We read with great interest the article by Creaney et al1 together with the associated editorial by Lee2 published in the July issue of Thorax. After the seminal paper by Robinson et al in 2003,3 there has been a lot of interest in the diagnostic value of soluble mesothelin in malignant pleural mesothelioma (MPM). Dr Lee emphasised the similarity between his results and those obtained by us.4 We completely agree with his statement that this new marker seemed therefore to be robust but could not be used as the sole diagnostic tool.2

We would like to comment on the finding of Creaney and coworkers that soluble mesothelin has no prognostic value. The same group of authors have previously suggested that an increasing serum level of this marker may reflect the tumour burden,3 thus suggesting that soluble mesothelin may have a prognostic value. The series reported by Creaney et al included only 52 cases of mesothelioma, and this figure may be insufficient to arrive at a firm statistical conclusion. In our first study which included 60 patients with MPM, we were also unable to find any relationship between patient outcome and soluble mesothelin assessed either in serum or in pleural effusion.4 However, when the same analysis was performed on a much larger series including almost 60% more patients with MPM,5 soluble mesothelin appeared as an independent prognostic factor along with the histological subtype, while tumour stage fell short as a significant parameter probably owing to the still low number of cases. Although both the Australian and French series may be subjected to bias, these data stress the urgent need for an international multicentre investigation on the value of soluble mesothelin in the management of malignant mesothelioma before we can firmly recommend the use of this marker in clinical practice.