A Generative Approach for Image-Based Modeling of Tumor Growth

Abstract

Extensive imaging is routinely used in brain tumor patients to monitor the state of the disease and to evaluate therapeutic options. A large number of multi-modal and multi-temporal image volumes is acquired in standard clinical cases, requiring new approaches for comprehensive integration of information from different image sources and different time points. In this work we propose a joint generative model of tumor growth and of image observation that naturally handles multimodal and longitudinal data. We use the model for analyzing imaging data in patients with glioma. The tumor growth model is based on a reaction-diffusion framework. Model personalization relies only on a forward model for the growth process and on image likelihood. We take advantage of an adaptive sparse grid approximation for efficient inference via Markov Chain Monte Carlo sampling. The approach can be used for integrating information from different multi-modal imaging protocols and can easily be adapted to other tumor growth models.

Imaging genomics in cancer research: limitations and promises

“…86 Segmentation methods that explicitly incorporate biophysical models of tumour growth, in a way to facilitate imaging-based segmentation, have also been proposed. 87,88 Although validation of these methods is a very challenging and effort-demanding task, some international efforts for creating validation platforms have started to emerge. A prime example is the Brain Tumor Segmentation challenge organized annually, which uses TCIA and other public data sets, along with ground truth, to evaluate a variety of algorithms.…”

Numerical search for a potential planet sculpting the young disc of HD 115600

“…We conclude that this work demon-strates that the MCMC scheme is a promising method to efficiently explore the parameter space of dynamical simulations and allows us to localize a better-fit region faster than the grid method, although we also stress that the limited number of simulations makes reaching the convergence state quite challenging. There are numerous schemes existing in the literature to explore multidimensional parameter spaces which could represent very good alternatives, such as the sparse grid MCMC (Menze et al 2011) -a scheme combining the use of both a grid and MCMC, or the use of grids of different resolutions successively (Horner et al 2013). The debris disc of HD 115600 shows interesting features which can be reproduced by simulating the gravitational influence of an inner massive planet on the disc structure.…”

Predicting the Location of Glioma Recurrence after a Resection Surgery

“…Tracqui et al [17] proposed 40% maximal tumor cell density to be visible in T2 MRIs, Konukoglu et al [9] used Tracqui's value, and Swanson et al [16] used a value of 2%. Menze et al [13] suggested the maximal tumor cell density that is visible in Flair MRIs to be 9.5%. We chose the tumor cell density threshold of visibility value as 20% because it is an intermediate value in literature for T2 MRIs, which includes Flair.…”