Japanese Journal of Clinical Oncology

Departments of ¹Internal Medicine and ²Pathology, National Kinki Central Hospital for Chest Diseases, Osaka, Japan

The level of hyaluronic acid (HA) was determined in the pleural fluid of 99 patients, including 19 with malignant mesothelioma, 27 with lung cancer, 1 with breast cancer, 1 with mediastinal tumor and 51 with non-malignant diseases. With a cut-off level at 100 µg/ml, the pleural fluid concentration of HA was high in 36.8% of patients (7 of 19) with malignant mesothelioma and 1.3% of patients (1 of 80) with lung cancer and other malignant and non-malignant diseases. The mean concentration of pleural fluid HA was significantly higher in patients with mesothelioma than in those with lung cancer and other malignant and non-malignant diseases. The pre-test probability of MM was 5.9% in this series. The LRs for >= 100, 50-99 and <= 49 µg/ml are 28.3, 3.3 and 0.5, respectively; these put the post-test probabilities at 64, 17 and 3%, respectively. Indeed, in cases of uncommon disease such as MM, the post-test probability is low even if the cut-off level of HA is >= 100 µg/ml. The discrimination between malignant mesothelioma and lung cancer needs special attention. In these two diseases, the LRs of MM for pleural fluid CEA >30, 10-30 and >10 ng/ml were 0.2, 1.9 and 2.4, respectively. The pre-test probability of MM for HA >= 100 µg/ml is 64%. Furthermore, because the LR for CEA is >10 ng/ml, the post-test probability is 81%. When the combination of two markers is considered, the high level of HA and the low level of CEA may be useful for the differential diagnosis of MM from pleuritis carcinomatosa.

Key words: hyaluronic acid - carcinoembryonic antigen - CEA - lung cancer - pre-test probability - post-test probability

Introduction

The differential diagnosis of pleural fluids is wide and may indicate the presence of pleural, pulmonary or extrapulmonary disease. When we attempt to identify the cause of the pleural effusion, most patients undergo various diagnostic procedures. These procedures include chest X-ray, CT scan of the thorax, thoracentesis with multiple chemical determinations, various kinds of tumor markers, microbiological studies, cytological study on pleural fluid and biopsy of the pleura. Malignant mesothelioma (MM) is an uncommon and usually fatal neoplasm for which there is currently no standard treatment (1 -3 ). Diagnosis of MM is often difficult. There are more than a few cases in which the biopsy by thoracotomy or thoracoscopy is necessary for diagnosis. In this disease, the concentration of hyaluronic acid (HA) is often elevated in pleural fluid, ascites and pericardial fluid (4 -7 ) and highly elevated HA has been considered a useful tumor marker. However, in many cases mesothelioma does not produce HA and other causes of elevated HA content occur (8 -10 ). Information concerning tumor markers is quantitative and clinical interpretation often depends upon the degree of positivity or negativity of a result. In order to evaluate its utility in the diagnosis of pleural fluid, we measured pleural fluid HA and assessed its usefulness as a tumor marker in MM, other malignant diseases and non-malignant diseases. In addition to standard sensitivity and specificity measurement, we described the likelihood ratios for MM in the evaluation of HA and carcinoembryonic antigen (CEA) in pleural fluid.

Patients and Methods

A study was carried out between February 1994 and January 1995 at National Kinki Central Hospital. Pleural fluid samples were examined in a prospective study of consecutive patients with pleural effusion. In all patients, thoracocentesis was performed for diagnostic purposes. These samples obtained at the time of thoracentesis were divided, with one part being sent for HA assay. Teijin Bio-Laboratories measured for HA by use of an electrophoretic method (Celaphore DP210 Cosmo). In addition, CEA was also measured in some of these patients. We measured CEA in the pleural fluid with a Glaozyme [New] CEA EIA kit (Sanyo Chemical Industries). The series consisted of 85 consecutively admitted patients with pleural effusion, and 14 pleural fluid samples of patients previously diagnosed as having MM were also examined (Table 1 ). The diagnoses of MM were based on biopsy specimens obtained at thoracotomy in eight patients, thorocoscopy in nine and transthoracic needle biopsy in two. All biopsy specimens were stained with hematoxylin, eosin, keratin, vimentin, CEA and alcian blue with or without pretreatment with hyaluronidase and confirmed as in the Osaka Mesothelioma Panel.

In total, 99 patients with pleural effusions were evaluated. Nineteen patients had pleural effusion due to malignant pleural mesothelioma and 27 had lung cancer (of which 18 had adenocarcinoma, 3 squamous cell carcinoma, 2 large cell carcinoma and 4 small cell carcinoma). Tuberculous pleuritis was found in 30 patients, pyothorax in 7, heart failure in 4, breast cancer in 1, mediastinal tumor in 1, pneumonia in 1, atypical mycobacteria in 1, idiopathic pulmonary fibrosis in 1, eosinophilic pneumonia in 1, pneumothorax in 1, hemothorax in 1, idiopathic eosinophilic pleural effusion in 2, collagen disease in 1 and unknown in 1.

Statistical analysis

The statistical significance of the difference between the two groups was assessed by unpaired t-tests. The significance level was set at P > 0.05.

The likelihood ratio (LR) for MM for each diagnostic category was calculated according to previously described methods (11 -14 ). An LR is a ratio of two probabilities: the probability of a given test result when the disease is present (the true-positive fraction) divided by the probability of the same test result when the disease is absent (the false-positive fraction). A positive test LR(+) can be expressed as follows:

LR(+) = true positive fraction/false positive fraction

Given the pre-test clinical probability of disease, the LR can be used to calculate the post-test probability of disease.

The LR is related to the odds of disease by the equation

post-test odds = pre-test odds * LR

The odds of disease and probability of disease are related by the following equations:

odds = probability/(1 - probability)

probability = odds/(1 + odds)

Table 1 . Causes of plural effusions in 99 patients

Cause of effusion	n
Malignant mesothelioma	19
Lung cancer	27
Adenocarcinoma	18
Squamous cell carcinoma	3
Large cell carcinoma	2
Small cell carcinoma	4
Tuberculous pleuritis	30
Pyothorax	7
Miscellaneous
Breast cancer	1
Mediastinal tumor	1
Congestive heart failure	4
Parapneumonic	1
Idiopatic eosinophilic pleural effusion	2
Hemothorax	1
Eosinophilic pneumonia	1
Atypical mycobacteria	1
Pneumothorax	1
Idiopathic interstitial pneumonia	1
Collagen disease	1
Unknown	1
Total	99

Results

The titers of pleural fluid HA concentrations in each patient are shown in Fig. 1 . The HA levels in pleural fluid were 515.8 ± 1201.1 (range 24-5100) µg/ml caused by MM, 32.8 ± 14.9 (range 13-90) µg/ml by lung cancer, 26.7 ± 11.7 (range 12-55) µg/ml by tuberculous pleuritis, 52.6 ± 95.9 (range 12-270) µg/ml by pyothorax and 30.5 ± 10.8 (range 17-124) µg/ml by other diseases (mean ± standard deviation). HA levels in patients with MM were significantly higher than in patients with other pleural effusions including lung cancer (p = 0.048). Using a cut-off level of 100 µg/ml, the evaluation showed a sensitivity toward mesothelioma of 36.8% and a specificity of 98.7%.

Likelihood Ratios and Post-test Probability

A likelihood ratio expresses the odds that a given level of a diagnosis test result would be expected in a patient with the target disorder (11 ). Likelihood ratios and post-test probabilities for the three-variable liner model were calculated and are listed in Table 2 .

The likelihood ratios for HA >= 100, 50-99 and <= 49 µg/ml were 28.3, 3.3 and 0.5, respectively. Of these 85 patients, 5 patients were MM at our hospital during this prospective study. Because the pre-test probability of the target disease is 5.9%, the post-test probability would be 64%.

Table 2 . Likelihood ratios for malignant mesothelioma for three levels of hyaluronic acid

Malignant mesothelioma
	Present		Absent
Hyaluronic acid (µg/ml)	Number	Proportion	Number	Proportion	Likelihood ratio
>= 100	7	0.368	1	0.013	28.3
50-99	4	0.210	5	0.063	3.3
le;49	8	0.421	74	0.925	0.5
Total	19		80

Table 3 . Likelihood ratios for malignant mesothelioma for three levels of CEA

	Malignant mesothelioma		Lung cancer
CEA (ng/ml)	Number	Proportion	Number	Proportion	Likelihood ratio
>30	2	0.154	16	0.640	0.2
10-30	1	0.077	1	0.040	1.9
>10	10	0.769	8	0.320	2.4
Total	13		25

Hyaluronic Acid and CEA as Diagnostic Markers for Malignant Mesothelioma and Lung Cancer

The HA and CEA of 13 patients with MM and 25 patients with lung cancer were measured at the same time. The median concentration of CEA in pleural fluid of patients with mesothelioma and lung cancer was 6.1 (range 0-33.1) ng/ml and 1045.2 (range 0-10 545.5) ng/ml, respectively. There were significantly lower concentrations of CEA in pleural fluid from patients with mesothelioma than from patients with lung cancer (p = 0.018). With a cut-off value of 40 ng/ml, the pleural fluid concentration of CEA did not rise in 13 patients with mesothelioma, but did so in 16 of 25 (64%) patients with lung cancer. In the patients with MM or lung cancer, the LRs of MM for pleural fluid CEA >30, 10-30 and >10 ng/ml were 0.2, 1.9 and 2.4, respectively (Table 3 ). The LR of MM increased with a decrease in the level of CEA.

Figure 1. Distribution of hyaluronic acid concentration in pleural effusion from patients with various diagnoses. (a) Malignant mesothelioma (n = 19); (b) lung cancer (n = 27); (c) tuberculosis (n = 30); (d) pyothorax (n = 7); (e) others (n = 16).

Discussion

HA, a member of the glycosaminoglycan family, is widely distributed in connective tissues, synovial fluid, the vitreous humor of the eye, urine and serum (15 ). Its function is an important constituent of the extracellular matrix. In recent years, this substance has been implicated in many cell functions, such as cell growth (16 ,17 ), locomotion and the metastatic process (18 -20 ). Some authors previously reported that an increased amount of HA has been recognized in pleural fluid with MM and it could serve as a useful tumor marker for this disease.

Roboz et al. (4 ) suggested that an effusion with an HA concentration of >0.25 mg/ml, confirmed by hyaluronidase susceptibility, is an indication of presence of MM. Pettersson et al. (5 ) reported that when the cut-off level was set at 100 mg/l, the HA concentration in effusions with MM was raised in 73% and in 23% with non-malignant inflammatory diseases, but in none with other cancers or with congestive heart failure. Nurminen et al. (6 ) estimated that hyaluronate-derived uronic acid in pleural fluid had a cut-off level of 75 mg/l in 85 patients and they showed that the sensitivity and the specificity for this disease were 73% and 90%, respectively.

During a 1-year study period, we evaluated 85 patients with pleural effusions in our hospital. The most common underlying disease in this series was tuberculosis (35.3%), followed by lung cancer (31.8%), pyothorax (8.2%) and mesothelioma (5.9%). The pre-test probability of MM was 5.9% in this series. The LRs for >= 100, 50-99 and <= 49 µg/ml are 28.3, 3.3 and 0.5, respectively; these put the post-test probabilities at 64, 17 and 3%, respectively (Fig. 2 ). Indeed, in the case of an uncommon disease such as MM, the post-test probability is low even if the cut-off level of HA is >= 100 µg/ml. If the pre-test probability of the target disease is assumed to be a high probability, e.g. 30%, and the LR is 28.3, the post-test probability is 92% [Fig. 2 A(d)]. Pre-test probabilities can be conveniently converted to post-test probabilities using the nomogram in Fig. 2 .

Figure 2. Nomogram for converting pre-test to post-test probability, using likelihood ratios (adapted from Fagan TJ. Nomogram for Bayes' theorem. N Engl J Med [Letter] 1975;293:257). (A) The pre-test probability of MM was 5.9% in this series. The LRs for HA >= 100, 50-99 and <= 49 µg/ml are 28.3, 3.3 and 0.5, respectively; these put the post-test probabilities at 64% (a), 17% (b) and 3% (c). If the pre-test probability of the target disease is assumed to be high, e.g. 30%, and the LR is 28.3, the post-test probability is 92% (d). (B) The pre-test probability of MM for HA >= 100 µg/ml is 64%. In the patients with MM or lung cancer, the LR for CEA >10 ng/ml, the post-test probability is 81% (a). If another patient has HA >= 100 µg/ml and CEA >30 ng/ml, the post-test probability is 26% (b).

Specifically, the discrimination between MM and lung cancer needs special attention. In malignant pleural effusion caused by MM, the concentration of HA is significantly higher than that in malignant pleural effusion caused by adenocarcinoma. This finding is useful for differential diagnosis between mesothelioma and adenocarcinoma (21 ). However, the practical clinical value of HA determinations in pleural effusion alone may be insufficient. We added the CEA value to this HA to increase the diagnostic value. For example, when one patient has HA >= 100 µg/ml and CEA >10 ng/ml in pleural fluid and the pre-test probability is 5.9%, if we use the level of HA alone, the post-test probability of the patient for MM would be 64%. To differentiate further between MM and lung cancer, we are able to use CEA levels in succession. The LR of MM for CEA >10 ng/ml is 2.4. For the differentiation between MM and lung cancer, the pre-test probability is 64% and the post-test probability would be 81% [Fig. 2 (a)]. On the other hand, if another patient has HA >= 100 µg/ml and CEA >30 ng/ml, the post-test probability would be 26% [Fig.2B(b)]. When a combination of the two markers is considered, the level of HA and CEA may be useful for the differential diagnosis of MM from pleural carcinomatosa. Similarly, some authors have emphasized that the elevated HA concentrations and low CEA concentrations in pleural fluid proved helpful in the differentiation between pleural mesothelioma and pleuritis carcinomatosa (5 ,22 ).

In previous studies, it has been demonstrated that HA synthetase activity in various biological systems can be stimulated with a variety of compounds, such as EGF (23 ), PDGF, TGF-[beta] (24 ) and IGF-I (25 ,26 ). Maeda et al. (27 ) suggested that the levels of TGF-[beta] in malignant pleural effusions caused by mesothelioma were significantly higher than those due to primary lung cancers. TGF-[beta] may contribute greatly to the formation of clinical features of mesothelioma, which are characterized by thickening pleura on chest films and large amounts of HA in malignant pleural effusion in mesothelioma.

Furthermore, the combined use of markers such as TGF-[beta], CEA and HA may be more useful for the diagnosis of mesothelioma.

References

1. Adams VI, Unni KK, Muhm JR, Jett JR, Ilstrup DM, Bernatz PE. Diffuse malignant mesothelioma of pleura: diagnosis and survival in 92 cases. Cancer 1986;58:1540-51. MEDLINE Abstract

2. Alberts AS, Falkson G, Goedhals L, Vorobiof DA, Van Der Merwe CA. Malignant pleural mesothelioma. A disease unaffected by current therapeutic maneuvers. J Clin Oncol 1988;6:527-35. MEDLINE Abstract

3. Sridhar KS, Doria R, Raub WA, Thurer RJ, Saldana M. New strategies are needed in diffuse malignant mesothelioma. Cancer 1992;70:2969-79. MEDLINE Abstract

4. Roboz J, Greaves J, Silides D, Chahinian AP, Holland JF. Hyaluronic acid content of effusions as a diagnostic aid for malignant mesothelioma. Cancer Res 1985;45:1850-4. MEDLINE Abstract

5. Pettersson T, Fröseth B, Riska H, Klockars M. Concentration of hyaluronic acid in pleural fluid as a diagnostic aid for malignant mesothelioma. Chest 1988;94:1037-9. MEDLINE Abstract

6. Nurminen M, Dejmek A, Mårtensson G, Thylen A, Hjerpe A. Clinical utility of liquid-chromatographic analysis of effusions for hyaluronate content. Clin Chem 1994;40:777-80. MEDLINE Abstract

7. Takeda K, Ohba H, Hyodo H, Shimizu T, Iwakura H, Hayashi T. et al. Pericardial mesothelioma: hyaluronic acid in pericardial fluid. Am Heart J 1985;110:486-8. MEDLINE Abstract

8. Friman C, Hellström P-E, Juvani M, Riska H. Acid glycosaminoglycans (mucopolysaccharides) in the differential diagnosis of pleural effusion. Clin Chim Acta 1977;76:357-61. MEDLINE Abstract

9. Boersma A, Degand P, Biserte G. Hyaluronic acid analysis and the diagnosis of pleural mesothelioma. Bull Eur Physiopathol Respir 1980;16:41-5.

10. Chiu B, Churg A, Tengblad A, Pearce R, McCaughey WTE. Analysis of hyaluronic acid in the diagnosis of malignant mesothelioma. Cancer 1984;54:2195-9. MEDLINE Abstract

11. Sackett DL, Haynes RB, Guyatt GH, Tugwell P. Clinical Epidemiology. A Basic Science for Clinical Medicine. 2nd ed. Boston: Little, Brown, 1985.

12. Radack KL, Rouan G, Hedges J. The likelihood ratio: an improved measure for reporting and evaluating diagnostic test results. Arch Pathol Lab Med 1986;110:689-93. MEDLINE Abstract

13. Giard RWM, Hermans J. Interpretation of diagnostic cytology with likelihood ratios. Arch Pathol Lab Med 1990;114:852-4.

14. Raab SS. Diagnostic accuracy in cytopathology. Diagn Cytopathol 1994;10:68-75. MEDLINE Abstract

15. Laurent TC, Fraser JRE. The properties and turnover of hyaluronan. In: Functions of the Proteoglycans, Ciba Foundation Symposium 124. Chichester: Wiley, 1986:9-29.

16. Tomida M, Koyama H, Ono T. Induction of hyaluronic acid synthetase activity in rat fibroblasts by medium change of confluent cultures. J Cell Physiol 1975;86:121-30. MEDLINE Abstract

17. Matuoka K, Namba M, Mitsui Y. Hyaluronate synthetase inhibition by normal and transformed human fibroblasts during growth reduction. J Cell Biol 1987;104:1105-15. MEDLINE Abstract

18. Turley EA. Hyaluronan and cell locomotion. Cancer Metastasis Rev 1992;11:21-30. MEDLINE Abstract

19. Knudson CB, Knudson W. Hyaluronan-binding proteins in development, tissue homeostasis and disease. FASEB J 1993;7:1233-41. MEDLINE Abstract

20. Laurent TC, Fraser JR. Hyaluronan. FASEB J 1992;6:2397-404. MEDLINE Abstract

21. Fraser RG, Pare JAP, Pare PD, Fraser RS, Genereux GP. The pleura. In: Fraser RG, Pare JAP, Pare PD, Fraser RS and Genereux GP, editors. Diagnosis of diseases of the chest. Philadelphia: W.B. Saunders, 1991:2750-93.

22. Ebert W, Leichtweis B, Drings P. Differentiation between pleural mesothelioma and pleuritis carcinomatosa. Comparison of CEA and hyaluronic acid determinations in pleural effusions with cytology. Eur Resp J 1991;4(Suppl. 14):174.

23. Lembach KJ. Enhanced synthesis and extracellular accumulation of hyaluronic acid during stimulation of quiescent human fibroblasts by mouse epidermal growth factor. J Cell Physiol 1976;89:277-88. MEDLINE Abstract

24. Heldin P, Laurent TC, Heldin CH. Effect of growth factors on hyaluronan synthesis in cultured human fibroblasts. Biochem J 1989;258:919-22. MEDLINE Abstract

25. Honda A, Iwai T, Mori Y. Insulin-like growth factor I (IGF-I) enhances hyaluronic acid synthesis in rabbit pericardium. Biochim Biophys Acta 1989;1014:305-12. MEDLINE Abstract

26. Honda A, Noguchi N, Takehara H, Ohashi Y, Asuwa N, Mori Y. Cooperative enhancement of hyaluronic acid synthesis by combined use of IGF-I and EGF and inhibition by tyrosine kinase inhibitior genistein, in cultured mesothelial cells from rabbit pericardial cavity. J Cell Sci 1991;98:91-8. MEDLINE Abstract

27. Maeda J, Ueki N, Ohkawa T, Iwahashi N, Nakano T, Hada T et al. Transforming growth factor-beta 1 (TGF-[beta]1)- and [beta]2-like activities in malignant pleural effusions caused by malignant mesothelioma or primary lung cancer. Clin Exp Immunol 1994;98:319-22. MEDLINE Abstract

---

Received December 16, 1996; accepted May 9, 1997
For reprints and all correspondence: Shinji Atagi, Department of Internal Medicine, National Kinki Central Hospital for Chest Diseases, 1180, Nagasone, Sakai, Osaka 591, Japan
Abbreviations: MM, malignant mesothelioma; HA, hyaluronic acid; CEA, carcinoembryonic antigen; EIA, enzyme immunoassay; LR, likelihood ratio

---

This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 19 May 1998
Copyright© Japanese Journal of Clinical Oncology, 1997.
CiteULike    Connotea    Del.icio.us    What's this?

This Article Abstract FREE Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (11) Request Permissions Google Scholar Articles by Atagi, S Articles by Yamamoto, S Search for Related Content PubMed PubMed Citation Articles by Atagi, S Articles by Yamamoto, S Social Bookmarking          What's this?