The incidence of malignant mesothelioma is increasing in the UK and continental Europe.1 In 1995 Petoet al predicted a progressive rise in the number of new cases recorded until the year 2020.2Specialist units are seeing more patients than ever and some general practitioners are seeing patients with mesothelioma for the first time. Effective treatment remains elusive. Radical surgery is rarely performed in the UK. The reasons for this are unclear but probably relate to the lack of randomised data to support its use and a perception that the mortality and morbidity of the preferred operation (extrapleural pneumonectomy) is excessive. Radiotherapy can be useful for relief of pain and shrinkage of chest wall masses3 but there is no evidence that it improves survival. Chemotherapy, however, is showing some promise: three recent phase II trials have reported tumour response rates exceeding 20% and in each of these studies symptoms were improved by treatment.4-6 In view of these encouraging data it is now an appropriate time to perform randomised clinical trials for patients with malignant mesothelioma. To increase confidence that treatment effects are real, clinical trials should stratify patients according to prognostic group.

There is no single accepted way in which to stage or predict prognosis for mesothelioma patients. The simple Butchart staging system has been widely used in the UK.7 Several systems based on TNM staging have been proposed but have been largely superseded by a new staging system proposed by the International Mesothelioma Interest Group (IMIG).8 The IMIG system has the advantage that resectable disease is described in a practical and more detailed fashion. A similar system has been described by Sugarbakeret al 9 and is known as the Brigham staging. The disadvantage of these latter two systems is that full staging information—for example, about resection margin and lymph node involvement—can only be obtained after pleural debulking surgery, either by total pleurectomy or extrapleural pneumonectomy, with lymph node sampling. This limits the usefulness of the systems, particularly in the UK where these operations are rarely performed. Computed tomography with magnetic resonance imaging in selected cases often allows a fair approximation of clinical stage and PET scanning may prove to be of additional value. Detailed staging by imaging is certainly required for identification of patients who may be suitable for multimodality therapy (radical surgery followed by chemotherapy and/or radiotherapy). However, prognostic scoring systems are the most readily available and practical method for assessing prognosis for most patients and may be valuable in stratification for clinical trials.

In this issue of Thorax, Edwardset al 10 confirm the reproducibility of two recently described prognostic scoring systems. The authors retrospectively analysed a large group of patients with mesothelioma treated within a single institution in Leicester between 1988 and 1999. Prognostic variables were analysed by univariate and multivariate models and were then incorporated into groups as described by two established scoring systems. The prognostic scoring systems used were the European Organization for Research and Treatment of Cancer (EORTC) system11 and the US Cancer and Leukemia Group B (CALGB) system.12 The EORTC system defines a high risk group of patients according to several factors: poor performance status, high white blood cell count at diagnosis, probable or possible (uncertain) histology, male sex, and sarcomatous cell type. The CALGB system defines six distinct prognostic subgroups with two year survival times ranging from 1.4 months to 13.9 months. The subgroup with best survival included patients with normal performance status and younger age (<49 years). Another indicator of better prognosis was a haemoglobin level at diagnosis of 14.6 g/dl or more. The worst survival was associated with patients with poorer performance status and baseline white blood cell count of more than 15.5 × 10⁹/l. Both scoring systems correctly identified patients with a poor prognosis in the Leicester series. Poor prognosis factors by multivariate analysis of the Leicester patients were sarcomatous cell type, low haemoglobin, high white blood cell count, poor performance status, and male sex. One year and two year survival rates were 21.3% and 3.5%, respectively. Survival rates for Leicester patients within prognostic groups were equivalent to patients in the EORTC and CALGB series.

The EORTC and CALGB prognostic scoring systems have thus been validated in a retrospective cohort. These systems could be considered for use prospectively in forthcoming phase II and III trials in malignant mesothelioma. Variation in prognostic factors may at least partially explain variation in survival between different phase II trials and information about such factors could be useful in interpreting the results. The British Thoracic Society and Medical Research Council are initiating a three arm phase III trial this year. Patients will receive single agent vinorelbine chemotherapy or combination chemotherapy with mitomycin, vinblastine and cisplatin (“MVP”) or best supportive care. This important trial should help to define the role of chemotherapy in terms of effects on quality of life and survival in malignant mesothelioma.