ESTRO 38 Abstract book

S434 ESTRO 38

and 85.3%, respectively. In multivariate analysis, para- aortic lymph node dissection (HR, 2.537; 95% CI, 1.247 – 4.859; p=0.009), harvesting more than 30 pelvic lymph nodes (HR, 2.167; 95% CI, 1.050 – 4.472; p=0.037) and laparoscopic surgery (HR, 2.548; 95% CI, 1.127 – 5.760; p=0.025) were identified as independent risk factors for chronic LEL. After propensity score matching, EBRT group showed significantly higher chronic LEL rate than brachytherapy group (3-year LEL rate: 27.8% vs 10.8%, p=0.034). Conclusion Chronic LEL was relatively common in patients with gynecologic cancer, especially during the first year after surgery. Para-aortic lymph node dissection, laparoscopic approach and harvesting more than 30 pelvic lymph nodes during surgery were risk factors for chronic LEL. Compared with brachytherapy, adjuvant pelvic EBRT was also significantly associated with an increased risk of chronic LEL. PO-0829 MRI-based texture analysis of lymph node for predicting clinical outcome in cervical cancer patients S. PARK 1 , M.H. Hahm 2 , S. Na 3 , G.O. Chong 4 , S.Y. Jeong 5 , J.E. Lee 1 , M.K. Kang 1 , M.Y. Kim 6 , J. Kim 1 1 School of Medicine- Kyungpook National University, Department of Radiation Oncology, Daegu, Korea Republic of ; 2 Kyungpook National University Hospital, Department of Radiology, Daegu, Korea Republic of ; 3 Kyungpook National University Chilgok Hospital, Department of Biomedical Engineering Center, Daegu, Korea Republic of ; 4 School of Medicine- Kyungpook National University, Department of Obstetrics and Gynecology, Daegu, Korea Republic of ; 5 School of Medicine- Kyungpook National University, Department of Nuclear Medicine, Daegu, Korea Republic of ; 6 Kyungpook National University Hospital, Department of Radiation Oncology, Daegu, Korea Republic of Purpose or Objective The aim of this study was to evaluate the feasibility of MRI-based texture analysis of metastatic lymph nodes for predicting survival of node-positive cervical cancer. Material and Methods We enrolled 89 patients with node-positive cervical cancer with initial magnetic resonance imaging (MRI) for staging. All patients were treated with definitive chemoradiotherapy in our institution, from 2006 to 2016. The criterion for a metastatic lymph node was defined as a maximum short axis diameter of ≥8 mm on pretreatment MRI. T2-weigted image (T2WI) and contrast-enhanced image (CE) of metastatic lymph nodes were segmented semi-automatically. Voxel based texture analysis of mathematical parameters in addition to first order statistics for semi-automatically segmented metastatic lymph nodes were done. The local progression-free survival (LRFS), regional progression-free survival (RPFS), distant metastasis-free survival (DMFS), disease-free survival (DFS), and overall survival (OS) were estimated using the Kaplan-Meier method, and the differences were compared using log-rank tests. P values of < 0.05 were considered statistically significant. Results First order statistics including median (p = 0.023), entropy (p = 0.021) and uniformity (p = 0.015) and parameters of gray level co-occurrence matrix including homogeneity (p = 0.023) and normalized inverse difference (p = 0.023) were significantly associated with RPFS. Entropy (p = 0.012) and uniformity (p = 0.030) were found to be associated with DFS. Median value from first order statistical method was associated with OS (p = 0.023). Conclusion MRI-based texture analysis of metastatic lymph node could provide quantitative and objective assessment for predicting clinical outcome in node-positive cervical

Conclusion When comparing FIGO based on gyn exam only and T- stages based on MRI only in the frame of a primarily unblinded FIGO assessment, about one third of pats. could be classified differently. Upstaging is more frequent than downstaging. To minimize these uncertainties, a comprehensive knowledge of the differential value of clinical and MRI findings for the different areas of tumour spread is necessary to integrate these findings into one comprehensive final local T-stage at diagnosis “T clinical, imaging ” as suggested by the recent ESGO ESTRO ESP recommendations. PO-0828 Chronic Lower Extremity Lymphedema in Gynecologic Cancer Patients : EBRT versus Brachytherapy W.I. Chang 1 , H. Kang 1 , H.J. Kim 1 , K.S. Seo 2 , H.S. Kim 3 1 Seoul National University College of Medicine, Radiation Oncology, Seoul, Korea Republic of ; 2 Seoul National University College of Medicine, Rehabilitation Medicine, Seoul, Korea Republic of ; 3 Seoul National University College of Medicine, Obstretics and Gynecology, Seoul, Korea Republic of Purpose or Objective The goal of this study is to compare the risks of chronic lower extremity lymphedema (LEL) between pelvic external beam radiation therapy (EBRT) and brachytherapy and further identify risk factors for chronic LEL in gynecologic cancer patients who underwent adjuvant radiotherapy after radical hysterectomy. Material and Methods FIGO stage I-III gynecologic cancer patients who underwent adjuvant RT between January 2013 and December 2017 after radical hysterectomy were reviewed retrospectively. After ruling out other causes of lower extremity edema, patients suspected for LEL were diagnosed as LEL when circumference difference between both legs was 2 cm or more. Patients diagnosed as LEL before RT were regarded as acute lymphedema. One-to- one case-matched analysis was conducted with propensity scores generated from patient, tumor and treatment characteristics. Results Of 252 patients who met the inclusion criteria, 216 (85.7%) patients were treated with pelvic EBRT with or without brachytherapy and 36 (14.3%) patients were treated with brachytherapy alone. Overall, 34 (13.5%) patients were diagnosed with LEL with a median follow-up of 30.6 months (range, 12.0 – 67.7 months). Median interval from surgery to the onset of LEL was 5.0 months (range, 1.1 - 31.7 months). The proportion of patients who were diagnosed with LEL in 6 months and one year was 55.9%