Commissioning and robustness evaluation of external laser sensor for respiratory gating in carbon-ion radiotherapy

Abstract

Background: Respiratory-induced tumor motion remains a significant challenge in carbon-ion radiotherapy (CIRT), potentially compromising dose accuracy and escalating dose in nearby organs. Although belt-based sensors are commonly used for gating, issues related to setup reproducibility and patient comfort increased the demand for laser-based, non-contact alternatives. Purpose: This study evaluates the performance of an ANZAI laser-based respiratory gating system for potential use in this modality based on total delay measurements and assessment of the accuracy and signal robustness of the sensor under a broad range of simulated and clinical conditions. Methods: The total system delay was quantified through segmentation of the signal pathway into four steps: motion-to-waveform, waveform-to-gate, gate-to-beam, and signal-to-irradiation. A programmable motion phantom was used to assess the tracking accuracy under various conditions, including changes in amplitude, period, and irregular breathing patterns, such as baseline shifts and phase variations. Signal robustness was evaluated with varying the sensor-to-surface distance (9-17 cm), sensor angle (0-45 degrees), ambient lighting, and interference from adjacent lasers. A volunteer study correlated the external laser signal with internal diaphragm motion obtained via four-dimensional computed tomography (4DCT) in supine and prone positions and assessed cross-compatibility with a belt-type sensor. Results: The total system delay was 59.3 ms for beam-on and 40.4 ms for beam-off, which is well below the 100-ms threshold. In the phantom tests, the sensor achieved precise tracking, with cross-correlation coefficients > 0.99 (regular and irregular respiratory patterns). The signal was robust across a 9-17 cm span and was stable in a dark environment. However, the measured amplitude increased at sensor angles > 20 degrees, and considerable signal distortion occurred due to interference from an adjacent laser. A strong correlation between the sensor signal and internal organ motion was observed in volunteer tests in supine (r = 0.986) and prone (r = 0.961) positions. The interoperability with the belt sensor was excellent (r = 0.99). Conclusions: The ANZAI laser sensor system is accurate and reliable, and it meets clinical latency requirements in respiratory-gated CIRT. It provides stable tracking under varied respiratory conditions. For optimal performance, maintaining a perpendicular sensor angle is recommended to ensure amplitude accuracy and mitigate crosstalk from adjacent lasers. These findings confirm the clinical readiness of the system and inform practical implementation.