Pleural fluid test for mesothelioma

A n estimated 150 000 patients develop a pleural effusion each year in the UK. Establishing the aetiology of pleural effusions can be challenging as they can be associated with over 50 systemic or pulmonary disorders. Exudative pleural fluids arise from plasma extravasation and contain proteins and cells from the systemic circulation, as well as those released locally from the pleura. This milieu of cellular products potentially holds vital clues that can unveil the underlying cause of the effusion. Decoding these clues has been a long and slow journey, and only in recent years have several disease-specific pleural fluid markers been incorporated into clinical practice (eg, adenosine deaminase for tuberculous pleuritis and brain natriuretic peptide for effusions from cardiac failure). It is logical that diagnostic clues for mesothelioma, a primary pleural malignancy, would rest within the pleural fluid. The incidence of mesothelioma is rising significantly in the UK and other European countries. Most patients with mesothelioma present with a pleural effusion, and this diagnosis should be considered in all patients with exudative effusions. However, fluid cytology has a notoriously low diagnostic yield, as differentiating mesothelioma cells from benign (or reactive) mesothelial cells is difficult. More invasive procedures such as thoracoscopy for tissue biopsy samples are frequently required. Finding a diagnostic marker for mesothelioma is a challenging endeavour. This is made difficult by the heterogeneity of mesothelioma, which comprises various histological subtypes (eg, epithelioid, sarcomatoid, desmoplastic) with dissimilar gene expression patterns, phenotypes and biological features. No unique molecule has been shown to reliably define mesothelioma from benign mesothelium or metastatic carcinomas, even by profiling the expression of tens of thousands of genes on mesothelioma tissues using microarray technology. SOLUBLE MESOTHELIN IN MESOTHELIOMA The Food and Drug Administration has recently approved the use of soluble mesothelin for disease monitoring in mesothelioma. This followed the findings of Robinson et al that blood levels of soluble mesothelin are significantly raised in patients with mesothelioma compared with controls with a wide range of pleuropulmonary diseases. This observation is robust and has been confirmed by other studies. Mesothelin is a 40 kDa glycoprotein predominantly found in normal mesothelial cells in the pleura, peritoneum and pericardium. Although mesothelin is bound to cell membrane, a circulating form termed soluble mesothelin-related peptides (SMRP or ‘‘soluble mesothelin’’) has been found in the serum. It has been suggested that the protein may be a splice variant of the normal mesothelin, with structural variation that alters its binding to cells. More recently it has been suggested that SMRP are mesothelin molecules released into the systemic circulation from malignant cells that overexpress mesothelin. 14 Most (.95%) patients with mesothelioma have a pleural effusion; it would be logical to assess the usefulness of soluble mesothelin in pleural fluid as a diagnostic test for mesothelioma. In this issue of Thorax, Creaney et al report that soluble mesothelin is detectable in pleural and peritoneal effusions and its levels are significantly higher (median 7-fold) in pleural mesothelioma (n = 52) than in effusions of benign pleuritis (n = 84) and non-mesothelioma malignancies (n = 56) (see page 569). Levels of soluble mesothelin were also higher in ascitic fluid from peritoneal mesothelioma than in benign controls. Using a cut-off that yielded 98% specificity, the sensitivity of raised soluble mesothelin in pleural fluid was 67% for all pleural mesothelioma and 77% for epithelioid mesothelioma. The results from the study by Creaney et al echoed those from a multicentre study from France which compared mesothelin levels in pleural fluid from patients with mesothelioma, metastatic carcinomas and benign asbestos pleural effusions. Scherpereel et al found a similar sensitivity (65%) for pleural fluid mesothelin with a specificity of 84%. To put this in context, soluble mesothelin has a much more favourable sensitivity than other tumour markers adopted in routine clinical practice, including prostate specific antigen (21%), carcinoembryonic antigen (11%) and cancer antigen (CA)-125 (57%) at similar specificity cut-off points (94–100%). In both studies the sensitivity of soluble mesothelin was limited by the lack of reactivity in sarcomatoid tumours 15 and some of the ‘‘mixed’’ mesotheliomas, presumably ones with a heavy sarcomatoid component. In both studies the sensitivity and specificity were calculated using a cutoff value determined in the same cohort. The cut-off values have not been tested in a separate cohort to confirm the sensitivity and specificity. False positives also occur: soluble mesothelin is released in significant quantities by ovarian and pancreatic carcinomas, as well as by a small proportion of lung carcinomas and, in the current series, a patient with lymphoma. Immunohistochemistry has revealed mesothelin staining in several types of non-mesothelioma cancers such as tumours of the gastrointestinal tract. It is likely that some of these tumours will shed mesothelin into the blood or serosal fluids. Because of the significantly higher incidence of malignant ascites from ovarian carcinomas, the value of soluble mesothelin in peritoneal fluid as a diagnostic marker for mesothelioma will be lower than in pleural effusions. Pleural fluid levels of soluble mesothelin were significantly higher than—but strongly correlated to—corresponding serum levels in the study by Creaney et al. This would be consistent with the belief that mesothelin is released from pleural mesothelioma cells into the pleural effusion and probably absorbed subsequently into the systemic circulation. As such, pleural fluid levels of soluble mesothelin may be a more sensitive test than serum levels in diagnosing mesothelioma. Measurement of soluble mesothelin levels in both pleural fluid and serum may further increase the sensitivity. The level of soluble mesothelin, either in the blood or pleural fluid, did not predict survival when adjusted for the histology. 15 However, it may reflect tumour load as a rapid reduction in mesothelin levels was observed after debulking surgery for peritoneal mesothelioma. A biomarker that reflects the tumour response to mesothelioma treatment would be extremely EDITORIALS 561