The Use of Model-Based Tumor-Size Metrics to Predict Survival
Abstract
Mixed-effect models are increasingly being used to analyze the time-course of tumor size and to identify tumor size metrics as predictors of overall survival in cancer patients. However, caution is needed interpreting the results of such analyses and applying them to clinical research. Sparse individual tumor data as a result of patient mortality can lead to falsely conclusions about tumor size metrics as predictors of overall survival.
Domains
Quantitative Methods [q-bio.QM]
Origin : Files produced by the author(s)