ESTRO 2024 - Abstract Book

S4045

Physics - Inter-fraction motion management and offline adaptive radiotherapy

ESTRO 2024

objectives and constraints. The prescribed dose of the targets was 45 Gy in 25 fractions. For optimization and evaluation, a robustness setting of 5 mm/3% was used. The constraint to the targets was V42.75Gy>95% on the voxelwise-minimum dose. Twenty treatments of 25 fractions were simulated per patient to evaluate CTV-LR coverage. For each fraction, the dose was recalculated on a randomly chosen reCT with additional simulated treatment uncertainties, differentiating between systematic and random uncertainties. Adequate coverage was defined as at least 90% of the treatments achieved a target coverage of V42.75Gy>95%. This was evaluated on the CTV-LR by accumulating the CTV-LR dose of the 25 fractions. After accumulation, the isodose line of the 42.75 Gy was expanded with different margin combinations. All margin combinations were sequentially tested, from 0 to 8 mm in left/right direction and from 0 to 15 mm for anterior/posterior direction in steps of 1 mm and from 0 to 15 mm for cranial/caudal direction in steps of 3 mm. We determined the Pareto-optimal margin settings for the CTV-LR in terms of the target coverage of the expanded dose and the increased target volume. Furthermore, we compared isotropic margin settings with anisotropic.

Results:

The study yielded 25 treatment plans, all of which met the EMBRACE-II planning aims. With these plans, 260 treatments were simulated on the reCTs. For these treatments, 746496 margin combinations were evaluated. Figure 1 depicts the Pareto-optimal margin combinations. A margin combination is considered Pareto-optimal when the best coverage is reached with a minimal increase in the internal target volume. Notably, when no margin is added, only 70% of the treatments achieved an adequate target coverage. To assure an adequate target coverage for 90% of the treatments, a 2 mm margin isotropically is sufficient, which increases the target volume on average with 104 cc. However, employing an anisotropic margin combination of {0, 0, 0, 3, 3, 0} mm in left, right, anterior, posterior, cranial, and caudal (LRAPCC), respectively, allowed for adequate target coverage with a mean internal target volume increase of 47 cc. Figure 2 shows the margin combinations that resulted in a Pareto-optimal coverage, emphasizing the critical role of the posterior and cranial margin in achieving adequate target coverage.

Conclusion:

By conducting a detailed analysis of simulations of fractionated treatment courses and anisotropic margin combinations, this study shows that with the addition of a small margin, the probability obtaining a CTV-LR V42.75Gy>95% is drastically increased. Consequently, adding a anisotropic margin of {0, 0, 0, 3, 3, 0} mm in LRAPCC directions ensured adequate coverage for 90% of the treatments. This margin setting is Pareto optimal in terms of coverage versus increase in internal target volume. This study underlines the potential of employing a combination of margins and robustness settings as method to secure adequate target coverage without unnecessary increase in the target volume.