ESTRO 2024 - Abstract Book

S4226

Physics - Intra-fraction motion management and real-time adaptive radiotherapy

ESTRO 2024

The implementation of online adaptive radiotherapy (oART) presents an opportunity to potentially reduce PRV margins, as it eliminates interfractional motion management. Nevertheless, the question of whether reducing PRV margins in the treatment of pancreatic cancer with cone beam CT (CBCT)-based oART is a valid approach remains unclear.

This study aims to evaluate the feasibility of reducing PRV margins by taking intrafractional motion into account, as assessed through CBCT scans obtained during practical oART sessions.

Material/Methods:

In this study, we assessed the CBCT data from four patients with pancreatic cancer who underwent oART treatment between December 2022 and May 2023. CBCT scans were obtained at two different points in our oART workflow: once for daily planning and again for position verification immediately prior to irradiation. For each CBCT scan, we retrospectively delineated the structures of the stomach and duodenum, designating one contour set as “base” during the planning oART phase and another as “pre” at the position verification stage. Additionally, we obtained the contour set that was clinically employed during the treatment and labeled it as “on couch”. This analysis involved a total of 12 CBCT sets. 1. To validate the reliability of the clinically utilized contour and to gauge intrafractional motion, we computed the Dice Similarity Coefficient (DSC) and Mean Distance to Agreement (MDA) for the stomach and duodenum structures in the “base”, “on couch”, and “pre” contours. 2. For the evaluation of PRV margins, we determined the Simpson Similarity Coefficient (SSC) for the stomach with margin sizes of 5, 7, and 10 mm at the “on couch” and “pre” time points. Additionally, we calculated the SSC for the duodenum with margin sizes of 3 and 5 mm at the “on couch” and “pre” stages. 3. To assess the change of intrafractional dose distribution: we made the reference plan using Ethos therapy solution emulator with initial planning CT and contour set of “on couch”. PRV margins of 5 mm was added to the duodenum and stomach. The prescription dose was 48 Gy in 15 fractions with a dose covering 95% of the target (D95%) to the planning target volume (PTV)-PRV, a volume that subtracted PRVs from the PTV, and PTV D98% ≥ 36 Gy. Thereafter, the adapted plan was generated under the same clinical goals as the reference plan with contours of “on couch”. Finally, evaluation of dose volume index of “pre” linked to initial planning CT was performed; whether dose constraint of V45 Gy < 1 cm3, V42 Gy < 5 cm3, and V39 Gy < 25 cm3 of stomach and duodenum were to be met. Both CT and CBCT scans were conducted during the end exhalation breath-hold phase. The contours were assessed at a 3 cm vertical and 1.5 cm horizontal offset from the PTV’s outer boundary. For CBCT sets acquired on the same day, consistent location information was maintained, and the registration process was conducted without any positional adjustments. We assessed the PRV margins as follows:

Results:

The median time elapsed between planning- and pre-CBCT was 18.3 minutes (ranging from 14.4 to 23.8 minutes). In terms of stomach volume, the median values for “base”, “on couch”, and “pre” were 113.1, 103.0, and 102.2 cm³, respectively, while for the duodenum, the corresponding values were 37.5, 40.3, and 41.6 cm³. Detailed DSC and MDA are summarized in Table 1. The SSC values are presented in Table 2, and it's worth highlighting that the stomach with margins of 5 mm, smaller than the margins outlined in our protocol, exhibited notably high SSC scores. Regarding