Abstract Background Malignant mesothelioma (MM) carries a poor prognosis and response rates to palliative chemotherapy remain low. Identifying patients with MM that are unlikely to respond to chemotherapy could prevent futile treatments and improve patient quality of life. Studies have suggested that soluble mesothelin is a potential biomarker for early diagnosis and prognosis of MM. We set out to explore the utility of serum mesothelin in routine clinical practice. Methods We conducted a prospective exploratory study of serum mesothelin levels in 53 consecutive patients with MM at our institution between April 2009 and February 2011. Survival was assessed and analysed by mesothelin level as both continuous and categorical variables using Cox regression models. Differences in response rate between treatment groups were assessed by the Kruskal-Wallis Test. Results All 53 patients, who had been given study information agreed to participate. The patients’ median age was 69 (range 24–90). Median mesothelin level was 2.7 nM and this value was used to dichotomize categories: ≤2.7 nM (low) and >2.7 nM (high). The progression free survival (PFS) for low vs high mesothelin was 8.0 vs 5.1 months (HR 1.8, p-0.058). When mesothelin was accessed as a continuous variable for PFS the HR was 1.03 (95% CI: 1.01 - 1.06; p = 0.013). The overall survival (OS) for low vs high mesothelin was 17.2 vs 11.3 months (HR 1.9, p = 0.088). When mesothelin was assessed as a continuous variable for OS the HR was 1.02 (95% CI: 0.99 - 1.04; p = 0.073). Thirty patients received chemotherapy of which 18 had a pre-chemotherapy serum mesothelin level. In these 18 patients, the pre-chemotherapy mesothelin level did not correlate with response. Conclusions A single random sample provides information about patient prognosis but does not predict treatment response. We suggest further prospective validation of mesothelin testing as a prognostic biomarker.