A data‐driven calibration of a nonlinear mechanistic model for DNA damage and repair: applications to radiosurgery and heavy ion irradiation treatment for glioblastoma

2014 
Current radiation treatment options for the invasive primary brain tumor glioblastoma multiforme (GBM), consist of two modalities which are very different in terms of the amount of dose delivered, the distribution of dose in space, and perhaps, the biological effectiveness. On one hand, the current standard of care for GBM consists of “conformal” or intensity modulated radiation therapy (IMRT) with daily treatments of 1.8 – 2 Gy over the course of several weeks to a total dose of approximately 60 Gy to a large treatment volume. On the other hand is stereotactic radiosurgery (SRS), which is a secondary radiation therapy treatment used after the disease has recurred. The spatial localization of the highly focused “radiosurgical” dose is achieved through the composition of small 3–5 mm spherical targets, created with multiple small beams. SRS is typically a single fraction treatment, with doses of up to 24 Gy or higher and used primarily for small lesions. It is a matter of contemporary debate as to whether or not biological response to radiation changes for doses higher than 10 Gy per fraction. Because of the different dose per fraction, dose delivery times, and potentially different biological responses to these two radiation treatments, mechanistic models of radiation‐induced DNA damage and repair are often used to quantify and translate radiation dose into biological effect. I present a mechanistic two‐compartment nonlinear ODE model of radiation‐induced DNA damage and repair, which includes sub‐lethal and fatal DNA classes of damage which is based on physically measurable quantities of the radiation treatment. Analytic solutions for this model can be found and demonstrates orders of magnitude differences from the linearized approximation used pervasively in the literature, particularly in the SRS high dose range. Further, data‐driven parameterization of the fully nonlinear model reveals superior model prediction and parameter stability across a wide range of experimental conditions compared to current model paradigms such as the linear‐quadratic model. The doseresponse data used to test the model includes a wide range of particles, energies, doses, dose rates and dose‐fraction timing, motivated by current trends in radiation oncology, including fractionated SRS and heavy ion therapy. This mechanistic modeling approach at the DNA level is connected to a patient‐specific tissue level reaction‐diffusion model for GBM which includes spatial and temporal delivery and response to radiation therapy to investigate the net effect of these novel radiation treatment strategies in silico.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    0
    References
    0
    Citations
    NaN
    KQI
    []