Talk Title:

Inference of field cancerization using stochastic models to improve cancer control

Talk Summary:

Early cancer detection strategies focus on screening individuals to find obvious precancerous lesions and removing them when possible. However, screening programs often suffer from both under-diagnosis due to inadequate screening, and over-diagnosis due to ineffective patient stratification in surveillance protocols. To overcome these shortcomings, I will present mathematical methods that focus on detecting aspects of “field cancerization” to better define and target the dynamics of cancerized phenotypes evolving with age within the design of early detection strategies. Here we will incorporate various multiscale data, ranging from population-level incidence to tissue-level (epi)genomic changes, into stochastic models to understand the process of clonal evolution leading to malignancy in the gastrointestinal tract. I will present a novel molecular clock in Barrett’s esophagus to infer patient-specific Barrett’s onset age (a clinically unobservable event) based on changes in DNA methylation, and end by discussing extensions of these methods to include additional tissue growth dynamics.