Impact of health-related quality of life (HRQoL) on short-term mortality in patients with recurrent ovarian, fallopian or peritoneal carcinoma (the NOGGO-AGO QoL Prognosis-Score-Study): results of a meta-analysis in 2209 patients

OBJECTIVE: Recurrent ovarian cancer is an incurable disease with variable but poor prognosis. Health-related quality of life (HRQoL) is a patient-reported outcome measure generally applied to measure effects of therapies. Our aim was the development and validation of a risk score for the prediction of short-term mortality using the combination of sociodemographic and clinical factors and HRQoL. METHODS: For exploratory and validation analysis, the North-Eastern German Society of Gynecological Oncology (NOGGO) and Working Group Gynecological Oncology (AGO) study databases were screened for trials. Only trials which obtained defined HRQoL measurements were included in the final analysis. Multivariable logistic regression analyses were used to identify risk factors and their weighting for the risk score. Modulation with cubic regression analyses revealed median survival and short-term mortality defined as 1-year mortality for each value. RESULTS: For exploration, 974 patients from three clinical studies of the NOGGO and for validation, 1235 patients from several clinical studies of the AGO were eligible. The risk score included platinum-free interval, performance status, age, global QoL and nausea/vomiting. Receiver operating characteristic analysis showed a good predictive value with an area under the curve of 0.81 for model 1 in the exploration and 0.74 in the validation. Short-term mortality in model 1 was 8.2%, 23.5% and 58.4% in the exploration sample, and 19.7%, 38.1% and 63.4% in the validation sample for patients under low, medium and high risk, respectively. CONCLUSIONS: This risk score discriminates well between recurrent ovarian cancer patients under low, medium and high risk of short-term mortality. It may help to identify a risk group under high risk for short-term mortality that can be used for randomization in clinical trials and may support decision making for palliative chemotherapy.