ESTRO 2024 - Abstract Book

S4382

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

ESTRO 2024

was used for gated treatment delivery for all patients and fractions. During treatment delivery, planar cine-MRI with visualization of the GTV contour and Boundary structure was available, with auto-beam off gating when the GTV was displaced outside the Boundary above a pre-specified threshold. Intra-fraction corrections were applied by couch shifts to bring the target (GTV) back into position (center of Boundary structure). Patients treated from 2016 to 2023 were included in this Ethics-approved retrospective study. The patient population comprised targets in the prostate (43%), lung (12%), pancreas (11%) kidney (10%), adrenal gland metastases (8%) and other groups (16%) of patients. Data of intra-fraction corrections was extracted from the MR-linac database and further analyzed to evaluate the number of and the applied 3D corrections. Preliminary dosimetric analysis was performed to ascertain the dosimetric implications of intra-fraction corrections.

Results:

Details of a total of 1689 patients and 8256 delivered fractions were extracted from the MR-linac database. In total, 3433 intra-fraction interventions were required to correct for anatomical changes, classified as 1) fast online couch correction based on planar cine-MRI during treatment delivery (73%); 2) repeated 3D MR imaging with couch correction (24%); 3) treatment compensation by means of an additional (partial) fraction due to intra-fractional target and OAR changes and a lack of intra-fraction replanning (3%). The number of patients requiring at least one intra fraction correction during the course of treatment was 1179 (ca. 70% of all patients) and the number of fractions in which intra-fraction corrections were applied was 1694 (ca. 20% of all fractions).

The number of corrections leading to a couch shift vector correction smaller than the applied PTV margin was 1987 (58%), whereas the corrections exceeding the PTV margin were 1446 (42%). Applied intra-fraction corrections reduced a potential underdose of target volume and/or an increased dose to OARs. However, a preliminary offline dosimetric analysis with deformable image registration (DIR) revealed the need to perform actual cumulative dose estimations and intra-fraction plan reoptimization due to the limitation of a 3D translation to account for intra-fractional changes.