ESTRO 2021 Abstract Book

S487

ESTRO 2021

Conclusion The interim analysis of this trial suggests that these two strategies of radiation dose escalation are feasible and lead to organ preservation in patients with operable rectal cancer. Early follow-up data suggests that TME-free survival is improved with the addition of brachytherapy compared to EBRT. The IDMC committee approved the continuation of patient recruitment in the phase III study as planned.

This trial was funded by Elekta and the Jewish General Hospital Foundation

OC-0622 Propensity- Score analysis of Proctectomy-TME vs Organ Preservation for Rectal cancer J. Gerard 1 , L. Montagne 2 , J. Doyen 2 , B. Thamphya 3 , S. Gourgou 4 , K. Benezery 2 , J. Hannoun-Levi 2 1 centre Antoine Lacassagne, Radiothérapie, Nice, France; 2 centre Antoine Lacassagne, Radiotherapie, Nice, France; 3 centre Antoine Lacassagne, Biostatistique, Nice, France; 4 institut De Cancerologie De Montpellier, Biostatistique, Montpellier, France Purpose or Objective Organ preservation (OP) using Contact X-Ray Brachytherapy (CXB) combined with chemoradiotherapy (CRT) is gaining interest for early T2-T3 adenocarcinoma of distal-middle rectum. Such OP approach may be detrimental for survival due to increased risk of local recurrences. To evaluate this risk we conducted a propensity -score matched cohort analysis comparing cohort using TME vs OP with CXB. Materials and Methods Oncological outcomes were analyzed comparing Accord 12 randomized trial (584 pts, 2004-2007)) treated using CRT and radical Proctectomy (RP-TME) and the single center (CAL- Nice) planned OP cohort (71pts, 2002-2018) using CXB and CRT. Both cohorts were using the same inclusion screening process, similar data reporting, same CRT and same primary investigator (JPG). The quality of data in both cohorts was strictly monitored. The variables included in in the propensity -score model were PS status, TN classification and rectal circumferential extension. Ages were not matched. Patients were matched 1:1 using the nearest neighbor method ( 0.1 caliper restriction). Propensity scores were calculated with logistic regression and multiple imputation for missing data. Cox proportional hazards model was estimating hazards ratios [95% CI]. The primary end point was 5-year specific survival. Results In Accord 12 cohorts after eliminating T4 and T > 66% circumference 311 pts were selected. In CAL cohort after eliminating T1 and non- standard radiotherapy 61 pts were selected. The one to one matching derived two groups of 36 pts each well matched for all confounding factors. Median follow-up time was 56 and 60 months respectively in CXB and Accord 12. The 5-year cancer specific rate was 89% [95%CI: 78-100] in Accord 12 and 82% [95% CI: 68-98] in CXB (no significant difference) (fig). The 5-year rate of local recurrence was 9% [95%CI: 0-19] and 10% [95%CI: 0-20] in Accord 12 vs CXB. The 5-year rate of metastases was 15% [95%CI: 2-26] and 22% [95%CI: 6-36] with no significant difference. A total of 33/36 patients preserved a normal rectum in CAL group.