Session Item

Ultra-high dose rate radiotherapy (FLASH-RT) involves few fractions of high doses delivered in a fraction of a second and aims to widening the therapeutic window by reducing normal tissue toxicity. Although the fast delivery entails largely reduced intrafractional motion, it is crucial to have proper motion management to ensure precise delivery of the high dose. For motion management, surface guided radiotherapy (SGRT) provides real-time monitoring, sub-millimeter accuracy and a large field of view (FoV). Rendering of surface images (SI) requires that the projected light is reflected and detectable by the surface scanning (SS) system’s camera. In a novel approach, we aim to utilize SGRT for motion management during FLASH-RT. 

Four canine cancer patients were treated with surface guided FLASH-RT at a modified linear accelerator (Precise, ELEKTA, Sweden). Prescribed doses of 28-35 Gy were delivered in 3.5 µs electron pulses (200 Hz) during a total treatment time of 60-125 ms. The patients were anesthetized prior to treatment. For SGRT, a single camera (Catalyst^TM, C-RAD, Sweden) unit was used. To account for the wide range of fur colors, the appropriate integration time and gain was individually set for each patient to optimize the rendered SI. Also, the FoV was adjusted due to the off-isocentric treatment. The reference SI was acquired in conjunction with the patient positioning. Recording of the motion during the whole treatment session, including the FLASH delivery was executed for one patient. The motion was analyzed by retrieving data from the SS system log files.

The individual camera exposure settings for all canine cancer patients and SI for two of them are presented (Table 1, Figure 1b, 1d). For the patient where surface guided motion management during beam-delivery was used, motion up to 3 cm between patient setup and treatment start was recorded (Figure 1e) and treatment was postponed until full anesthesia was reached. The total vector offset from the reference SI was 0.1 mm at the time of radiation delivery indicated by the SS system (Figure 1e). The same offset was also observed 1.72 s before beam-on and 1.63 s afterwards. For patients with very dark fur, it was challenging to obtain an acceptable SI due to hardware limitations. When encountering the maximum exposure time, the signal could not be further amplified which resulted in either a decreased speed of the SS system or decreased reconstructed surface regions (Dalmatian vs Labrador, Figure 1b and 1d).

We have demonstrated that motion management during FLASH-RT is important and feasible using SGRT. It was shown that the surface scanning system can cover the required FoV and capture images during the off-isocentric FLASH-RT treatment. It can be used to estimate the motion prior to treatment and to assure that the patient remains in the correct position during beam delivery.