A prognostic index for progression-free survival in malignant mesothelioma with application to the design of phase II trials: A combined analysis of 10 EORTC trials.

Francart, Julie [UCL] Vaes, Evelien [UCL] Henrard, Séverine [UCL] Legrand, Catherine [UCL] Baas, P Gaafar, R. van Meerbeeck, J P Sylvester, Richard J. [UCL] Robert, Annie [UCL] et al.[show all]

PURPOSE: For cytostatic agents or when the response assessment is difficult, adaptations to phase II designs may allow a better assessment of therapeutic activity: first by using the progression-free survival rate (PFSR) as primary end-point instead of the response rate, and second by considering progression-free survival (PFS) risk groups based on a prognostic index (PI). In mesothelioma, current treatments yield disappointingly poor results and there is a need to investigate new regimens. The purpose of this report is to provide a PI for PFS in mesothelioma and reference values for the PFSR. MATERIALS AND METHODS: Data on 523 patients included in 10 European Organisation for Research and Treatment of Cancer (EORTC) mesothelioma studies were analysed to identify prognostic factors using a multivariate Cox regression model. Subsequently, a PI and a nomogram for PFS were developed. The PFSRs at 3, 4, 5 and 6 months were estimated. RESULTS: A performance status>0, stage IV disease and mixed or sarcomatous histological type were indicators of a poor prognosis for PFS. From the PI, based on these three variables, four risk groups were defined. The median progression-free survival ranged from 5.3 to 2.1 months in these risk categories. The PFSRs at 3 months were 70.6%, 62.4%, 54.2% and 42.1% in the four categories, respectively. CONCLUSION: The PI allows dividing patients into homogeneous risk categories in which PFSRs can be calculated and used to design future phase II mesothelioma trials. Defining homogeneous categories of patients avoids dilution of results between groups and improves the assessment of therapeutic activity.