Predicting Survival for Patients With Metastatic Disease

Area of Science:

• Oncology
• Medical Physics
• Biostatistics

Background:

• Accurate prognostication is crucial for guiding treatment decisions in metastatic cancer.
• Radiation oncologists' ability to predict patient survival is an area requiring further investigation.

Purpose of the Study:

• To assess the accuracy of radiation oncologists in predicting survival for patients with metastatic disease undergoing radiation therapy.
• To identify factors influencing the accuracy of these survival predictions.

Main Methods:

• Prospective study involving 22 radiation oncologists estimating survival for 689 patients.
• Survival predictions were categorized (0-6, >6-12, >12-24, >24 months) and compared to observed survival (OS).
• Statistical analyses included Stuart's Tau C and logistic regression models to identify predictors of accuracy.

Main Results:

• Overall accuracy in survival prediction was 39.7%, with 26.5% underestimations and 33.9% overestimations.
• Karnofsky performance status was the most significant predictor of accuracy, with better accuracy for patients with shorter OS and lower performance status.
• Accuracy varied by patient age, primary site, and race; physician experience did not correlate with accuracy.

Conclusions:

• Radiation oncologists demonstrated approximately 40% accuracy in predicting patient survival.
• Further research is needed to understand the impact of survival estimates on treatment choices and to enhance prediction accuracy.