S2266
Physics - Intra-fraction motion management and real-time adaptive radiotherapy
ESTRO 2026
Purpose/Objective: Intrafractional motion can distort radiotherapy doses [1]. Real-time image-guidance can trigger patient repositioning based on positioning errors [2] but does not consider motion-induced dosimetric errors. Real-time motion-including dose calculations [3] could enable real-time dose guidance but the optimal action levels and potential treatment improvements are unknown. This study simulated two dose-guided correction strategies and compared their accuracy and efficiency with geometry-guidance. Material/Methods: Nineteen prostate cancer patients with implanted fiducial markers were treated using two-arc VMAT delivering 40Gy in five fractions. Planning objectives were prostate CTV D95%=100% and bladder V36Gy<10cm ³ . During treatment, paired MV–kV images acquired every 20° of gantry rotation were used to monitor prostate positioning errors. The bladder was assumed to follow the prostate motion. Patients were repositioned when errors exceeded 1.5mm in any direction on two consecutive images (Figure1a). Post-treatment, the sub-doses delivered between MV–kV imaging were calculated in the TPS by splitting VMAT fields into sub-arcs. Motion was emulated by shifting and accumulating these sub- doses according to the prostate motion.In simulated real-time dose-guided treatments, the final prostate CTV D95% and bladder V36Gy were continuously estimated by assuming either constant prostate position as in the latest MV-kV image during the remaining treatment or repositioning to the planned position. Two dose-guided strategies were evaluated: a prostate-only strategy with repositioning triggered when the predicted prostate CTV D95% gain exceeded 1.0% (Figure1b), and a combined prostate–bladder strategy, where repositioning was additionally triggered when the final predicted bladder V36Gy was either >10cm ³ or >3cm ³ above the planned value and would be reduced by 0.5cm ³ by the repositioning.For comparison, two geometry-guided scenarios were evaluated with repositioning triggered at prostate positioning errors ≥ 1.5mm (current practice) and ≥ 3mm. The final motion-induced CTV D95% reduction compared to the plan ( Δ D95%) and number of corrections were compared (Figure1c) with current clinical practice (paired Wilcoxon rank test).
Results: The mean motion-induced CTV Δ D95% was higher for geometry-guidance with a 3mm threshold compared to 1.5mm, though not statistically significant. Both dose-guided strategies significantly reduced motion- induced CTV Δ D95%, and the combined prostate– bladder strategy eliminated all fractions violating bladder V36Gy >10cm ³ . Across all dose-guided approaches, the mean number of treatment interruptions per fraction was notably lower than in current clinical practice (Figure 2).
Conclusion: Online dose guidance can improve target coverage and reduce high bladder doses while substantially reducing treatment interruptions, highlighting its potential for more efficient and accurate treatment delivery. References: [1] Sengupta C, Skouboe S, Ravkilde T, Poulsen PR, Nguyen DT, Greer PB, et al. The dosimetric error due
Made with FlippingBook - Share PDF online