Session Item

In the radiotherapy treatment of oesophageal cancer (EC), the lungs and heart are in close proximity of the target volume. Reducing lung dose was historically the primary aim in treatment planning. However, due to increased awareness of the risk of cardiac complications after radiotherapy, the optimisation process nowadays is more focused on reducing heart dose. The result is an increased lateral dose contribution which may lead to less robust treatment plans. In this study, we explored the trade-off between heart and lung dose, and corresponding target robustness.

Two EC patients were selected for this study, who presented diaphragm displacements during treatment (Table 1). Target volumes were delineated on the phases and the average of the planning 4DCT, and of two repeated 4DCTs. For each patient, two volumetric modulated arc therapy (VMAT) plans were created to cover the planning target volume; one with the aim to reduce mean lung dose (MLD) as much as possible (VMAT_lung) and one with the aim to reduce mean heart dose (MHD) as much as possible (VMAT_heart). Both plans had to obey priority 1 and 2 constraints as defined in our clinic (MHD <26 Gy; MLD <16 Gy). First, we investigated the presumed static dose cloud of the plans by shifting the isocenter 8mm in all directions on the planning image. The resulting 14 dose scenarios were summarized in a voxel-wise minimum dose distribution to evaluate target coverage. Additionally, the plan was robustly evaluated (including 2mm shifts to account for residual uncertainties) on the repeated 4DCTs, using the average and using accumulative dose of all the phases.

For both patients, the optimisation window of heart dose was substantially larger than that of lung dose comparing VMAT_lung and VMAT_heart (Figure 1). The average MHD reduction was 12.5 Gy comparing VMAT_heart to VMAT_lung, while on average the MLD increased by 3.3 Gy. Robust evaluation at baseline suggested reduced target robustness for the VMAT_heart plans, compared to the VMAT_lung plans (Table 1). Variations in the diaphragm position resulted in reduced target coverage on the repeated CT. This was especially true for the VMAT_heart plans if the diaphragm moved cranially as observed for patient 2. On the opposite, the diaphragm moved caudally for patient 1 and hotspots were observed in the heart region for VMAT_heart. For both patients, the changes in diaphragm position were consistent along the treatment. When all phases were considered, target coverage slightly improved, as the dose averaged out. However, target underdosage was still present for patient 2 (V95: 74-79%).  

In VMAT radiotherapy for EC, the heart can be spared without a substantial increase in MLD. However, this results in increased risk of underdosage if the diaphragm shifts in cranial direction or cardiac hotspots if it shifts in caudal direction. Robustness evaluation of the treatment plan seems to indicate these risks already at baseline.