diagnosis between mesothelioma, LC, and benign asbestos disease. In our recent article,6 the differentiating power was found to be suboptimal, using mesothelin alone. This is because of the fact that there is a considerable overlap between serum mesothelin levels in mesothelioma, LC, and benign asbestos disease, impeding the differential diagnosis. The resulting sensitivity and specificity rates with a calculated AUC of approximately 0.708 are not sufficient to differentiate between either MPM and LC, or between MPM and benign asbestos disease. This is in accordance with the findings of Scherpereel et al.,4 where the AUC values of the ROC curves for mesothelin were found to be significantly higher in patients with MPM compared with controls with benign disease (AUC = 0.872), whereas differentiation of MPM from other malignant diseases was significantly lower (AUC = 0.693). Besides this finding, the poor sensitivity of mesothelin clearly limits its added value to early diagnosis and emphasizes the need for further biomarker research.18

CEA is a negative predictor for MPM and CEA expression and seems to be down-regulated in MPM when compared with LC or benign asbestos disease.8,19 However, a low level of CEA (median, 0.3 ng/mL) alone can neither prove, nor exclude MPM. Usually, low CEA levels suggest benign disease. In contrast, LC patients are known to have increased CEA serum levels.9

Therefore, we decided to combine the two markers mesothelin and CEA in a quotient. In MPM we expected an increased quotient (mesothelin/CEA) because of the positive- correlated marker mesothelin with the negative- correlated marker CEA. In contrast we expected a decreased quotient for LC. As a result, the quotient mesothelin/CEA significantly improved the differentiation between MPM and LC, as well as between MPM and benign asbestos disease. At 95% specificity, an increase in sensitivity between 14% (versus benign asbestos disease) up to 63% (versus LC) could be achieved. For the comparison of MPM and LC the AUC was found to be 0.987.

These results are in good accordance with the AUC of 0.94 described for MPM and non–small-cell cancer by van den Heuvel et al.19 In addition, the differentiation between MPM and benign asbestos disease was likewise improved by the use of the quotient, with an AUC of 0.887 compared with an AUC of 0.715 for mesothelin alone.

In a systematic review, the frequently investigated serum markers SMRP (mesothelin), CEA, Ber-EP4 (antihuman epithelial antigen), and calretinin were most valuable in discriminating mesothelioma from other malignant diseases.20 The markers epithelial membrane antigen and mesothelin were most valuable in differentiating mesothelioma from nonmalignant diseases. However, no marker performed sufficiently well in discriminating mesothelioma from all other diseases. Therefore, the use of single-tumor markers remains of limited diagnostic value.

Various marker combinations and complex algorithms have been suggested for differential diagnosis between benign and malignant thoracic diseases. For example, a fuzzy-classifier using a marker panel of cytokeratin fragment 21-1 (CYFRA 21-1), neuron specific enolase (NSE), and C-reactive protein (CRP) significantly improved detection of LCs in asbestosis patients.21 A similar approach might probably be useful for MPM diagnosis.

Van den Heuvel et al.19 evaluated a combination of mesothelin, CEA, and CYFRA 21-1. Although CYFRA 21-1 was able to discriminate between normal and malignant disease, only the two serum markers, CEA and mesothelin, proved capable of discriminating MPM from non–small-cell lung cancer, with a sensitivity of 66% at 96% specificity. A sensitivity rate of 93% (69%) at 95% (100%) specificity was found for the differentiation between MPM and LC, and may therefore, be a useful diagnostic tool to distinguish between both malignancies more accurately.

In a different study, a combination of mesothelin with osteopontin was analyzed in MPM compared with healthy persons and patients with benign lung diseases.22 An increased sensitivity and specificity was reported for the combination. However, no data in relation to differential diagnosis to LC were presented. In our view, this seems to be a most important diagnostic issue. Unfortunately, the measurement of osteopontin suffers for some technical and analytical problems, on which we had commented in our recent article.6

In pleural effusions, Blanquart et al.17 recently introduced a combination of chemokine chemokine (C-C motif) ligand 2 (CCL2), Galectin-3, and SMRP as a favorable tool to differentiate MPM from AC and benign pleural effusions. The AUC of the ROC curve of their logistic regression model was found to be 0.968 and close to the AUC reported by us. The usefulness of their model in serum samples, which might offer a diagnostic solution for a broader patient group, remains to be demonstrated.

Ostroff et al.23 reported on a biomarker discovery assay by measurement of more than 1000 proteins simultaneously in biological samples, applying a DNA aptamer-technology. After multivariate approaches 64 candidate protein biomarkers were identified and a 13-protein classifier was validated. The classifier accuracy for detection of MPM in the asbestos-exposed population was maintained in a blinded validation set, with a sensitivity of 90%, a specificity of 89%, and an overall accuracy of 92%. The proteins were related to inflammation and regulation of cellular proliferation.To date, none of the classifier biomarkers have previously been associated with MPM. Mesothelin was confirmed as a potential biomarker by this group, but other markers proved to be superior. In paired samples, the random forest classifier AUC of 0.99 and 91%/94% sensitivity/specificity was superior to that of mesothelin with an AUC of 0.82 and 66%/88% sensitivity/specificity. Differential expression of CEA was not statistically significant.

Our data of the marker combination mesothelin/CEA in the detection and differentiation of patients with MPM and primary LC in comparison with benign diseases related to asbestos exposure resulted in a comparable outcome. Mesothelin and CEA are relatively easy to measure. Complex mathematical procedures are not required. On the basis of the data reported here, the use of both tests can, therefore, be recommended for diagnosis and differential diagnosis.

Back to Top | Article Outline


The technical assistance of Monika Philipp (Giessen), and the maintenance of mesothelioma databases at the Thoraxklinik Heidelberg by Christa Stolp are gratefully acknowledged. We thank Maren Schuhmann for language revision.

This work was partly funded by BMBF Germany No.82DZL00402.

Back to Top | Article Outline


1. Butz M. Beruflich verursachte Krebserkrankungen. Dokumentation des Berufskrankheiten-Geschehens in Deutschland. Deutsche gesestzliche Unfallversicherung. 2012 Sankt Augustin, Germany:1–79

2. Krismann M, Müller KM, Jaworska M, Johnen G. Pathological anatomy and molecular pathology. Lung Cancer. 2004;45S:S29–33

3. Robinson BW, Creaney J, Lake R, et al. Mesothelin-family proteins and diagnosis of mesothelioma. Lancet. 2003;362:1612–1616

4. Scherpereel A, Grigoriu B, Conti M, et al. Soluble mesothelin-related peptides in the diagnosis of malignant pleural mesothelioma. Am J Respir Crit Care Med. 2006;173:1155–1160

5. Pass HI, Wali A, Tang N, et al. Soluble mesothelin-related peptide level elevation in mesothelioma serum and pleural effusions. Ann Thorac Surg. 2008;85:265–72; discussion 272

6. Schneider J, Hoffmann H, Dienemann H, Herth FJ, Meister M, Muley T. Diagnostic and prognostic value of soluble mesothelin-related proteins in patients with malignant pleural mesothelioma in comparison with benign asbestosis and lung cancer. J Thorac Oncol. 2008;3:1317–1324

7. Mezger J, Lamerz R, Permanetter W. Diagnostic significance of carcinoembryonic antigen in the differential diagnosis of malignant mesothelioma. J Thorac Cardiovasc Surg. 1990;100:860–866

8. Fuhrman C, Duche JC, Chouaid C, et al. Use of tumor markers for differential diagnosis of mesothelioma and secondary pleural malignancies. Clin Biochem. 2000;33:405–410

9. Schneider J. Tumor markers in detection of lung cancer. Adv Clin Chem. 2006;42:1–41

10. Travis WD, Brambilla E, Müller-Hermelink HK, Harris CC (Eds).. Tumours of the lung, pleura, thymus and heart. World Health Organization Classification of Tumours. Pathology and Genetics. 2004 Geneva WHO Press:Pp. 1–344 Available at: Accessed March 30, 2013.

11. Sobin LH, Gospodarowicz MK, Wittekind Ch Eds TNM Classification of Malignant Tumours,. 20097th Ed Chicester, UK John Wiley & Sons

12. Parker JE. Radiological criteria: The use of chest imaging techniques in asbestos-related diseases. In Proceedings of an International Expert Meeting on Asbestos, Asbestosis and Cancer, People and Work, Finnish Institute of Occupational Health, Helsinki, Finland. Research Reports. 1997;14:28–40

13. Baur X, Clasen M, Fissler-Eckhoff A, Heger M, Hering KG, Hofmann-Preiss K, et al. Diagnostics and expert opinion of Asbestos-induced occupational diseases. Interdisciplinary guideline of the German respiratory society and the German society of occupational and environmental medicine. Pneumologie. 2011;65:159–188

14. Hauser-Heidt G, Schneider J, Hackstein N, Litzelbauer D, Rau WS, Woitowitz HJ. Rundherdatelektasen als Pseudotumoren der Lunge: Eine neue MdE-relevante Folge arbeitsbedingter Asbestfaserstaub-Einwirkung. Zbl Arbeitsmed. 2002;52:295–304

15. International Labour Office (ILO. Guidelines for the Use of the ILO International Classification of Radiographs of Pneumoconiosis. 2002Revised Edition 2000 Geneva International Labour Office (Occupational Safety and Health Series No 22).

16. Youden WJ. Index for rating diagnostic tests. Cancer. 1950;3:32–35

17. Blanquart C, Gueugnon F, Nguyen JM, et al. CCL2, galectin-3, and SMRP combination improves the diagnosis of mesothelioma in pleural effusions. J Thorac Oncol. 2012;7:883–889

18. Hollevoet K, Reitsma JB, Creaney J, et al. Serum mesothelin for diagnosing malignant pleural mesothelioma: an individual patient data meta-analysis. J Clin Oncology. 2012;30:1541–1549

19. Van den Heuvel MM, Korse CM, Bonfrer JM, et al. Non-invasive diagnosis of pleural malignancies: the role of tumour markers. Lung Cancer. 2008;59:350–354

20. van der Bij S, Schaake E, Koffijberg H, Burgers JA, de Mol BA, Moons KG. Markers for the non-invasive diagnosis of mesothelioma: a systematic review. Br J Cancer. 2011;104:1325–1333

21. Schneider J, Bitterlich N, Kotschy-Lang N, Raab W, Woitowitz HJ. A fuzzy-classifier using a marker panel for the detection of lung cancers in asbestosis patients. Anticancer Res. 2007;27(4A):1869–1877

22. Cristaudo A, Bonotti A, Simonini S, et al. Combined serum mesothelin and plasma osteopontin measurements in malignant pleural mesothelioma. J Thorac Oncol. 2011;6:1587–1593

23. Ostroff RM, Mehan MR, Stewart A, et al. Early detection of malignant pleural mesothelioma in asbestos-exposed individuals with a noninvasive proteomics-based surveillance tool. PLoS ONE. 2012;7:10–e46091

Soluble mesothelin-related proteins; Mesothelioma; Lung cancer; Asbestosis; Carcinoembryonic antigen.

Copyright © 2013 by the European Lung Cancer Conference and the International Association for the Study of Lung Cancer.


Article Tools



Search for Similar Articles
You may search for similar articles that contain these same keywords or you may modify the keyword list to augment your search.

Other Ways to Connect



Visit on your smartphone. Scan this code (QR reader app required) with your phone and be taken directly to the site.

 For additional oncology content, visit LWW Oncology Journals on Facebook.