ESTRO 2024 - Abstract Book

S4078

Physics - Inter-fraction motion management and offline adaptive radiotherapy

ESTRO 2024

2231

Digital Poster

Breast Cancer Proton Therapy: DSC as Metric for Treatment Precision & Decision-making

Bojan Štrbac, Vincent C. Hamming, Pietro Pisciotta, Jeffrey Free, John H. Maduro, Johannes A. Langendijk, Stefan Both

University Medical Center Groningen, Department of Radiation Oncology, Groningen, Netherlands

Purpose/Objective:

In this study, the Dice similarity coefficient (DSC) was employed as a crucial metric to assess the alignment between the planned clinical target volume (CTV) and the observed CTV during patient positioning in proton radiotherapy. The DSC fulfills two functions: it serves as an evaluation tool for treatment accuracy and assists in clinical decision-making. Evaluating treatment involves assessing how closely the delivered treatment aligns with the planned CTV. Clinical decision-making, on the other hand, is based on establishing a DSC threshold that signals significant deviations— whether anatomical or positional. Such deviations trigger the consideration for a repeat CT (rCT). The primary objective of this study was to explore the use of the DSC as a method to assess treatment accuracy and direct clinical decisions for breast cancer patients with non-reconstructed chest walls undergoing proton radiotherapy.

Material/Methods:

In this study, we analyzed 173 daily online CBCTs from ten patients. Hybrid deformable registration was used to map the CTV and heart from the planning CT (pCT) to the CBCTs. The method was automated using the RayStation scripting platform. In the context of intensity-modulated proton therapy (IMPT), a 5 mm robustness margin was incorporated to address delivery uncertainties. Additionally, a ±5 mm on-treatment limit was established, which was determined by monitoring swelling structures that account for both anatomical and positioning deviations. This limit served as a criterion for rCT acquisition. The DSC's value depends on the changes in the geometry's size, position, and the distance to its center of mass. For thin chest wall targets, a swelling variation of ±5 mm translates to a DSC of 0.75. However, for precise fraction delivery, a higher DSC threshold of 0.85 has been set as the benchmark for successful positioning.

Results:

Our analysis demonstrated a broader DSC range for variations in the CTV compared to the heart. Out of the 173 evaluated fractions, 93% met or surpassed the DSC threshold of 0.85 for the CTV(Figure 1). Notably, a deliberate outlier case where a patient underwent adaptation on the third fraction but retained the initial CT illustrates the high sensitivity of the method (Figures 1 and 2, Patient 1)