MR-Linac provide soft-tissue contrast, enabling direct visualization of liver lesions during treatment with the potential to increase SBRT precision. Imaging, however, is hampered by breathing motion. We therefore developed a 4D-MRI guided liver SBRT workflow, which is in clinical use since October 2019. The purpose of this study was to evaluate this motion management strategy and toxicity in the first cohort of patients. Eight patients with 10 liver oligometastases received 4D-MRI guided liver SBRT, after signing informed consent. For each patient, a mid-position (midP) planning CT was derived from a 4D-CT. An SBRT plan was created on the midP-CT (step-and-shoot IMRT, 15 beams, 3x20 Gy), with a patient-specific anisotropic PTV margin tailored to the respiratory tumor amplitude in the 4D-CT. On the MR-Linac, a daily 4D-MRI (multi 2D TSE, TE/TR = 60/316 ms, 25 slices, 2x2x5mm3, 30 repetitions, 4min) was acquired. In each 4D-MRI phase, the tumor was loco-rigidly registered to the midP-CT. The time-weighted average of the 10 resulting registrations was applied to adapt the treatment plan, using adapt-to-position with segment-weights optimization. Before irradiation, a new 4D-MRI was used to visually verify if the tumor was still accurately aligned. Additional adaptions were applied if needed. During irradiation, motion monitoring was performed by repetitive acquisition and visual inspection of 4D-MRI scans. The time of the patient in the treatment room was recorded. The tumor in all 4D-MRI scans was retrospectively registered to the midP-CT to quantify intra-fraction mid-position and peak-to-peak amplitude variability. Additionally, acute toxicity was reported. With our 4D-MRI guided workflow, all lesions were visible prior and during treatment. No additional fiducial markers were required for guidance. Treatment times were available for 12 fractions with a median (IQR) of 45 (41 – 50) minutes. Median daily tumor misalignment before plan adaption was 2.6 (-6.5 – 10.6) mm (LR), 4.2 (-6.2 – 17.7) mm (CC) and -0.3 (-4.5 – 4.1) mm (AP). 4D-MRI tumor amplitude was 4.0 (2.5 – 5.9) mm (LR), 14.9 (10.8 – 23.7) mm (CC) and 6.6 (2.2 – 13.9) mm (AP). The intra-fraction tumor mid-position group mean (M), systematic (Σ) and random (σ) deviations were M = 0.2, Σ = 1.7, σ = 1.9 mm (LR), M = 3.6, Σ = 2.9, σ = 2.9 (CC) and M = -2.0, Σ = 1.5 σ = 2.5 mm (AP). Tumor amplitude variability with respect to the midP-CT was M = 1.4, Σ = 1.7, σ = 1.7 mm (LR), M = -5.0, Σ = 6.1, σ = 3.6 (CC) and M = -0.4, Σ = 4.0 σ = 3.5 mm (AP). Grade 2 toxicity (fatigue and nausea) was observed in four patients, with a median (IQR) follow up time of 36 (15 – 51) days. 4D-MRI guided liver SBRT was successfully delivered to our first cohort of 8 patients, providing image guidance without marker implantation. Median treatment time was 45 minutes. Considerable intra-fraction position and amplitude variability stress the need for intra-fraction motion management strategies, such as trailing. No grade ≥3 toxicity was observed.
Read full abstract