Personalized Radiotherapy Planning Based on a Computational Tumor Growth Model

Area of Science:

• Medical Physics
• Computational Biology
• Oncology

Background:

• Personalized radiotherapy planning for brain tumors requires accurate tumor modeling and dose distribution.
• Current methods often lack patient-specific personalization and uncertainty quantification.

Purpose of the Study:

• To present a proof of concept for automatic, personalized radiotherapy planning for brain tumors.
• To integrate computational tumor growth and cell survival models with patient-specific MRI data.
• To account for uncertainties in model parameters and MRI segmentation.

Main Methods:

• Developed a computational model combining glioblastoma growth and exponential cell survival.
• Personalized the model using patient MRI data, considering single or multiple acquisitions.
• Incorporated uncertainty quantification for model parameters and segmentation.
• Defined prescription dose distribution based on computed tumor cell densities and survival models.
• Applied two methods (MAP or expected density) to minimize integral tumor cell survival.

Main Results:

• Demonstrated patient-specific radiotherapy planning conformal to tumor infiltration.
• Showcased the ability to spare adjacent organs at risk through dose re-distribution.
• Validated the approach on two high-grade glioma patients.

Conclusions:

• The proposed computational approach offers a proof of concept for personalized radiotherapy planning in brain tumors.
• This method has the potential to improve tumor targeting and organ-at-risk sparing in clinical practice.