Reducing treatment machine downtime with a preventative MLC maintenance procedure

The persistent use of MLCs through VMAT and IMRT is causing additional wear and tear on these mechanical parts, leading to an increase in MLC interlocks, breakdowns and failures. This study investigates the effect of an MLC clean and service procedure on MLC performance and positional accuracy demonstrated through in-house service logbook reports, Varian MLC backlash test results, daily MLC position QC results and Varian TrueBeam trajectory log file data. A service and clean of each individual MLC leaf was carried out on 3 standard MLC and 2 high definition MLC (HDMLC) Varian TrueBeam linacs. In-house service logbook reports, Varian backlash test results, daily MLC picket fence QC results and beam hold data from patient delivery trajectory log files were analysed for up to 6 months pre and 2 years post the MLC service/clean to assess the impact on unplanned MLC maintenance work and MLC performance. The median (range) of hardware faults reduced from 12 (1-17) pre clean to 0 (0-1) and 4 (1-11) at 12 and 24 months respectively for all linacs with the exception of a HDMLC linac where faults increased from 4 pre-clean to 13 and 20 again at 12 and 24 months respectively. The reduction in faults in the alternative 4 linacs was consistent with the reduction in the number of MLCs reaching the 0.3 mm and the 0.4 mm backlash recommendations in the first 12 and 24 months following the service/clean. The increase in faults in the HDMLC linac was also consistent with the increase in MLCs reaching the 0.3 mm backlash recommendation in the first 12 and 24 months. The median (range) of MLCs reaching a daily picket fence QC position tolerance of 0.25 mm reduced from 14 (1-40) pre clean to 0 (0) at both 12 and 24 months post-clean. This demonstrates the improvement in MLC performance caused by the MLC service/clean but also reveals factors other than MLC position accuracy influence hardware faults. Additionally, the number of beam hold-offs determined from patient delivery trajectory log files were found to have no correlation with the MLC service/clean. The MLC service/clean improves MLC performance and MLC position accuracy, reducing reactive repair work for engineering and physics staff. The results were maintained for 1 year post the MLC service/clean, with a trend back towards pre-clean levels in the subsequent 12-24 months. This suggests this preventative maintenance work could be performed at a frequency of > 2 years. This period of reduced faults and improved performance is significant given the 10 year expected lifespan of a linac.