Performance of a Multi Leaf Collimator System for MR-Guided Radiation Therapy
B. Cai,H. Li,D. Yang,V. Rodriguez,A. Curcuru,Y. Wang,J. Wen,R. Kashani,S. Mutic,O. Green
DOI: https://doi.org/10.1002/mp.12571
IF: 4.506
2017-01-01
Medical Physics
Abstract:Purpose: The purpose of this study was to investigate and characterize the performance of a Multi Leaf Collimator (MLC) designed for Cobalt‐60 based MR‐guided radiation therapy system in a 0.35 T magnetic field. Methods: The MLC design and unique assembly features in the ViewRay MRIdian system were first reviewed. The RF cage shielding of MLC motor and cables were evaluated using ACR phantoms with real‐time imaging and quantified by signal‐to‐noise ratio. The dosimetric characterizations, including the leaf transmission, leaf penumbra, tongue‐and‐groove effect, were investigated using radiosensitive films. The output factor of MLC‐defined fields was measured with ionization chambers for both symmetric fields from 2.1 × 2.1 cm2 to 27.3 × 27.3 cm2 and asymmetric fields from 10.5 × 10.5 cm2 to 10.5 × 2.0 cm2. Multi leaf collimator (MLC) positional accuracy was assessed by delivering either a picket fence (PF) style pattern on radiochromic films with wire‐jig phantom or double and triple‐rectangular patterns on ArcCheck‐MR (Sun Nuclear, Melbourne, FL, USA) with gamma analysis as the pass/fail indicator. Leaf speed tests were performed to assess the capability of full range leaf travel within manufacture's specifications. Multi leaf collimator plan delivery reproducibility was tested by repeatedly delivering both open fields and fields with irregular shaped segments over 1‐month period. Results: Comparable SNRs within 4% were observed for MLC moving and stationary plans on vendor‐reconstructed images, and the direct k‐space reconstructed images showed that the three SNRs are within 1%. The maximum leaf transmission for all three MLCs was less than 0.35% and the average leakage was 0.153 ± 0.006%, 0.151 ± 0.008%, and 0.159 ± 0.015% for head 1, 2, and 3, respectively. Both the leaf edge and leaf end penumbra showed comparable values within 0.05 cm, and the measured values are within 0.1 cm with TPS values. The leaf edge TG effect indicated 10% underdose and the leaf end TG showed a shifted dose distribution with 0.3 cm offset. The leaf positioning test showed a 0.2 cm accuracy in the PF style test, and a gamma passing rate above 96% was observed with a 3%/2 mm criteria when comparing the measured double/triple‐rectangular pattern fluence with TPS calculated fluence. The average leaf speed when executing the test plan fell in a range from 1.86 to 1.95 cm/s. The measured and TPS calculated output factors were within 2% for squared fields and within 3% for rectangular fields. The reproducibility test showed the deviation of output factors were well within 2% for square fields and the gamma passing rate within 1.5% for fields with irregular segments. The Monte Carlo predicted output factors were within 2% compared to TPS values. 15 out of the 16 IMRT plans have gamma passing rate more than 98% compared to the TPS fluence with an average passing rate of 99.1 ± 0.6%. Conclusion: The MRIdian MLC has a good RF noise shielding design, low radiation leakage, good positioning accuracy, comparable TG effect, and can be modeled by an independent Monte Carlo calculation platform.