Abstract

Early cancer detection strategies in esophageal adenocarcinoma focus on screening individuals at particular ages to check for the precursor Barrett’s esophagus (BE), and then monitoring those patients found with BE over time. Due to lack of randomized control trial data, screening programs suffer from levels of both under-diagnosis and over-diagnosis as the consequence of sub-optimal risk stratification and timing of initial screening and surveillance. To address these shortcomings, I will present mathematical methods for inferring dynamics of field cancerization in BE, such as estimating the patient’s age at onset of BE (a clinically unobservable event). Here we will incorporate multiscale data, ranging from population-level incidence to tissue-level epigenetic changes, into stochastic models of clonal evolution. I will present a Bayesian molecular clock model to infer patient-specific Barrett’s onset age based on changes in DNA methylation and discuss extensions of these methods to include spatial growth dynamics.

Short bio: Dr. Kit Curtius obtained a bachelor’s degree in Mathematics at UCLA and then completed a PhD in Applied Math at University of Washington in 2015. During her PhD, Kit’s research focused on mathematical modeling of cancer within the Computational Biology Program at the Fred Hutchinson Cancer Research Center, Seattle, WA. In 2016, she moved to the UK to do a postdoc with Prof Trevor Graham in the Evolution and Cancer Lab at Barts Cancer Institute, London. Kit held an MRC Rutherford Fellowship at BCI before returning to the US in July 2020 to start her group, the Quantitative Cancer Control Lab, as an Assistant Professor at UCSD in the Department of Biomedical Informatics and Moores Cancer Center.