An international study of multitrial data investigating quality of life and symptoms as prognostic factors for survival in different cancer sites.
6002 Background: The prognostic value for survival of HRQOL data derived from self-report questionnaires, has been well documented in cancer research. The objective of this study was to examine the prognostic value of HRQOL parameters for different cancer sites using one standardized and validated patient self-assessment tool. Methods: A total of 11 different cancer sites, pooled from 30 European Organisation for Research and Treatment of Cancer (EORTC) Randomized Controlled Trials (RCTs), were selected for this study. For each cancer site, univariate and multivariate Cox proportional hazard modeling was used to assess the prognostic value (p<0.05) of 15 HRQOL parameters, assessed with the EORTC QLQ-C30 at baseline before randomization, for overall survival. Models were adjusted for the parameters age, gender, distant metastasis, World Health Organization performance status and stratified by clinical study. Results: A total of 7,417 patients completed the EORTC QLQ-C30 before randomization. For brain cancer cognitive functioning (CF) (hazard ratio (HR) =0.95; p<.0001) was prognostic. For breast cancer nausea and vomiting (NV) (HR=1.17; p=0.0011) was a prognostic indicator. For colorectal cancer physical functioning (PF) (HR=0.93; p<.0001), NV (HR=1.07; p<.0001), and appetite loss (AP) (HR=1.07; p<.0001) predicted survival. For esophageal cancer PF (HR=0.88; p=0.0072) and for head and neck cancer NV (HR=1.14; p=0.0097) were prognostic. For lung cancer PF (HR=0.94; p=0.0006) and pain (HR=1.08; p<0.0001), for melanoma dyspnea (HR=1.06; p<.0001), for ovarian cancer NV (HR=1.2; p<.0001), for pancreatic cancer global QOL (HR=0.83; p=0.0073), for prostate cancer role functioning (RF) (HR=0.96; p=0.006) and AP (HR=1.07; p<.0001), and for testis cancer RF (HR=0.81; p=0.0144) were predictors of survival. Conclusions: Our findings show that different HRQOL parameters provide prognostic information for survival for patients with different tumor sites and that no single HRQOL scale can predict survival in all cancer patients. Thus, each cancer site needs careful examination and no single QOL paramenter can predict survival in all cancer diseases.