Virtual 5 mm-width multileaf collimation

Purpose/Objective: The use of multileaf collimators (MLC) in place of standard cerrobend blocks has many advantages in static radiotherapy treatments. However, limitations still exist due to the increased effective penumbra and stepped isodose lines produced by conventional MLC, which may not be acceptable in certain clinical situations. These limitations are mainly attributed to the 1 cm leaf widths currently available on commercial MLC. We have implemented a virtual 5 mm-width MLC technique in an attempt to improve these limitations. Materials and Methods: Using the MLC on a Varian 2100C linear accelerator in clinical use, dose distributions were obtained with film dosimetry for 6 and 18 MV photon beams in a water phantom system over a range of block orientations at different depths for conventional MLC, virtual 5 mm-width MLC and cerrobend blocks. The MLC leaf widths on this system are 1 cm at the isocentre. The virtual 5 mm-width MLC technique uses two superimposed MLC fields to irradiate a field. The first field is the standard MLC field with the collimator centre aligned with the field centre. The other MLC field has the centre of the collimator shifted away from the field centre by 5 mm in the direction perpendicular to the leaf movement direction. The composite of these two MLC fields is equivalent to that of a 5 mm-leaf-width MLC. Results: The virtual 5 mm-width MLC resulted in an improvement of the effective penumbra by approximately 1 mm. A more significant effect was found with the 'peak' to 'trough' distances of specified isodose lines which reflects the 'smoothness' of an isodose line. This effect was most pronounced when maximal stepped isodose lines (ie. 45 degree block orientation) occurred. A much smoother field edge was obtained with this technique as compared to conventional MLC. 'Peak' to 'trough' distances were measured to be as great as 4 to 5 mm using conventional MLC blocks, whereas the corresponding 'peak' to 'trough' distances with virtual 5 mm-width MLC were 1 mm or less. This was comparable to the distances measured with cerrobend blocks of 0.5 mm or less. Conclusion: We have demonstrated a method of improving the dose distribution of MLC fields using a virtual 5 mm-width MLC technique. The effective penumbra can be decreased and the stepped isodose lines can be virtually eliminated. This would allow MLC to be used more frequently in place of cerrobend blocks especially when the MLC blocks are near critical normal structures. We have implemented this technique in our clinic and our preliminary experience has found it to be a practical approach