Adaptive intensity modulated radiation therapy planning optimization with changing tumor geometry and biology enforcing both cumulative and fraction size dose constraints

Intensity Modulated Radiation Therapy (IMRT) is a modern technique of delivering radiation treatments to cancer patients. In IMRT technology, intensities must be chosen for the many small unit grids into which the beams are divided to produce a desired distribution of dose at points throughout the body with the goal of maximizing dose delivered to the tumor while sparing healthy tissues from excessive radiation and keeping dose homogeneous across the tumor. Although IMRT plans are optimized as a single overall treatment plan, they are delivered over 30--50 treatment sessions (fractions) and both cumulative and per-fraction dose constraints apply.;The extended time period of treatment allows for periodic re-imaging of the changing tumor geometry and for adapting the treatment plan accordingly. This research presents promising iterative optimization approaches that re-optimize and update the treatment plans periodically by incorporating the latest tumor geometry information. Two realistic lung cases simulating practice, based on anonymized archive datasets, are used to test the effectiveness of the proposed adaptive planning approaches. The computed optimal plans both satisfy cumulative and per-session dose constraints while improving the objective (average tumor dose) as compared to non-adaptive treatment.;In addition to tracking tumor geometrical changes through the treatment, recent advances in imaging technology also provide more insight on tumor biology which has been traditionally disregarded in planning. The current practice of delivering homogeneous physical dose distributions across the tumor can be improved by nonhomogeneous distributions guided by the biological responses of the tumor points. This research is one of the first efforts in developing radiation therapy planning optimization methods with tumor biology information while maintaining both cumulative and per-fraction dose constraints. The proposed biological optimization models generate treatment plans reacting to the tumor biology prior to the treatment as well as the changing tumor biology throughout the treatment. The optimization models are tested on a simulated head and neck test case. Results show computed biologically optimized plans improve on tumor control obtained by traditional plans ignoring biology, and also with adaptive over non-adaptive methods.