Physical and dosimetric characteristic of high-definition multileaf collimator (HDMLC) for SRS and IMRT

Physical and dosimetric characteristics of HDMLC were studied for SRS6, 6, and 10 MV X-rays from Novalis Tx. This in-built tertiary collimator consists of 60 pairs (32 x 0.25 cm; 26 x 0.5 cm and 2 x 0.7 cm) of leaves. Properties of HDMLC studied included alignment, readout and radiation field congruence, radiation penumbra, accuracy and reproducibility of leaf position and gap width, static and dynamic leaf shift, tongue-and-groove effect, leaf transmission and leakage, leaf travel speed, and delivery of dynamic conformal arc and IMRT. All tests were performed using a calibrated ionization chamber, film dosimetry and DynaLog file analysis. Alignment of leaves with isocenter plane was better than 0.03 cm at all gantry and collimator positions. The congruence of HDMLC readout and radiation field agreed to within +/- 0.03 cm for filed sizes ranging from 1 x 1 to 20 x 20 cm(2). Mean 80% to 20% penumbra width parallel (perpendicular) to leaf motion was 0.24 +/- 0.05 (0.21 +/- 0.02) cm, 0.37 +/- 0.12 (0.29 +/- 0.07) cm, and 0.51 +/- 0.13 (0.43 +/- 0.07) cm for SRS6, 6, and 10 MV X-rays, respectively. Circular field penumbra was comparable to corresponding square field. Average penumbra of 1 x 20 cm(2) field was effectively constant over off-axis positions of up to 12 cm with mean value of 0.16 (+/- 0.01) cm at 1.5 cm depth and 0.38 (+/- 0.04) cm at 10 cm depth. Minimum and maximum effective penumbra along the straight diagonal edge of irregular fields increased from 0.3 and 0.32 cm at 70 degrees steep angle to 0.35 and 0.56 cm at 20 degrees steep angle. Modified Picket Fence test showed average FWHM of 0.18 cm and peak-to-peak distance of 1.99 cm for 0.1 cm band and 2 cm interband separation. Dynamic multileaf collimation (DMLC) output factor remained within +/- 1% for 6 MV and +/- 0.5% for 10 MV X-rays at all gantry positions, and was reproducible within +/- 0.5% over a period of 14 months. The static leaf shift was 0.03 cm for all energies, while dynamic leaf shift was 0.044 cm for 10 MV and 0.039 cm for both SRS6 and 6 MV X-rays. The dose depression and corresponding tongue-and-groove size were 24% and 0.17 cm for 6 MV and 19% and 0.20 cm for 10 MV X-rays. Average transmission through HDMLC was 1.09%, 1.14% and 1.34% for SRS6, 6 and 10 MV X-rays. Analysis of DynaLog files for leaf speed test in arc dynamic mode, delivery test of dynamic conformal arc, and step-and-shoot and sliding window IMRT showed at least 95% or more of the error counts had misplacements < 0.2 cm, with maximum root mean square (RMS) error value calculated at 0.13 cm. Accurate and reproducible leaf position and gap width, and less leakage and small consistent penumbra over the fields demonstrate HDMLC suitable for high-dose resolution SRS and IMRT.