ESTRO 2024 - Abstract Book

S4370

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

ESTRO 2024

Purpose/Objective:

Breast cancer radiotherapy has evolved over the past decades, both in dose prescription and in radiation techniques. Fractionation schemes have been reduced to 5 fractions of 5.2 Gy for many patients, and wedge-based 3D conformal treatments using tangential beams have been replaced by modulated beams using photons or protons, sometimes combined with deep-inspiration breath-hold (DIBH) techniques for heart dose reduction. Patient positioning, anatomical variations but also variation in DIBH may potentially result in less uniformly delivered dose to the breast than initially anticipated, which may not average out for hypofractionated treatments. This study aims to quantify the dosimetric consequences of DIBH reproducibility in short-course breast radiotherapy for modulated beam delivery techniques.

Material/Methods:

Clinical hypofractionated non-DIBH plans (5 x 5.2Gy) of 26 breast cancer patients recently treated on Ethos machine (6MV-FFF) with an average CTV volume of 737 cc (range 141 – 1422 cc) were included. Treatment planning for all plans (11 IMRT, 15 VMAT) was performed in Eclipse 17.0 (Varian Medical Systems). For IMRT, a tangential beam setup is used with 2 medio-lateral beams and 2 lateral-medial beams (spaced by 15 degrees). If constraints to nearby organs at-risk were not met, VMAT plans with 3 partial arcs of 220-240 degrees were created. All plans were optimized using a virtual bolus, subsequently recalculated without virtual bolus. As part of the clinical workflow, plan robustness was checked by applying a 5 mm isocenter shift in medial and dorsal direction, plan recalculation and evaluation based on PTV V95>95% and V110%<1%. Intrafractional positioning errors as a result of repeated DIBH were simulated by perturbing the isocenter of individual beams in 3 directions by a random number drawn from a Gaussian distribution, representative for intrafraction DIBH variations 1 . Distributions with standard deviations (SD) of 1.0 mm (group A) and 1.5 mm (group B) were used to perturb isocenter positions, representative for typical and worse DIBH reproducibility, respectively. Dose distributions of perturbed beams were calculated using Monte Carlo dose engine in SciMoCa 1.7.2.5381 (Scientific RT) and beam doses were summed. Plans were dosimetrically evaluated on breast CTV retracted 5 mm from the skin, normalized to the mean reference dose, for mean dose, D95% and V95%. Per group and plan, 25 simulations were performed. For 6 patients, additional VMAT plans were optimized using 6 arcs to limit each beam dose delivery to a maximum of 30 seconds. Simulations of these plans were compared to simulations of the clinical plans in which each beam was simulated to be delivered in 2 DIBHs.

Results:

For group A, all beam displacements were within 3 mm (69% within 1 mm, 96% within 2 mm), whereas 5% of the beams was displaced in a direction larger than 3 mm for patients in group B (50% within 1 mm, 80% within 2 mm). In total, 1295 plan simulations were completed. Over all patients, VMAT plans showed smaller variations in CTV mean dose and CTV D95% compared to IMRT plans (see Figure). For example, CTV mean dose SD measured 0.13 Gy for IMRT and 0.11 Gy for VMAT for group A, and increased for group B to 0.20 Gy (IMRT) and 0.13 Gy (VMAT), respectively.