ESTRO 38 Abstract book

S788 ESTRO 38

values for CEA and CA19.9 were 3.5 ng/mL and 8.6 U/mL. Tumors were located at inferior in 47.5% and at medial in 47.5%. Clinical staged cT3N+ in 72.5% and 87.8% was cN+. Acute toxicity was present in 81.6%, G3-4 in 10.0%. The majority was hematologic (44.7%; 1.9% G3-4), cutaneous (44.7%; 4.1% G3-4), diarrhea (24.7%; 3.4% G3-4). Following surgery, positive margins (R+) were obtained in 9.7%, more often in tumors of the inferior (p<0.001); linfovascular invasion (LVI) was found in 9.7% of pathological specimens. Pathological assessment found ypT0N0 in 15.9%, ypT1-2N0 in 28.4% and ypT3-4 or ypN+ in 55.6%.Lower pre-treatment CEA was related to better response (median of 2.0ng/mL to ypT0N0, 3.05ng/mL to ypT1-2N0, 4.6ng/mL to ypT3-4/N+;p<0.001). Better response was linked to lower rate of R+ and LVI (for both ypT0N0: 0%, ypT1-2N0: 2.2%, ypT3-4/N+: 16.3%; p<0.001). With a median follow-up of 52 months, 5 years LR disease- free survival (LRDFS), disease-free survival (DFS), overall survival (OS) and cancer-specific survival (CSS) were 91,8%, 68.5%, 70.8% and 80.4%, respectively. Pathological response (data in 5 years, ypT0N0 vs ypT1-2N0 vs ypT3- 4/N+) related to better LRDFS (100% vs 100% vs 84.9%; p<0.001), DFS (88.9% vs 85.4% vs 54.1%; p<0.001), OS (92.5% vs 82.5% vs 60.3%; p<0.001) and CSS (97.3% vs 88.7% vs 69.7%; p<0.001). Prognostic factors identified at univariate analysis were tumor location on DFS (p=0.012) and CSS (p=0.014), CEA at diagnosis on DFS (p=0.002), OS (p=0.002) and CSS (p=0.001), pre-treatment CA19.9 on LRDFS (p=0.036), DFS (p=0.027), OS (p=0.017) and CSS (p=0.001), LVI on LRDFS (p=0.004), DFS (p<0.001), OS (p=0.004) and CSS (p<0.001), R+ on LRDFS (p<0.001), DFS (p<0.001), OS (p<0.001) and CSS (p<0.001).In multivariate analysis R+ kept prognostic impact on LRDFS (p=0.009); pathological response (p=0.002), LVI (p=0,023) and R+ (p<0,001) on DFS; pathological response (p=0.023 and p=0.010, respectively) and R+ (p<0,001 for both) on OS and CSS. Conclusion RTCT LC in LARC is well-tolerated with manageable toxicities. Better response to treatment was related to better surgical and survival outcomes. Multivariate analysis identified R+, LVI and pathological response as prognostic factors. EP-1453 Machine learning prediction of early distant progression after SBRT for colorectal cancer H. Chung 1 , P. Lang 1 , M. Kayvanrad 2 , R. Thompson 3 , W. Chu 1 , E. Gennatas 4 , G. Valdes 4 , P. Cheung 1 1 Odette Cancer Centre- University of Toronto, Radiation Oncology, Toronto, Canada ; 2 University of Toronto, Rotman Research Institute, Toronto, Canada ; 3 Saint John Regional Hospital, Radiation Oncology, Saint John, Canada ; 4 University of California San Francisco, Radiation Oncology, San Francisco, USA Purpose or Objective While SBRT for oligo-metastases (OM) and oligo- progression (OP) has been increasingly utilized, it remains a challenge in identifying patients who would benefit from SBRT, due to complex interactions of patient, tumour and treatment factors. This study examines the ability of machine learning (ML) based classifiers to identify colorectal (CRC) patients who develop early distant progression (DP, ≤ 90 days since treatment completion) after completing SBRT for OM and OP, and thus received little benefit. Material and Methods All CRC patients treated with SBRT to any extracranial site at a single institution for OM/OP in 2009 - 2016 were retrospectively reviewed. Clinical characteristics included age, gender, pre-SBRT CEA, RAS status, ECOG performance, treatment indication (OM/OP), primary in situ, SBRT location, disease-free interval (DFI) since last treatment, number of prior lines of systemic of therapy, PTV volume and mean PTV BED. Univariable and

Simon Hospital, Digestive Surgery, Paris, France ; 7 Saint- Louis Hospital, Digestive surgery, Paris, France Purpose or Objective Squamous cell cancer of the anus is associated with multiple risk factors, including infection with human papillomavirus [HPV] and human immunodeficiency virus, immunosuppression, multiple sex partners, receptive anal sex and tobacco smoking. The aim of our study was to identify prognostic factors associated with poor outcome after radiotherapy for anal cancer. Material and Methods We analysed retrospectively the medical records of 171 patients treated by (chemo)-radiotherapy for a non- metastatic anal cancer in our institution from 2000 to 2015. Patients and tumor characteristics, treatments (chemotherapy (CT), radiotherapy (RT), and surgery) and outcomes were analyzed. Colostomy-free, disease-free and overall survivals at 5 years were studied. Univariate and multivariate analyses were performed by logistic regression in order to determine factors associated with poor progression free survival (PFS). Results Patients characteristics were as follow: median age: 62 years (range 36-89); gender: 45 males (26%) and 126 females (74%); HIV serology: positive 21 patients (12%), negative or unknown 150 patients (88%); tobacco smoking: 86 Pts (50%) among whom 28 Pts and 58 Pts were current and former smokers respectively. Tumors were classified as locally limited (T0-1-2, N0, M0) for 86 pts (50%) and locally advanced (T3-4 or N+, M0) for 85 pts (50%). Median total dose was 64.4 Gy (range 54-76.6), 146 patients were treated by concurrent chemoradiotherapy. Factors associated with poor PFS in univariate analysis were: tumor size >4cm, lymph node involvement; tobacco smoking, no initial surgical excision and anal warts at diagnosis. In multivariate analysis, only tobacco smoking status was significantly associated with poor PFS (HR=2.85 [1.25-6.50], p=0.013). 5-years-PFS for non-smokers, former smokers and current smokers was 88.1%, 76,7% and 73.8% respectively (p=0.038). Tobacco smoking was also associated with poor overall survival (p=0.03) and colostomy free survival (p=0.02). Conclusion Tobacco smoking status is associated with poor overall survival, PFS and colostomy free survival in patients treated for anal cancer by (chemo)-radiotherapy. EP-1452 Response after neoadjuvant radiochemotherapy as prognostic factor in locally advanced rectum cancer S. Couto Gonçalves 1 , A. Ponte 1 , M. Marques 1 , J. Casalta- Lopes 1 , I. Nobre-Góis 1 , T. Teixeira 1 , A. Barros 2 , M. Borrego 1 1 Centro Hospitalar e Universitário de Coimbra, Radiation Oncology, Coimbra, Portugal ; 2 Centro Hospitalar e Universitário de Coimbra, Oncology, Coimbra, Portugal Purpose or Objective Locally advanced rectal cancer (LARC) treatment includes neoadjuvant (NA) radiotherapy (RT) to achieve downsizing and downstaging and subsequently increase loco-regional (LR) control. The well-known long-course (LC) associates pelvic RT, 45-50.4Gy/25-28F/5-5.5weeks (1,8Gy/F), and radiosensitizing chemotherapy (RTCT). This study aims to evaluate pathological response and survival of patients with LARC after LC. Material and Methods Patients with LARC cT3-4 treated with NA RTCT LC prior to surgery between 2002-2017 were included. Tumor response was evaluated by AJCC staging system; toxicity by CTCAE5.0; survival by Kaplan-Meier; α=0.05. Results 320 patients were included, median age of 65 years, 64.4% male, karnofsky >80% in 94.4%. At diagnosis the median