Biocomputing: numerical simulation of glioblastoma growth and comparison with conventional irradiation margins

Object: Estimation of glioblastoma (GBM) growth patterns is of tremendous value in determining tumour margins for radiotherapy. We have previously developed a numerical simulation model for the pattern of spread of glioblastoma tumours. This model involved the creation of a digital atlas of the brain with elasticity and resistance-to-invasion values for specific brain structures and also included probable direction of tumour spread as estimated by Diffusion Tensor Imaging (DTI). The current study is aimed at comparing the outcome of such simulation with conventional irradiation margins currently in use.

Methods: Actual patient data were used to simulate the direction of microscopic extension, using a variety of margin-, proliferation- and diffusion-rate scenarios to generate growth patterns, which were then compared with current standard radiotherapy margins.

Results: Our patient growth pattern simulations showed microscopic invasion beyond irradiation margins for both combinations of high-diffusion/low-proliferation and low-diffusion/high-proliferation rate scenarios. The model also indicated that some healthy brain tissue that was projected to be safe from recurrence fell inside treatment margins.

Conclusion: These results may explain the current inadequacy of our treatment techniques in preventing locoregional recurrences of GBM.