2406 - Expiratory - Phase CMM Gating for Liver SBRT on an MR-Linac: Unity Operational Performance

Purpose/Objective(s): Liver SBRT is highly sensitive to respiratory motion. MRI-guided treatment allows real-time visualization and gating of targets. Comprehensive Motion Management (CMM) on the Unity MR-Linac provides anatomy-based tracking with automatic beam gating and drift correction. This study evaluates the operational performance of Unity’s expiratory-phase CMM gating in liver SBRT with evaluation of geometric and dosimetric stability of adaptive plans.

Materials/Methods: We analyzed 10 liver SBRT patients (5 fractions each, total 50 fractions) treated with expiratory-phase CMM gating. Patients underwent a respiratory correlated 4D-CT simulation. The end expiratory phases (40%,50%, and 60%) were identified as the most stable treatment window. The Gross Tumor Volume was contoured on the 40%,50%,and 60% CT phase datasets. These volumes were merged to create the internal GTV (iGTV). A Planning Target Volume was generated by adding a 3 mm margin to the iGTV. Pretreatment MRI was acquired & co-registered to the planning 4D-CT. Online plan adaptation was used in 60% of fractions (Adapt-To-Shape) and 40% (Adapt-To-Position). Unity AuditLog.csv data were processed in MATLAB, using time-aligned logs to compute metrics. We calculated duty cycle, gating (VOICE) efficiency, LINAC active and beam-on times, and beam-hold durations. Target motion was quantified via 3D translation RMS; intrafraction drift was assessed by moving-median baselines; tracking robustness was summarized by APM tracking states.

Results: All fractions met institutional dosimetric targets. For the cohort, mean duty cycle was 93 ± 5% (VOICE efficiency 90 ± 6%, typical threshold ~96%). LINAC active time was 9.5 ± 1.5 min per fraction; beam-on time was 8.1 ± 1.0 min. Beam-holds were frequent but brief (median 0.23 s; occasional holds up to ~37 s). Residual motion during beam-on was small (Z-axis RMS 2.5 ± 0.5 mm, total 3D RMS 2.8 ± 0.6 mm). Measurable baseline drift occurred (mean longitudinal drift +3.1 ± 1.5 mm per fraction). All patients successfully completed the prescribed SBRT course. Tracking remained robust, with zero APM tracking losses during beam-on. (Predicted-translation log fields were present but unused, so prediction metrics were not evaluated.)

Conclusion: Expiratory-phase CMM gating on the Unity MR-Linac is feasible for liver SBRT, yielding a high duty cycle (~90%) and constrained residual motion during delivery. The observed intrafraction drift suggests a benefit for drift-correction strategies. Unity’s real-time gating provides robust target tracking in liver SBRT with acceptable efficiency.