Dynamic Prediction of Near-Term Overall Survival in Patients with Advanced NSCLC Based on Real-World Data

SIMPLE SUMMARY: Patients near the end of life often receive aggressive care, which may be of low value. For patients with advanced cancers, it is standard clinical practice to estimate the prognosis to inform treatment decisions and improve end-of-life care. However, clinical estimates of prognosis may be imprecise and rapidly become out-of-date if clinical factors that evolve over time are not incorporated. Patient prognosis is commonly estimated based on a clinician’s subjective assessment of patient reserve, such as performance status. We propose a spline-smoothed landmarking approach to dynamically estimate survival probabilities based on objective, evolving patient features. The proposed method allows predictions at any time during the patient disease course and demonstrates dramatically improved prediction accuracy compared to methods using clinical features at a fixed time. The proposed approaches can assist clinicians and patients in appropriately regulating treatments to improve outcomes and quality of life. ABSTRACT: Patients with terminal cancers commonly receive aggressive and sub-optimal treatment near the end of life, which may not be beneficial in terms of duration or quality of life. To improve end-of-life care, it is essential to develop methods that can accurately predict the short-term risk of death. However, most prediction models for patients with cancer are static in the sense that they only use patient features at a fixed time. We proposed a dynamic prediction model (DPM) that can incorporate time-dependent predictors. We apply this method to patients with advanced non-small-cell lung cancer from a real-world database. Inverse probability of censoring weighted AUC with bootstrap inference was used to compare predictions among models. We found that increasing ECOG performance status and decreasing albumin had negative prognostic associations with overall survival (OS). Moreover, the negative prognostic implications strengthened over the patient disease course. DPMs using both time-independent and time-dependent predictors substantially improved short-term prediction accuracy compared to Cox models using only predictors at a fixed time. The proposed model can be broadly applied for prediction based on longitudinal data, including an estimation of the dynamic effects of time-dependent features on OS and updating predictions at any follow-up time.