Todd H. Steinberg

A dosimetric comparison of various multileaf collimators

The dosimetric characteristics of three multileaf collimator (MLC) systems (Elekta, Siemens and Varian) having 10 mm leaf width are compared. A 6 MV photon beam was used from each unit for measurements. Film dosimetry was performed for the measurements and the analysis techniques were exactly duplicated in each system. Two of the collimators have rounded leaf ends (Elekta and Varian) and the third (Siemens) has a flat end that follows beam divergence. A scanning densitometer (Wellhöfer with 0.45 mm spot and 0.5 mm step size) was used for film analysis. The dosimetric characteristics studied include: penumbra width (80-20%) as a function of position of the leaf end in the field, inter- and intra-leaf radiation leakage, dose distribution of the tongue and groove, and isodose curves for stepped leaves forming 45 degrees angle beam edge. Results show that MLC designs with divergent and non-divergent leaves produce penumbra (80-20%) widths that are within 2.0 mm of each other. However, the distance of the collimator from the x-ray target plays an important role, and the smallest penumbra width was noted for the Varian MLC despite its rounded leaf-end design. Compared to the other systems, this collimator is positioned about 15 cm closer to the patient which affects the skin dose. The MLC with flat leaf end, although closer to the target, showed slightly poorer penumbra width. Inter-leaf leakage through the leaves is 1.3% for two of the collimators (Elekta and Varian) with the backup jaws and is nearly 1% for the third system (Siemens). The Siemens MLC produces reduced tongue-and-groove effect compared to the other two collimators (Elekta and Varian). The isodose undulation for a stepped edge is found to be significant for the collimator closest to the patient (Varian) and does not depend on the leaf-end shape. There is no perfect MLC system that can be recommended, rather each one has unique advantages and disadvantages that should be weighed with comfort, ease and cost effectiveness for clinical use.

Role of multileaf collimator in replacing shielding blocks in radiation therapy.

To facilitate the use of multileaf collimator (MLC) in field shaping, we tested the hypothesis that the changes in the penumbra due to MLC replacing a Cerrobend block can be related to a change in the margin of the block. We also investigated if it is possible to estimate the effect of MLC replacing a block in terms of a change in the block margin. Calculations were performed for a single field as well as a multiple field setup. For the single field setup, blocks with equal areas were drawn at the four corners of a 16 × 20 cm2 field at angles of 20°, 40°, 60°, and 80° with the horizontal axis. The blocks were then replaced with MLC leaves. For 6 MV x‐rays, dose profiles in the penumbra regions of the blocks at 5‐ and 10‐cm depths were compared with those obtained with the corresponding MLC setup. For multiple fields, the same sets of blocks were set up on the anterio‐posterior (AP‐PA) pair of a four‐field setup. The margins of the blocks were increased (i.e., block shaved) in 1 mm steps to a maximum of 6 mm. The similarity between MLC and the change in the block margin was examined by comparing the dose–volume histogram (DVH) of the normal tissues in the penumbral regions for the different setups. To correlate the effect of MLC with a change in the block margin, difference dose–volume histograms (DDVH) of the normal tissues relative to the original block were compared for the MLC setup with those for the changes in the block margin. The correlation obtained was used to predict the effect on the penumbra region of the MLC setup for the lateral fields of a patient irradiated with a four‐field setup. The calculations were carried out with 15 MV x‐rays. For the single field setup, dose undulation is largest for the 50% isodose line (IDL) as reflected in the largest increase in the 50% to 20% isodose region compared with the 90% to 10% and the 80% to 20% regions. The increase in the penumbral width is largest for the 20° block when replaced by the MLC and is smaller as the angle increases. The increase in the penumbral width also increases with depth. The effect of replacing a Cerrobend block with an MLC is similar to an increase in the block margin. For 15 MV x‐rays, the increase varies inversely with the angle of the blocks, from > 6 mm increase in block margin for the 20° block to about 1 mm for the 80° block. In the clinical example, replacing the blocks in the lateral fields of a four‐field irradiation with MLC is similar to changing the margin of the blocks. For the posterior block, MLC is similar to a 1‐ to 2‐mm increase in the margin of the block, whereas for the anterior block the effect is similar to 1 mm for the straight portion of the block to about 6 mm in the superior portion of the block. Characterization of an MLC setup replacing a Cerrobend block is necessary for adequate coverage of target volume. The effect of MLC replacing a Cerrobend block is similar to a change in the block margin. It is possible to estimate with reasonable accuracy the effect of MLC replacing a Cerrobend block.

Dependence of virtual wedge factor on dose calibration and monitor units.

One of the important features of the Siemens Virtual Wedge (VW) is that the VW factor (VWF) is approximately equal to unity for all beams with a total deviation for a given wedge no greater than 0.05, as specified by Siemens. In this note we report the observed dependence of VWF on dose calibration (cGy/MU), monitor units (MU), and beam tuning for a Primus, a linear accelerator with two dose-rate ranges available for VW operation. The VWF is defined as the ratio of doses measured on the beam central axis for the wedge field to the open field; the open field dose is always measured with the nominal high dose-rate beam. When VW operates in the high dose-rate range, the VWF is independent of calibration (cGy/MU). When VW works in the low dose-rate range, the VWF varies linearly with the calibration of the low dose-rate mode. For a linear accelerator that has only one dose-rate range for VW, there is no observable dependence of VWF on the calibration. We also studied the monitor unit dependence of VWF. A discontinuity in VWF was observed at the switching point between the high and low dose-rate ranges. Working with Siemens, we have investigated causes of this discontinuity. As a result of this investigation, the discontinuity in VWF as a function monitor unit is practically removed.

Performance and beam characteristics of the Siemens Primus linear accelerator

Siemens Primus is a small footprint, klystron driven medical linear accelerator incorporating a compact solid state modulator. A double focused multileaf collimator (MLC) replaces the lower jaw. The first Primus in the world was installed at St. Jude Childrens Research Hospital in early 1997 with x-ray energies of 6 and 15 MV and electron energies of 8, 10, 12, 15, 18, and 21 MeV. The 10 cm depth dose for a 100 cm SSD 10 X 10 cm2 beam is 68% and 77% for 6 and 15 MV x rays, respectively. For both x-ray energies, beam flatness is slightly better than the manufacturers specification of 3% and beam symmetry is considerably better than 1%. The double focus design of the MLC produces a sharp penumbra (5-7 mm at 6 MV and 6-8 mm at 15 MV), increasing modestly with beam size. MLC leaf leakage is less than 1.25%. The depths of the 80% depth dose for the six electron energies of 8, 10, 12, 15, 18, and 21 MeV are 2.6, 3.2, 4.0, 4.9, 6.0, and 7.4 cm, respectively. Beam flatness is typically 2%-3% for all electron energies except 21 MeV, where it reaches 4% for a 25 X 25 cm2 cone. Electron beam symmetry is better than 1% for all energies except 21 MeV, where it is equal to 1%. The results are stored electronically and may be retrieved using anonymous ftp from the American Institute of Physics, Physics Auxiliary Publication Service.