Prognostic factors in patients with advanced cancer: a comparison of clinicopathological factors and the development of an inflammation-based prognostic system
A partir de données portant sur 2 456 patients atteints d'un cancer de stade avancé, cette étude évalue l'intérêt d'un score basé sur deux marqueurs inflammatoires pour prédire la survie des patients
Purpose: In advanced cancer, oncological treatment is influenced by performance status (PS), however this has limitations. Biomarkers of systemic inflammation may have prognostic value in advanced cancer. The study compares key factors in prognosis (PS, patient reported outcomes) against an inflammation-based score (Glasgow Prognostic Score - mGPS). A new method of prognosis in advanced cancer (combining PS and mGPS) is tested and then validated.
Experimental Design: Two international biobanks of advanced cancer patients were analysed. Key prognostic factors (PS, patient reported outcomes (EORTC QLQ-C30) and mGPS (using C-reactive protein and albumin concentrations)) were examined. The relationship between these and survival was examined using Kaplan-Meier and Cox regression methods, in a test sample before independent validation.
Results: Data were available on 1825 patients (test) and 631 patients (validation). Median survival ranged from 3.2 months (test) - to 7.03 months (validation). On multivariate analysis, PS (HR 1.62-2.77) and mGPS (HR 1.51-2.27) were independently associated with, and were the strongest predictors of survival (p<0.01). Survival at 3 months varied from 82% (mGPS 0) to 39% (mGPS 2) and from 75% (PS 0-1) to 14% (PS 4). When used together survival ranged from 88% (mGPS 0, PS 0-1) to 10% (mGPS 2, PS 4), p<0.001.
Conclusion: A systemic inflammation-based score, mGPS, and PS predict survival in advanced cancer. The mGPS is similar to PS in terms of prognostic power. Used together, PS and mGPS act synergistically improving prognostic accuracy. This new method may be of considerable value in the management of advanced cancer patients.
Clinical Cancer Research , résumé, 2013