Evaluating Latency In Dual-Gating MR-Guided Radiotherapy Using a Double-Motion Phantom on a 0.35T MR-Linac
Abstract
Purpose
A novel phantom configuration is designed to measure the multi-target tracking latency of a 0.35 T MT-guided Radiotherapy System.
Methods
An MRI motion phantom, initially designed for single target tracking, has been modified to have two independently moving tracking targets. The beam-on and beam-off latency of various multi-target tracking algorithms (5 total) of a 0.35 MR-guided RT system were measured. All algorithms were tested with the combination of either the system’s motion correction (MC) toggled on or off target while tracking either the target or organ-at-risk (OAR) objects. The latency of each of these combinations are averaged over at least 10 beam-on and beam-off events for a total of 40 latency results.
Results
Across all tracking algorithms, the average dual beam-off latency with MC-off was 230.6 ±102.0 ms (range: 173.4 ~ 279.7 ms). The average dual beam-on latency for all tracking algorithms with MC-off was 336.8 ±133.5 ms (285.7 ~ 438.0 ms). With MC on, the average dual beam-off latency for all tracking algorithms was -135.8 ±101.9 ms (-178.2 ~ -99.8 ms). Negative time values were measured for all MC on, beam-off configurations since, while using motion correction, the beam-off signal is sent earlier than when the expected beam-off signal is received. The average dual beam-on latency with MC on across all tracking algorithms was 447.3 ±143.9 ms (328.6 – 546.2 ms). Only one configuration (MC-on, beam-on, and Standard algorithm) breaches our 500 ms limit (adopted from AAPM 76) due to the nature of the motion control extending the beam-on time longer than expected.
Conclusion
Our novel phantom design was successful in measuring the average dual beam-off and on latency while tracking 2 targets simultaneously. All latency measurements with MC off were with the AAPM 76 limit of 500 ms.