Development and demonstration of end-to-end testing for intra-fraction motion-managed workflows.

Background: Intra-fraction motion management techniques, including beam gating and intra-fraction drift correction (IDC), have recently been introduced on the Unity MR-linac (Elekta AB, Stockholm, Sweden) to mitigate the dosimetric impact of motion during treatment. However, residual motion (e.g., within the gating window) still affects the delivered dose, causing deviations from the statically planned dose. Conventional end-to-end (E2E) testing does not incorporate such (known) motion, hampering evaluation of motion managed workflows.

Purpose: This study develops and demonstrates novel methods that incorporate known motion before treatment delivery. Using such a reference dose distribution allows for E2E testing of intra-fraction motion-managed workflows.

Methods: A novel approach was developed to assess the E2E accuracy for motion-managed delivery techniques by comparing the measured dose distribution to a reference dose distribution that incorporates the applied motion during the delivery. Two motion-included reference dose distributions were generated and evaluated: (1) A Priori Motion-Included (APriMI) dose distribution which uses the known (periodic) motion to estimate the influence of anatomical motion on the dose distribution, and incorporates this into a new dose distribution; and (2) the Posteriori Motion-Included (PostMI) dose distribution, which adds an external trigger to relate the beam-on/off time to the motion of the setup. This allows for evaluation of non-periodic motion, or a drift motion during IDC workflows. In addition to these, the conventionally used static treatment planning system (TPS) dose distribution was used as a reference dose distribution. Several scenarios were evaluated: static (no phantom motion), two unmanaged, and two motion-managed scenarios using the Comprehensive Motion Management (CMM) software (Elekta AB, Stockholm, Sweden) for gated and IDC workflows, with and linear drift motion patterns. All measurements were performed on a clinical Unity MR-linac equipped with CMM software, using film dosimeters for high spatial resolution dose distribution assessment. The geometric and dosimetric E2E accuracy of the workflow were evaluated for all scenarios.

Results: First, the static benchmark scenario was evaluated and showed high agreement between the measured dose distribution and all reference dose distributions (i.e., static, APriMI, and PostMI). For the motion-included scenarios, excellent agreement was observed between the measured and calculated dose distributions in both unmanaged and managed cases when using either APriMI or PostMI. The largest geometric shift in the motion included scenarios was 0.3 mm, comparable to the static scenario. Dosimetric accuracy, evaluated using a global gamma index (2%/2 mm), exceeded 95.5%. As expected, larger deviations occurred when the static dose distribution was used as a reference, with geometric shifts up to 9.0 mm and gamma pass rates as low as 17.8%.

Conclusions: E2E testing of intra-fraction motion-managed workflows is possible using APriMI and PostMI dose distributions. Strong agreement was observed with these motion-included distributions, while larger deviations were seen with the static dose distribution. These findings highlight the need for reference dose files that account for actual motion in the measurement setup to assess E2E accuracy of motion-included workflows.