A comparison of multileaf-collimator and alloy-block field shaping

The purpose of this report was to compare the dose distribution at a field edge defined with divergent alloy blocks to the distribution obtained with a multileaf collimator (MLC). The comparison is made for simple block replacement situations.

A tertiary multileaf collimator mounted on a linear accelerator operating at 6 MV was compared to divergent alloy blocks positioned at the level of the blocking tray. The leaves of the MLC were positioned to give maximum stepping (leaf displacement equals leaf width), and the blocking produced the same field shape. Three different treatment plans were compared: single field, opposed fields, and a four-field "box." Dose distributions were determined using radiographic film scanned with a laser densitometer with a 0.45-mm spot size. One experiment was repeated using radiochromic film with reduced energy dependence. Dose distributions were examined on the isocenter plane, and on planes displaced by 1.0 and 2.5 cm. The effect of daily setup variations was also studied by comparing a single fraction treatment with a fractionated treatment consisting of 15 fields slightly displaced relative to each other. The magnitude of these displacements was determined using available literature on treatment reproducibility.

For a single field plan, maximum stepping of an MLC-defined edge produces an obvious undulating dose pattern compared to an alloy block edge. At the isocenter plane, this pattern is unchanged when parallel opposed fields are used. However, blurring occurs for both MLC and block edges when planes displaced from the isocenter are examined. The gradient for the block edge is 8%/mm for opposed fields and a plane 2.5 cm from the isocenter, compared to 15%/mm for the isocenter plane. Adding two additional fields does not change the dose pattern in the isocenter plane, but does reduce the gradient across the steepest portion of the penumbra to 8%/mm, and shifts the isodose line with the most pronounced stepping to higher values (from 50 to 80%). Introducing daily setup variations results in a reduction of the sharp dose gradient along the sides of a single field, and around the periphery of the beam at the isocenter plane of opposed fields. Smaller changes are found for edges already blurred by other factors. Radiochromic film was generally noisier than radiographic film, but comparison of the two films did not show a significant difference, indicating that the energy dependence of the radiographic film was not a problem.

The obvious dose stepping seen on a portal image of a single field with MLC shaping is shown to be partially erased by the addition of other fields, and for planes away from the isocenter. However, the effects of daily setup variations must be included to more effectively blur dose stepping along the external envelope of a single field or near the isocenter plane of opposed fields. This result conflicts with attempts to improve immobilization.