|
Abstract:
|
Many studies have been conducted to establish genomic correlates of cancer risk. These studies open the door to precision approaches for cancer prevention and early detection. But knowing that one is at a high risk of specific cancer does not immediately translate into a strategy for mitigating the risk. Without a direct path from risk to action, and given the costs of screening an entire healthy population for their risk, can we be confident that all the work on genomic risk prediction will fulfill its promise? In this presentation, I will review the history of genomic risk prediction, showcase some prominent genomic risk studies, and discuss the process of evaluating the performance of genomic risk predictions. For risk prediction tools to fulfill their promise, I propose that in addition to being accurate they should be meaningful and actionable. There must be a process for translating their results into a targeted cancer control strategy that is preferred over a one-size fits all approach. I will share a process we have piloted for making a risk prediction tool actionable in the context of prostate cancer early detection and derive some general lessons for broader application.
|
