Session Item

Breathing motion may result in large PTVs and poor CBCT image quality for patients treated with liver SBRT. To overcome this, we developed and implemented a markerless breath-hold liver SBRT technique using non-invasive nasal high flow therapy (NHFT) for breath-hold prolonging and surface-guided radiotherapy (SGRT) for breath-hold monitoring. In this study, we evaluate the first two years of clinical experience.

Heated and humidified air was administered via a nasal cannula (40 L/min, 80% oxygen, 34°). Patients performed voluntary inspiration breath-holds supported by visual feedback within a 2 mm gating window. After a training session, 4 or 5 contrast-enhanced breath-hold CT scans were acquired to enable delineation of an ITV to account for inter- and intra-breath-hold variations. Patients were treated in 3–8 fractions using VMAT with 6 or 10 MV FFF beams using SGRT-controlled beam-hold. Patient setup was performed using SGRT and 4-DoF CBCT-CT matching based on the liver. After treatment, a CBCT scan was acquired for evaluation purposes.

Between June 2020–June 2022, 17 liver SBRT patients started with the training session: 10 completed treatment in breath-hold, 5 continued in free-breathing (reasons: inability to understand procedure or position the arm above the head for a longer time and unreproducible or unstable breath-holds), 2 were not treated due to an increase in metastases. The average maximum breath-hold duration during treatment ranged per patient from 47–108 s. CBCT imaging took 60 seconds and half of all 100 scans were acquired during a single breath-hold. All patients could hold their breath for at least 60 seconds during one or more treatment fractions. For the 49 treatment fractions, the average number of breath-holds required for treatment was 6.6±4.1 (range: 2–19). The average time from CBCT before treatment to CBCT after treatment was 19±11 minutes (range: 9–51 min).
The difference in centroid position of the GTVs on the multiple planning CT scans was on average 3±3 mm for all directions (range: 0–10 mm). However, this difference was caused by a combination of breath-hold and delineation variations (volume differences ranged from 0.5–14.4 cm³). For all patients with a GTV difference >4 cm³, tumor visibility decreased over time due to the washout of contrast, resulting in smaller volumes on subsequent CT scans. The ITV was on average 6.3 cm³/28% larger (range: 1.1–23.9 cm³/5–72%) than the largest GTV. The 3D displacement vector for the 43 CBCT scans acquired after treatment was on average 5.0 mm (range: 0.7–12.9 mm; Figure 1) and for all but 2 scans less than our 1 cm PTV margin.

Liver SBRT in breath-hold using SGRT and prolonged breath-holds by NHFT is feasible for the majority of patients. Breath-holding resulted in high-quality CBCT imaging (Figure 2) that showed breath-hold variations smaller than the PTV margin. However, to further decrease the PTV, internal anatomy-based inter- and intra-breath-hold monitoring is desired.