PD-0878 Reduction of Post-Operative Toxicity in Esophageal Cancer Patients after Model Based Proton Therapy

Purpose or Objective During the COVID-19 pandemic, health insurance providers permitted to treat esophageal cancer (EC) patients with proton therapy (PT) when treated in the neo-adjuvant setting. This decision was based on the results of the MD Anderson randomized study showing significantly lower rates of the Total Toxicity Burden (TTB), especially of postoperative toxicity and consequently decreased hospital and intensive care unit (ICU) stay. The TTB combines a large range of toxicities with different incidence rates and severities into a single sum score. For patient selection, we used a model-based approach (MBA) using a model for TTB (>60). The aim of this study was to investigate if the post-operative TTB decreased with PT. Materials and Methods In accordance with the MBA, a plan comparison (photon vs proton radiotherapy (RT)) was performed in all EC patients that were to be treated with neo-adjuvant chemoradiotherapy (nCRT) according to the CROSS regimen. For patient selection, we developed a normal tissue complication probability (NTCP)-model for a TTB > 60, which corresponds to at least one grade > III or 2 grade > II complications. Patients, who were eligible for PT in terms of target motion ( 5%, were selected and treated with PT. Postoperative TTB and hospital and ICU stay were compared with a prospective dataset of patients that were treated with photon RT between 2014 and 2018 (n=224). Results Since March 2020, 26 out of 32 patients were selected for PT. The average reduction in mean lung dose was 5.2 Gy with PT compared to VMAT, with a corresponding average NTCP-reduction for TTB of 9.7%. Mean heart, spleen and liver dose were reduced by 7.9, 7.4 and 7.1 Gy, respectively. At the time of analyses, 14 out of 19 patients underwent esophagectomy. Three patients developed intercurrent metastases, in one a wait and see policy was applied and one switched to definitive CRT. In the resected PT patients, the expected proportion of patients with a TTB>60 was 29.7% for photon and 20.8% for proton RT. After resection, only one out of 14 patients (7.1%) had an TTB>60 ((p=0.051), compared to the expected proportion based on the photon plans). On average, the observed post-operative TTB was 15 (SEM: 6) after PT, compared to the historical photon cohort in which the observed post-operative TTB was 45 (SEM: 5). Hospital and ICU stay were also reduced by PT compared to the photon cohort;12.7 (SEM 1.4) vs. 20.0 (SEM 1.2) and 1.9 days (SEM 0.4) vs. 4.9 (SEM 0.6), resulting in an average cost reduction of about 10,000 euro per patient. Follow up will continuously be updated in our dynamic database. Conclusion In the neo-adjuvant setting, the first results indicate that proton-based nCRT reduces the post-operative TTB compared to historic photon-based nCRT. Moreover, there seems to be a trend towards reduced hospital and ICU stay.