The impact of fluctuations in intensity patterns on the number of monitor units and the quality and accuracy of intensity modulated radiotherapy

The purpose of this work is to examine the potential impact of the frequenc y and amplitude of fluctuations ("complexity") in intensity distributions o n intensity-modulated radiotherapy (IMRT) dose distributions. The intensity -modulated beams are efficiently delivered using a multileaf collimator (ML C). Radiation may be delivered through a continuous (dynamic mode) or discr ete (step-and-shoot) sequence of windows formed by the leaves. Algorithms a nd software that convert optimized intensity distributions into leaf trajec tories apply approximate empirical corrections to account for the various e ffects associated with MLC characteristics, such as the rounded leaf tips, tongue-and-groove leaf design, leaf transmission, leaf scatter, and collima tor scatter upstream from the MLC. Typically, the difference between inter- and intraleaf transmissions is ignored. In this paper, using a schematic e xample of IMRT for head and neck carcinomas, we demonstrate that complex an atomy and severe optimization constraints produce complex intensity pattern s. Using idealized intensity patterns we also demonstrate that, for complex intensity patterns, the average window width tends to be smaller and, for the same dose received by the tumor, the number of MUs is larger. We found that as the complexity increases, so does the contribution of radiation tra nsmitted through and scattered from the leaves ("indirect radiation") to th e total delivered dose. As a consequence, the lowest deliverable intensity in complex intensity patterns may be significantly greater than that requir ed to provide adequate protection for some normal tissues. Furthermore, sin ce corrections for leaf transmission and scatter effects are approximate an d the difference between inter- and intraleaf transmission is ignored, the accuracy of the delivered dose may be affected. Using the results of a simp le experiment and a typical intensity-modulated beam for a head and neck ca se as examples, we show the effect of window width and complexity on the ac curacy and deliverability of intensity patterns. Some possible strategies f or improving the accuracy and for relaxing the lower limit on deliverable i ntensity are discussed. (C) 2000 American Association of Physicists in Medi cine. [S0094-2405(00)00306-0].