Interleaf leakage for 5 and 10 mm dynamic multileaf collimation systems incorporating patient motion

Use of dynamic multileaf collimation (DMLC) for intensity modulated radiation therapy (IMRT) is accelerating. Delivery systems have the ailment of interleaf leakage (IL). This is compounded by the inefficiency of IMRT delivery, estimated to be a factor of 5 for DMLC. With IL on the order of 4%, it is possible to deliver as much as 20% of the prescribed dose to nonprescribed regions. However, IL is characterized by narrow Gaussian peaks of approximately 0.5-1.0 mm full-width-half-maximum (FWHM). We performed a leakage study for 5 and 10 mm leaf systems, accounting for intratreatment and intertreatment motions. In solid phantoms, film was placed perpendicular to beams. DMLC patterns delivered step-wedged distributions. The same field was duplicated using a collimating jaw in a segmented fashion to obtain baseline data of primary and scatter contributions. Longitudinal shifts up to 4 mm and angulations up to 4 degrees were introduced during beam delivery by running multiple patterns, to arrive at a composite delivery. The intent of these rigid body motion experiments was to replicate patient motion. Clinical IMRT fields using segmented MLC were also tested. Films were scanned and converted to dose. A microionization chamber confirmed film data at discrete points. In all cases shifts diminished IL peak values. In the step-wedge case, the net 18 MV IL peaks diminished from 3.6% to 3.2% for the 10 mm system. The 5 mm system IL values decreased from 4.0% to 3.2% with a 2 mm shift but increased to 4.0% with 4 mm shifts. The clinical field data followed the same pattern with a washing out of peak values, but the overall transmission to shielded regions slightly increased. Therefore nonprescribed regions are influenced by an effective transmission value rather than discrete peak IL values. The 5 mm leaf system does not introduce increased IL and is an appropriate system for IMRT.