The Optimization of Inverse Planning and IMRT

All inverse planning involves weight optimization. Beamlet-based Inverse Planning (BBIP) irradiates complex target volumes with a large number of small beams. The number of small beams or beamlets is often more than one thousand. The optimization process adjusts the weights of these beams to produce a desired dose distribution. Delivery of this large number of small beams is challenging, but many clever techniques have been devised to handle this problem. For example, the delivery step can use tomotherapy with a continuously moving gantry and a fast binary collimator system. Or, fixed gantry angles can be employed with a standard multileaf operating in a Dynamic Multileaf Collimator (DMLC) mode. It is also possible to move the gantry and MLC simultaneously for Intensity Modulated Arc Therapy (IMAT). Some of these approaches can directly model the large number of small beams needed to produce complex dose distributions, but it is also possible to reduce the complexity of the problem with the use of interpreter (also known as sequencer or translator) software that combines the beamlets into a smaller number of larger apertures prior to delivery. Depending on a number of factors like the spacing of the intensity levels used in the reduction process, the final number of apertures often exceeds 100 and can go considerably higher. This large number of fields per patient can extend treatment times to unacceptable levels, and can result in a total number of field changes that exceeds the design capabilities of any multileaf collimator and lead to unnecessary system failures. Taking the example of 20 IMRT patient treatments per day and 100 apertures per patient gives 2000 field changes each day if these patient’s are treated on the same accelerator using a step-and-shoot delivery approach.