Predicting the growth probability function of tumors in medical images

The dynamics of a tumor can be studied using a feature-based stochastic method to predict the temporal and spatial growth of the tumor. A posterior probability of growth function is incorporated into the interacting particle model definition, and the probability influences growth direction at each location. Using features derived from images and data-mining, the growth probability function is predicted and tested to investigate the ability of the derived feature values to explain the tumor evolution.