Image guided personalization of reaction-diffusion type tumor growth models using modified anisotropic eikonal...

Area of Science:

• Computational Biology
• Medical Imaging
• Mathematical Oncology

Background:

• Reaction-diffusion models are key for simulating brain glioma growth.
• Integrating medical imaging enhances spatial accuracy in these models.
• Patient-specific model adaptation remains a challenge.

Purpose of the Study:

• To develop and validate a parameter estimation method for reaction-diffusion tumor growth models using time-series medical images.
• To enable accurate adaptation of general models to individual patient cases.
• To improve the simulation of patient-specific tumor evolution.

Main Methods:

• Proposed a parameter estimation technique for reaction-diffusion models based on sequential medical images.
• Formulated tumor evolution consistent with observed image data and reaction-diffusion dynamics.
• Analyzed parameter identifiability and estimation accuracy using synthetic and real patient data.

Main Results:

• Demonstrated unique identification of several model parameters when fixing the proliferation rate.
• Showed accurate estimation of tumor growth speed, irrespective of the fixed proliferation rate value.
• Successfully simulated patient-specific tumor evolution using estimated parameters.
• Achieved promising preliminary results when applying the method to two real clinical cases.

Conclusions:

• The proposed method effectively estimates patient-specific parameters for reaction-diffusion tumor growth models.
• Accurate tumor growth simulation is achievable for individual patients.
• This approach holds potential for personalized treatment planning in brain gliomas.