Session Item

Friday
May 07
14:15 - 15:30
21st century brachytherapy: is it available, affordable and relevant?
0210
Symposium
00:00 - 00:00
Performance in clinical conditions of the dual-layer multileaf collimator from a Halcyon 2.0 linac
PO-1381

Abstract

Performance in clinical conditions of the dual-layer multileaf collimator from a Halcyon 2.0 linac
Authors: Hermida Lopez|, Marcelino(1)*[mhermida@vhebron.net];Sánchez-Artuñedo|, David(1);Cereijo-Graña|, Francisco(1);
(1)Hospital Universitario Vall d'Hebron, Servei de Fisica i Protecció Radiològica, Barcelona, Spain;
Show Affiliations
Purpose or Objective

The Halcyon platform (Varian Medical Systems) introduced a multileaf collimator (MLC) with a novel dual-layer design. In the Halcyon 1.0, only the lower (distal) layer was used for beam modulation. In the recently introduced Halcyon 2.0, both distal and proximal layers modulate the beam, thus allowing more complex fluences to be delivered.

This works aims to characterize the performance in clinical conditions of the MLC of one Halcyon 2.0 unit, by analyzing the treatment logs of delivered clinical plans, and the results produced in the daily machine performance checks (MPC).

Material and Methods

Trajectory logs of 5043 treatment fractions were analyzed, corresponding to 271 IMRT and VMAT plans. Treatments were delivered from Mar. 6, 2019 to Oct. 2, 2019. Logs were analyzed with home-made Python scripts, based on the package pylinac [https://pylinac.readthedocs.io/en/stable/]. MPC analysis was based on the results of 160 MPC checks. The analysis was done with home-made R scripts.

MLC RMS and 95th percentile values were obtained from the logs and were grouped by treatment site and modality (VMAT or IMRT). For the MPC checks, we report accuracy and repeatability grouped by MLC layer and bank. Time of unscheduled machine maintenance due to MLC malfunctions was recorded.

Results

The maximum RMS error (0.164 mm) and the 95th percentile RMS error (0.110 mm) for all analyzed logs (both specified at isocenter distance) are far below the tolerance of 3.5 mm proposed by the TG-142 report. Boxplots and point clouds in figure 1 show the distribution of maximum RMS error values for each treatment session, grouped by treatment site and modality. The number of analyzed logs by site ranged from 65 for bladder to 1391 for prostate. Although a one-way ANOVA test found statistically significant differences of mean RMS error among sites, these differences probably do not have practical relevance, as the mean values of RMS maximum errors for all sites ranged from 0.055 mm to 0.076 mm.


Boxplots in figure 2 show the accuracy (positioning error at isocenter distance) of the MLC grouped by layer and bank, as obtained from the analyzed MPC checks. Repeatability (not shown) was better than 0.5 mm and 0.6 mm for the distal and proximal layers, respectively. The obtained values are far below the tolerances suggested by Varian (accuracy: 0.6 mm for the distal layer, and 0.55 for the proximal layer; repeatability: 0.8 mm for distal layer, and 0.9 mm for proximal layer). No temporal trends were observed in the MLC MPC results along the studied period.



No MLC-related machine breakdowns occurred during the 7-month analyzed period.


Conclusion

The MLC of the Halcyon 2.0 performed far below the tolerances recommended by the TG-142 report and by the manufacturer, along a period of 7 months, with no MLC-related machine breakdowns. These results suggest that more stringent tolerances can be applied for the quality assurance of the dual-layer MLC from the Halcyon 2.0.