Intensity modulated radiation therapy (IMRT) is a new modality of radiation therapy and shows significant promise for improving dose conformation to the tumor'target and dose avoidance to the sensitive structures. At present, clinical IMRT treatment plans are still based on primitive dose optimization algorithms. Many uncontrollable degrees of freedom exist in the system and constitute an additional multi-dimensional parameter space coupled to the beam profiles, making rigorous optimization intractable. As thus, the IMRT plans used for clinical treatments are often sub-optimal and the clinical endeavors to deliver the best possible treatment to cancer patients are compromised. This proposal is aimed to establish a novel dose optimization framework using a statistical analysis method and to provide superior computational tools for IMRT planning. We hypothesize that the statistical analysis-based inverse planning scheme will provide substantially improved dose distributions that are required to achieve greater local tumor control while reducing complications of sensitive structures.
The specific aims of the work are (1) to establish a statistical analysis-based dose optimization framework for IMRT treatment planning;(2) to implement an effective beam's eye view dosimetrics (BEVD) to aid the beam placement in IMRT planning and develop a BEVD- guided optimization algorithm to automate the selection of beam orientations;and (3) to demonstrate the impact of the new inverse planning system for IMRT through the study of a series of phantom cases and a number of previously treated patients of various disease sites. Successful completion of these aims will provide radiation oncologists with a substantially improved means of generating highly conformal dose distributions and fully realize the technical capacity of IMRT. This will significantly reduce the toxicity associated with radiation therapy and/or allow us to escalate tumor dose to the levels determined by the physical limit of IMRT other than by current sub-optimal optimization algorithms. Thus it will likely result in a measurable improvement in treatment outcome and have a widespread impact on cancer management.
Showing the most recent 10 out of 44 publications
