Abstract book - ESTRO meets Asia

S93 ESTRO meets Asia 2018

matologiche, Roma, Italy 4 Hospice distretto socio sanitario 1, ASL Lecce, Lecce, Italy Purpose or Objective Upper gastro-intestinal (GI) acute toxicity can be a limiting factor in radiation therapy for abdominal neoplasms when treatment interruptions occur. Constraints for treatment planning optimization in abdominal radiotherapy mostly derive from scientific publications of pelvic tumors because optimization parameters defined directly from upper GI treatments are lacking or poor. The aim of this study is to evaluate dosimetric and clinical parameters regard to toxicity outcome in upper GI cancer radiotherapy. Material and Methods We retrospectively analyzed patients with upper GI cancer treated since 2009 to 2018 with 3D-conformal or intensity modulated radiotherapy, with concurrent chemotherapy or not, and after abdominal surgery or not operated. During the treatment, acute upper GI toxicities such as nausea, vomit and weight loss were reported and have been coded according CTCAE v4.03 scale. In all patients bowel bag (BB), liver, stomach (if present), were delineated by radiation oncologists on simulation CT. Dose Volume Histograms (DVH) of the delivered treatment were obtained and analyzed for detecting the VDose related to organs at risk (OARs) and the impact of other clinical factors that better predict the toxicity rising during the treatment. Results The records of 199 patients have been analyzed, median age was 66 years (range 35-84), 115 (57.8%) resected and 84 (42.2%) not resected. Primary tumors were located in the following subsites of upper GI tract: gallbladder 8 (4%), junction 9 (4.5%), pancreas 102 (51.3%), stomach 48 (24.1%), biliary tract 32 (16.1%). Median prescription dose was 50.4Gy (range 30-55.8) with median fractionation 1.8Gy (range 1.8-3.0). Linear quadratic correction for mildly hypofractionated treatments was not applied because the observed outcome was the acute toxicity, less influenced by fraction size and more by total dose. The DVHs of GI OARs were analyzed against the outcome (CTCAE toxicity grade) by fitting univariate logistic regression on Vdoses calculated at step of 0.1 Gy from 5 Gy to 50 Gy. The model showing in each case the lowest Akaike Information Criterion was selected for choosing the best VDoses. Upper GI CTCAE toxicity Grade >= 1 and Grade >= 2 best VDoses were V24Gy on BB, while Grade = 3 VDose on BB was V30Gy. Other OARs Vdoses didn't return significant results. Results of multivariate analysis are summarized in table 1. VDoses are the best predictors of toxicity for all toxicity grades.

significant deviation in Hosmer Lemeshow test in caliìbration. The number of radiotherapy treatment interruptions according toxicity grade is significantly higher for higher levels of toxicity (Fig. 1: χ2 P-Value = .00781).

Conclusion Our analysis shows a strong correlation between toxicity grade, interruptions and VDoses on BB. External validation of this model could offer further reliability for model clinical application in daily planning procedures.

Clinical: Lower GI

PO-227 MRI Down staging, Pathological Response & MSI in Signet Ring Cell Carcinoma Rectum Undergoing LCCRT B. Rajkrishna 1 , R. Thomas Samuel 1 , D. Saikat 1 , M. Dipti 2 , P. Tharani 3 , R. Rajat 4 1 Christian Medical College, Radiotherapy, Vellore, India 2 Christian Medical College, Genreal Pathology, Vellore, India 3 Christian Medical College, Radiodiagnosis, Vellore, India 4 Christian Medical College, Colorectal Surgery, Vellore, India Purpose or Objective To evaluate MRI down staging, pathological response and correlation of Microsatellite instability status with Radiotherapy response in Signet Ring Cell Carcinoma Rectum. Material and Methods 22 patients with signet ring cell carcinoma rectum who underwent preoperative LCCRT from 2013 to 2016 were enrolled in the study. All patients had MRI pelvis, serum CEA, CT thorax and abdomen or chest X ray and USG abdomen and pelvis and MSI immunohistochemistry (IHC). After 6 weeks of radiotherapy, response assessment was done using MRI pelvis and operable patients underwent Total Mesorectal Excision followed by adjuvant chemotherapy. Radiotherapy response was correlated with post radiotherapy MRI down staging, pathological response and MSI status. Results There were thirteen male and nine female patients. The mean age was 38 years, (18 - 62 years). Majority had T3 lesion (16) and N2 nodal status (14). All received LCCRT 50.4Gy with concurrent Capecitabine, except one due to co-morbidity. In post radiotherapy response assessment MRI, five showed MRI TRG 5 (no response), twelve TRG4 (minimal response), three TRG 3 (moderate response) and one TRG 2 (near complete response). No patients showed complete MRI response. Due to clinical progression one patient did not had post radiotherapy response assessment MRI. Among 22 patients, 15 were operable. Post operative histopathology showed six patients had pathological complete response (TRG 1), four had near complete

The best performing model is the one for CTCAE >= G2, showing an AUC of ROC = .76 (95% C.I. = .68 - .83), not