691 Background: In the absence of mature overall survival (OS) endpoints, interim clinical trial data can be used to predict long-term survival in patients (pts) with advanced malignancies and inform trial continuation, treatment preference, and reimbursement decisions. Research shows that HRQoL assessments can be associated with OS, offering potential utility for validated HRQoL scales. We describe a predictive model used to determine the extent to which HRQoL data could predict CS (survival conditional on progression at either 6 or 12 months). Methods: Using pt-level data from a large phase III trial of nivolumab vs everolimus, a simulation approach with a survival random forest algorithm identified factors statistically important in predicting CS from a large number of covariates measured at baseline. Stepwise Cox proportional hazard survival models were fitted using covariates identified as important. Baseline scores and change over time were tested to determine the influence on the predictive power of the HRQoL data. Results: For both nivolumab and everolimus, baseline FKSI values were significant predictors of CS; median survival times roughly doubled for pts with baseline FKSI scores ≥30 vs pts with scores < 30 (nivolumab, 31.3-16.6; everolimus, 26.6-11). Baseline FKSI scores were the most important predictor vs the other baseline covariates from the survival random forest simulation, and a statistically significant covariate when fitting a stepwise Cox proportional hazard survival model. Change in scores over time influenced CS for pts with high baseline scores, and pts who demonstrated improvement in scores vs baseline had significantly higher CS vs pts with low baseline and no improvement in scores vs baseline. However, when examining change in HRQoL score over time, the statistical importance of the covariate begins to diminish due to high correlations with factors such as adverse events and weight change. Conclusions: We found that HRQoL data, specifically the FKSI, could be useful in predicting CS, especially at the onset of a trial. The importance of the FKSI score in predicting CS could be a powerful complement to existing clinical prognostic factors.