Tumor growth modeling based on dual phase CT and FDG-PET

In this paper, we presented a method to deal with tumor growth prediction using multimodality non-invasive clinical imaging data. We developed a reaction-diffusion tumor growth model that (1) relates cell metabolic rate and tumor growth, and (2) is driven by clinical imaging data. The metabolic rate was incorporated into the model through cell proliferation rate of the model via energy conservation law. FDG-PET scan was employed to provide non-invasive measurement of the metabolic rate. To bridge the gap between the model prediction and the clinical observation, we introduced intracellular volume fraction (ICVF) using dual phase CT scans to link them. The patient specific model parameters were estimated by minimizing the deviation between the predicted ICVF and the measured ICVF with deformation corrected. The experiment was conducted on two pancreatic neuroendocrine tumors. The average surface distance between the predicted tumor and the reference tumor was 2.1 mm and 2.7 mm, respectively, and the mean square difference of the ICVF map was 5.4% and 3.8%, respectively.