Investigation Of The Intensity Profile Optimization In IMRT

Radiation therapy has become a normal technology to cure cancer, and Intensity-modulated radiation therapy (IMRT) has been a major research focus in medical physics for the last decade. IMRT is based on the use of optimized non-uniform radiation beam intensity incident on the patient to make the high-dose treated volume to encompass the 3D target volume, at the same time minimizing the dose to surrounding organs at risk. The advantage of IMRT is to protect organs at risk and to create enough dose in the target volume. However, the intensity beam optimized by IMRT is more complicated and difficult to deliver by multileaf collimator (MLC), because of the large number of segments or the much more delivery time. Now many researchers focus the smoothing of intensity profile and develop a method of smoothing intensity profile to optimize intensity profile to meet some evaluated parameters to dose distribution when the intensity profile complexity is much less as possible. On the optimization algorithm, Multiple-shifted-beamlet-matrices (MSBM) method is presented as a new method of IMRT to produce the smoothing beam, and a new algorithm of segmentation is brought forward for incident beam intensity profile to decrease the number ofsegments.Firstly, we have finished verifying the existence of the degenerated solutions in IMRT optimization. In this process, intensity profile optimizations for a simulated case have been performed by starting with different initial conditions in simulated space conditions, which have soft constraints, dose-volume constraints, dose-volume and hard constraints, respectively. The resultant DVHs for the PTVs are the same and the DVHs for OARs satisfy given constraints, although the intensity profiles could be different. These results verify that there exist degenerated solutions for theoptimization problem even with above constraints included.Second, a new method , based on the degenerated solutions for the optimization problem, was brought up to improve spatial resolution and smoothness of intensity profiles in IMRT treatment planning, and is named as multiple-shifted-beamlet-Matrices (MSBM). In contrast to the conventional approach, the single beamlet matrix for each beam, we create several beamlet matrices for each beam. Those matrices are shifted relatively in the direction of the MLC's leaf travel. This method achieves finer spatial resolution for the intensity distribution than the given beamlet size. The results show that this method can produce optimized dose distributions that are similar to those produced by the conventional inverse planning method with the benefit of smoother intensity profiles that are easier to deliver with a computer controlled MLC. The method produce segments no less than the given beamlet size, so all segments can be implemented.At last, based on Freeman in the computer graphics and images to obtain the outline of image, we have advanced an algorithm of segmentation for 'step-and-shoot mode' in IMRT to decrease the number of segments and improve delivery effect. We applied Freeman to estimate the concavity complexity of the intensity level matrices in orthogonal direction of the row and column in order to choose the travel direction of the MLC's leaves and decrease the number of the segments. Thus this way gets rid of the conventional method of the given direction of the MLC's leaves.There is much significance of the MSBM and the segment algorithm for clinic application.