Abstract Book

S87

ESTRO 37

Purpose or Objective To describe tumor recurrence patterns in patients with resectable gastric cancer treated with preoperative chemotherapy, surgery, and postoperative chemotherapy or chemoradiotherapy in the international multicenter randomized CRITICS trial. Material and Methods Event-free survival and tumor recurrence patterns were determined for 788 patients with adenocarcinoma of the stomach who were randomized to preoperative chemotherapy, surgery and postoperative chemotherapy (CSC group), or preoperative chemotherapy, surgery and postoperative chemoradiotherapy (CSCRT group). Event- free survival was defined as time from randomization until disease progression before surgery, irresectable disease at surgery, tumor recurrence after potentially curative resection or death from any cause. Sites of progressive or recurrent disease were categorized as loco-regional, peritoneal, distant or occurring at multiple sites. Event-free survival was compared between the two arms using the log-rank test. Time to first site-specific progression or recurrence accounted for competing risks and was summarized as cumulative incidence functions. Results Of the 788 patients (393 in CSC and 395 in CSCRT) included between 2007 and 2015, 636 (81%) patients (310 in CSC and 326 in CSCRT) underwent surgery with curative intent, and 478 (61%) patients (233 in CSC and 245 in CSCRT) received postoperative treatment. At the time of the analysis median follow-up was 5.1 years and 474 patients experienced an event (233 patients in CSC and 241 patients in CSCRT). Event-free survival at 2 and 5 years were 52% vs. 51% and 39% vs. 38%, for CSC vs. CSCRT, respectively (stratified log-rank p = 0.92). Cumulative incidences at 2 and 5 years from the competing risks analysis are shown in Table 1. Loco- regional recurrence was detected within 5 years in 19% of patients (9% loco-regional only + 10% in combination with another site) in the CSC arm vs. 19% (7% + 12%) of patients in the CSCRT arm. In the subset of 245 patients who underwent surgery with curative intent and started postoperative chemoradiotherapy, the 2-year and 5-year cumulative rates for loco-regional recurrence were 12% (4% loco-regional only + 8% in combination with another site) and 15% (6% + 9%), respectively.

Radiotherapy, Utrecht, The Netherlands 2 University Medical Center Utrecht, Image Sciences Institute, Utrecht, The Netherlands Purpose or Objective As part of the development of MR-guided HDR brachytherapy, we have been working on HDR source localization. HDR source localization can be used for 2 main goals: I) real-time MR-guided treatment verification, and II) detection of the dwell positions after catheter insertion, by localization of a dummy source, to replace manual catheter reconstruction. To be clinically relevant, high temporal and spatial resolutions are required. This study describes a validation (accuracy and precision) of our proposed MR-based HDR source localization method, by a comparison with CT. Material and Methods Data acquisition A non-active Ir Flexisource (Elekta) was positioned in the center of a platform in a cylindrical phantom filled with doped water. MR imaging was performed on a 1.5T MRI system (Ingenia, Philips), using a 2D dynamic spoiled gradient echo sequence (10 dynamics), see scan parameters in Table 1. Two intersecting slices (coronal/sagittal) were scanned (source cable in-plane). For an increased temporal resolution, the spatial resolution was varied and all acquisitions were repeated with parallel imaging (SENSE=2). The influence of these acceleration steps on the localization accuracy was assessed. Besides, in subsequent MR acquisitions, the phantom was rotated to vary the angle between the source and B0 (0º-90º). A 3D CT scan was made with voxel size: 0.34x0.34x0.40mm³, using O-MAR (Philips). The CT source position was considered as the gold standard. Post processing The HDR source position was determined by simulation of the MR artifact (complex data) and matching the MR images to the simulations in a phase correlation algorithm to find the translation between the two images[1], for the coronal and sagittal slices (see Fig.1). The MR images were registered (translation) to the CT data set. Next, the registered 2D positions were combined to the 3D MRI source position. Accuracy and precision The accuracy was calculated as the mean Euclidean distance between the source positions from MRI and CT (over 10 dynamics). The precision was analyzed as the standard deviation over the distances between the positions from MRI and CT (over 10 dynamics).

Conclusion In the CRITICS gastric cancer trial most tumor recurrences occurred within the first two years after treatment. Cumulative incidences of recurrent disease at different sites were comparable between the two study arms, i.e. postoperative chemotherapy versus postoperative chemoradiotherapy. On behalf of the CRITICS Investigators

Proffered Papers: BT 2: Brachytherapy physics and technology

OC-0169 Accuracy and precision of high frame rate MR-based HDR brachytherapy source localization E. Beld 1 , M.A. Moerland 1 , M.A. Viergever 2 , J.J.W. Lagendijk 1 , P.R. Seevinck 2 1 University Medical Center Utrecht, Department of