ESTRO 2024 - Abstract Book

S2270

Clinical - Upper GI

ESTRO 2024

2712

Digital Poster

Can we deliver what we plan? The impact of dose accumulation on model-based predictions in EC

Sabine Visser, Erik W. Korevaar, Margriet Dieters, Johannes A. Langendijk, Stefan Both, Christina T. Muijs

University Medical Centre Groningen, Radiation Oncology, Groningen, Netherlands

Purpose/Objective:

Multiple critical organs-at-risk (OAR) are exposed to dose when treating esophageal cancer (EC) with radiotherapy. Prediction models are being used to translate these OAR doses to complication risks, and to select patients that benefit most of proton therapy. For EC, patients are selected according to the model-based approach using the 2 years mortality probability (2YM) model based on gross target volume size and mean heart dose (MHD) at baseline. 1 However, the delivered dose to the patient might deviate from the planned dose, affecting the estimated risk. In this study, we aimed to quantify differences between the planned and delivered dose for the most relevant OAR and evaluate the consequences for the 2YM predictions.

Material/Methods:

The dataset included 161 EC patients treated with IMPT (n=143) or VMAT (n=18) (Table 1). MHD, mean lung dose (MLD) and effective dose to the circulating immune cells (EDIC) 2 were collected at baseline (planned dose) and were reconstructed using calculations of the treatment plan on at least three repeated CTs, resembling the treatment course dose. For this purpose, two strategies were used: (1) The dose was warped by deformable image registration to the planning CT and accumulated (accumulated). (2) Weekly dose data was collected and the mean was calculated (weekly mean). Subsequently, the 2YM estimates were calculated based on the corresponding MHDs. Differences between the planned, accumulated and weekly mean evaluation were compared using the Friedman test and post-hoc Wilcoxon signed rank tests.

Results:

The median dose in the planned, accumulated and weekly mean evaluation was 9.0 Gy, 9.0 Gy and 9.3 Gy for MHD, 3.5 Gy, 3.4 Gy and 3.4 Gy for MLD, and 1.1 Gy, 1.1 Gy and 1.2 Gy for EDIC (Figure 1). A statistical significant difference (p<0.05) was found comparing the planned to the reconstructed dose for MHD, with corresponding differences for 2YM and EDIC. Differences >2 Gy were seen in 21/161 (13%) patients. The reconstructed MHD was higher in 15 of these patients (resulting in an increased 2YM of 3.1% on average) and lower in 6 patients (resulting in a decreased 2YM of 2.7% on average). The difference between the accumulated and weekly mean evaluation was smaller; only one patient showed a difference >2 Gy, which translated in a 2YM difference >2%. For this patient, the MHD on the repeated CT was consistently lower than of the warped dose on the baseline CT, as was the volume of the heart (Figure 2).

Conclusion: