Modeling dynamic, nutrient-access-based lesion progression using stochastic processes


Conference Paper

© SPIE. Downloading of the abstract is permitted for personal use only. Simulation methods can be used to generate realistic, computational lesions for insertion into anatomical backgrounds for use in a virtual clinical trial framework. Typically, these simulation methods rely on clinical lesion images|with resolution many times the size of a cell|to produce a lesion that is time- and anatomical- location{invariant. Though, in reality, a lesion's morphology and growth rate are dependent on both of those things. The goal of this work was to produce a lesion model starting from simple assumptions about the behavior of proliferating cells, simulate their states over time, and produce a lesion model for which the morphological features are determined by known cellular properties. Each cell of each lesion that is simulated can exist in one of several states depending on its access to nutrients and potential for proliferation at a given time. Running these simulations a sufficiently large number of times under the same conditions yields the most probable lesion for a given set of constraints|or, specifically for this work, a given anatomical environment. These lesions can be used in studies in which detection of subtle pathological features on small scales are essential to obtain meaningful results.

Full Text

Duke Authors

Cited Authors

  • Sauer, TJ; Samei, E

Published Date

  • January 1, 2019

Published In

Volume / Issue

  • 10948 /

International Standard Serial Number (ISSN)

  • 1605-7422

International Standard Book Number 13 (ISBN-13)

  • 9781510625433

Digital Object Identifier (DOI)

  • 10.1117/12.2513201

Citation Source

  • Scopus