Session Item

In respiratory gated radiotherapy, a low latency between target entrance/exit of the gating window and actual beam on/off is crucial for high treatment accuracy. However, no standard method to determine the latency exists. We used an in-house developed method (Thomsen et al, ESTRO 2021) for accurate latency measurements at a clinical proton and photon accelerator and simulated the geometrical consequences for liver SBRT treatments. 

Gating latencies were measured at a Varian ProBeam (protons, RPM gating system) and TrueBeam (photons, TrueBeam gating system) accelerator, respectively. A motion-stage performed 1cm, 1Hz vertical sinusoidal motion of a marker block that was optically tracked by the gating system. Amplitude gating levels were set to cover approx. half the motion. Gated beams were delivered to a 5mm cubic scintillating ZnSe:O crystal that emitted visible light when irradiated. During beam-delivery, a GoPro video camera acquired images at 120Hz of the moving marker block and light emitting crystal. After treatment, the marker block position (black dot, Fig.1A) and crystal light intensity were determined in all video frames (Fig.1A). The video was then synchronized with the gating system log files by least squares-based alignment of the marker block motion in the video and in the gating log files (Fig. 1B-C). The gating latencies were measured over 15 breathing cycles and defined as the time difference between marker entrance/exit of the gating window and actual beam on/off as detected by the crystal signal (Fig.1B-C). For TrueBeam, the dose rate (and thus crystal signal) increased/decreased over an extended period of ≈30ms at each beam on/off and the latencies were measured relative to the mid-time of this 30ms interval (Fig.1B).

Simulations of geometrical consequences of the gating latencies were based on the cranio-caudal motion during the first treatment field of 15 patients previously treated with liver SBRT guided by internal electromagnetic motion monitoring (Calypso). Amplitude-based respiratory gating with approximately 50% duty cycle around the exhale phase was simulated for all patients. Geometrical errors caused by gating latencies were quantified as the percentage of the total time beam-on time spent outside the gating window (Fig.2A).

The mean (±SD) gate-on latencies were 270.8±37.2ms (ProBeam) and 86.2±12.8ms (TrueBeam), respectively. Gate-off latencies were 103.6±16.2ms (ProBeam) and 44.5±11.6ms (TrueBeam). For simulated treatments based on these latencies, the mean time with beam-on outside the gating windows was 6.7% [patient range:3.7-10.0] for ProBeam and 3.5% [1.8-5.5] for TrueBeam (Fig.2B). For gate-off latencies ≤ 104ms, the beam-on time with errors above 1mm were below 5% for all patients (Fig.2C).   

Gating latencies were measured based on beam visualization by a scintillating crystal. The TrueBeam had smaller latencies than the ProBeam. For both accelerators, the latencies only resulted in minor geometrical errors.