Predicting brain tumour growth patterns using a novel MRI-based tumour spread map: application to radiotherapy planning.

BACKGROUND

The treatment of glioblastomas (GBM) with radiation therapy is extremely challenging due to their invasive nature and high recurrence rate within normal brain tissue.

PURPOSE

In this work, we present a new metric called the tumour spread (TS) map, which utilizes diffusion tensor imaging (DTI) to predict the probable direction of tumour cells spread along fiber tracts. We hypothesized that the TS map could serve as a predictive tool for identifying patterns of likely recurrence in patients with GBM and, therefore, be used to modify the delivery of radiation treatment to pre-emptively target regions at high risk of tumour spread.

METHODS

In this proof-of-concept study, we visualized the white matter fiber tract pathways within the brain using diffusion tensor tractography and developed an algorithm which mathematically calculates a relative probability index in each voxel, resulting in the generation of the TS map. Based on the information provided by the TS map, the original radiotherapy target volume was then modified to include areas with a higher probability of tumour spread and exclude other areas with a lower probability of spread. A volumetric modulated arc therapy (VMAT) treatment plan was then developed utilizing the modified target volumes and subsequently compared to that using the original target volumes. Follow-up anatomical imaging obtained 8 months post-surgery was assessed to validate our findings.

RESULTS

A TS map was generated on a glioblastoma patient demonstrating a relative probability of tumour spread along fiber tracts throughout the brain. The modified planning target volume better covered brain regions with a higher risk of tumour spread while still demonstrating a 21% reduction in volume compared to the original planning target volume, resulting in greater preservation of normal tissue. The modified VMAT plan resulted in an average mean dose to four identified recurrences of 80% of the prescription dose, while the original VMAT plan delivered only 63% of the prescription dose as the average mean dose to the recurrences.

CONCLUSION

The utilization of tractography and the generation of corresponding TS maps offer a promising approach to predicting patterns of tumour recurrence and optimizing treatment delivery. Further research is needed to validate the predictive value of the TS map on a larger cohort of patients and explore its potential in personalized treatment strategies for GBM patients.