GEOMETRIC AND DOSIMETRIC ANALYSIS OF MULTILEAF COLLIMATION CONFORMITY

Background and purpose: There is concern over the stepped edges of mul tileaf collimator (MLC) fields for target coverage compared with those of cerrobend. Despite recent reports dispelling this concern, users a re still cautious when using MLCs for small fields. Leaf orientation c an be a problem if one is required to orient the leaves along an axis not ideal for conformity (such as dynamic or universal wedge cases). I n this study we examined the dependence of MLC field conformity on fie ld size and elongation. Materials and methods: We examined circles of varying diameter and ellipses of varying eccentricity with leaves orie nted along the major and minor axes of ellipses. The tests were both g eometric, comparing areas of overblocking or underblocking (leaves ins ide and outside the field), and dosimetric, using radiographic films a t depth in the beam's eye view plane. Results: For the geometric compa rison there is a rapid increase in non-conformity, defined as the perc entage of overblocking or underblocking area, as the circle diameter d ecreases. For ellipses, when the leaves move along one axis direction, the conformity does not depend on the diameter of, the same axis, but instead improves as the dimension of the axis in the non-leaf motion direction increases. The best conformity is achieved when the maximum number of leaves is used to shape the field. When the dosimetry is ana lyzed, the predictability of these trends decreases due to the impact of undulations (scatter), leaf inaccuracies and dosimetric uncertainti es. Conclusions: We recommend that for small round fields MLC should b e used with caution and that for ellipses the direction of leaf moveme nt should be aligned with the minor axis whenever possible. Though the se experiments are for idealized geometries, the observations can be a pplied to clinical fields. An MLC with a thinner leaf width could be b eneficial for small round fields. (C) 1998 Elsevier Science Ireland Lt d.