Abstract
Prediction is ubiquitous across the spectrum of cancer care from screening to hospice. Indeed, oncology is often primarily a prediction problem; many of the early stage cancers cause no symptoms, and treatment is recommended because of a prediction that tumor progression would ultimately threaten a patient's quality of life or survival. Recent years have seen attempts to formalize risk prediction in cancer care. In place of qualitative and implicit prediction algorithms, such as cancer stage, researchers have developed statistical prediction tools that provide a quantitative estimate of the probability of a specific event for an individual patient. Prediction models generally have greater accuracy than reliance on stage or risk groupings, can incorporate novel predictors such as genomic data, and can be used more rationally to make treatment decisions. Several prediction models are now widely used in clinical practice, including the Gail model for breast cancer incidence or the Adjuvant! Online prediction model for breast cancer recurrence. Given the burgeoning complexity of diagnostic and prognostic information, there is simply no realistic alternative to incorporating multiple variables into a single prediction model. As such, the question should not be whether but how prediction models should be used to aid decision-making. Key issues will be integration of models into the electronic health record and more careful evaluation of models, particularly with respect to their effects on clinical outcomes.