S2249
Physics - Intra-fraction motion management and real-time adaptive radiotherapy
ESTRO 2026
Brussels, Belgium. 2 Institut de Recherche Expérimentale et Clinique (IREC), Pôle de Pneumologie, ORL (airways) & Dermatologie (skin), Groupe Recherche en Kinésithérapie Respiratoire, Université Catholique de Louvain, Brussels, Belgium. 3 Service de kinésithérapie et ergothérapie, Service de Pneumologie, Cliniques Universitaires Saint-Luc, Brussels, Belgium. 4 Service de radiothérapie, Cliniques Universitaires Saint-Luc, Brussels, Belgium Purpose/Objective: Mechanically assisted non-invasive ventilation (MANIV) is a breathing management technique that enables stable and reproducible deep-inspiration breath-holds (DIBHs) using a ventilator delivering sustained inhalation pressure and enhanced oxygenation without sedation [1–3]. It is used in clinical practice during stereotactic (SABR) and fractionated radiotherapy (RT) for thoracic and abdominal cancers, ensuring consistent DIBHs within and across treatment fractions. This study evaluates intra-fraction motion in patients treated with MANIV-DIBH and investigates whether its high reproducibility permits reduction of intra-fraction PTV margins. Material/Methods: This retrospective analysis included all patients treated since May 2022 with MANIV-DIBH for pulmonary or upper abdominal tumours using SABR (3-5 fractions) or fractionated RT (analysis of first 5 fractions). For each fraction, three cone-beam CTs (CBCTs) were acquired: pre-treatment (CBCT1), during treatment (CBCT2), and post-treatment (CBCT3). Positional shifts along the anteroposterior, craniocaudal, and mediolateral axes were recorded (AP/CC/ML), using CBCT1 and CBCT2 as references for CBCT2 and CBCT3, respectively. Treatment with fewer than three recorded deviations per fraction were excluded. Mean deviations, three-dimensional displacement vectors (V3D), systematic error ( Σ : standard deviation SD of population means), and random error ( σ : root mean square of intra-patient SD) were computed. PTV margins were derived using the simplified van Herk formula: PTV = 2.5 Σ + 0.7 σ [4].
Results: Of 84 treatments using MANIV-DIBH, 72 were eligible for analysis (69 SABR, 6 fractionated RT), targeting 12 pulmonary and 60 upper-abdominal tumours. A total of 622 CBCTs (369 CBCT2, 253 CBCT3) yielded 1107 intra-fractional and 757 post-fraction deviations. From CBCT2, mean population deviations were -0.01 cm (AP), 0.06 cm (CC), and -0.01 cm (ML), with mean V3D of 0.34 ±0.19 cm. Systematic and random errors were 0.13 cm, 0.14 cm, and 0.10 cm and 0.23 cm, 0.27 cm, and 0.16 cm, respectively (AP/CC/ML). From CBCT3, the mean population deviations were 0.03 cm, 0.00 cm, and 0.02 cm (AP/CC/ML), with a mean V3D of 0.33 ±0.16 cm. Systematic and random errors were 0.10 cm, 0.11 cm, and 0.07 cm (AP/CC/ML) and 0.22 cm, 0.23 cm, and 0.18 cm (AP/CC/ML), respectively. The PTV margins were 0.48 cm, 0.53 cm, and 0.36 cm (AP/CC/ML) for CBCT2, and 0.41 cm, 0.45 cm, and 0.30 cm (AP/CC/ML) for CBCT3.
Made with FlippingBook - Share PDF online