Session Item

In July 2021, we started with model-based selection (MBS) of oesophageal cancer (EC) patients for proton therapy (PRT), in neo-adjuvant (nCRT) as well as in definitive setting (dCRT). Selection was based on prediction models for 2-years mortality, which included the mean heart dose (MHD) and gross tumour volume of primary tumour (GTVp) as predictors. The aim of this study was to evaluate the selection of EC patients for PRT and to quantify the dose and 2-years mortality risk reductions that could be obtained

MBS requires a plan comparison (PhRT vs PRT) which was performed in all EC patients. To select patients for PRT, the following eligibility criteria had to be met: cT1-T3, cN0-N2, WHO performance score <2, target motion <15 mm (established on 4DCT) and a ΔRisk >5% for 2-years mortality. We used the following prediction models for 2-years mortality;                                         

for dCRT: -1.0421 + 0.059 * SQRT (GTV) + 0.263 * SQRT (MHD),

and for nCRT: -3.0352 + 0.100 * SQRT (GTV) + 0.4457 * SQRT (MHD).

IMPT robustly optimized plans were created using a 3 beam approach including 5 times repainting to reduce the impact of intra-fractional motion during treatment. To verify patient positioning, daily 2DKV imaging, online CBCT’s and weekly repeat CT’s were performed.

Since July 2021, 30 patients were treated with nCRT (n=16) or dCRT (n=14) for EC at our department. In total 21 patients (70%) qualified for PRT, including 88% of patients treated with nCRT and 50% of those treated with dCRT. Reasons for not qualifying were mostly N3-status in the dCRT group and breathing movement and limited clinical benefit in the nCRT group (Table 1). In patients who qualified, the mean heart (MHD), lung (MLD), liver (MLiD) and spleen dose (MSD) were significantly lower for PRT compared to PhRT. Although, the ΔMHD (12.4 vs 8.3 Gy) and ΔMLD (5.3 vs 3.8 Gy) between the PhRT and PRT plans were significantly larger in dCRT group compared to the nCRT group, the difference in the 2-year mortality risk was on average 12% in both groups, in favour of PRT. This might be explained by the difference in the steepness of and the position on the curves of the prediction models for dCRT or nCRT (Figure 1).

PRT reduces the dose in multiple organs-at risk, including the MHD. Using the applied model, this is expected to result in reduced mortality risk, both after nCRT and dCRT. Criteria, other than ΔRisk, seem to impact the selection of EC patients for PRT more often in the dCRT group.