ESTRO 38 Abstract book

S88 ESTRO 38

of Radiation Oncology, Aarhus, Denmark ; 6 Leiden University Medical Center, Department of Radiation Oncology, Leiden, The Netherlands Purpose or Objective EMBRACE-II is a prospective interventional multicentre study of IMRT and MRI-based image-guided adaptive brachytherapy (IGABT) in cervix cancer. The EMBRACE-II RTQA process for IGABT is mandatory for centres which did not participate in the first EMBRACE study. The RTQA process comprised 1) a compliance questionnaire confirming that IGABT was established at the centre, 2) contouring on two benchmark cases by the principal investigator, and 3) submission of dosimetric data from 5 dummy run patients treated according to the study protocol. This abstract reports the results of the contouring assessments for the benchmark cases. Material and Methods Delineation was performed using a bespoke online contouring tool. Case 1 was a IIIB tumour with 40mm anterior vaginal extension at diagnosis. There was moderate response to chemoradiotherapy (CRT) with a small amount of residual anterior vaginal disease and palpable proximal parametrial disease at BT. Case 2 was also a IIIB tumour but with no vaginal involvement at diagnosis. There was good response to CRT with no residual palpable or radiological parametrial disease. The regions of interest (ROIs) assessed were the residual GTV (GTV res ), high-risk CTV (CTV-HR), intermediate-risk CTV (CTV-IR), bladder, sigmoid, rectum and bowel. Each ROI was scored 0-10 by 2 assessors - a score of ≥6 was required to pass. Individualised feedback was also provided. A pass on every ROI for both cases was required for accreditation. Results 49 clinicians submitted contours for evaluation. 4 (8%) passed at the first attempt and 27 (62%) on resubmission after individualised feedback. 23% of submissions required revision of one ROI, 29% two ROIs, and 30% three or more ROIs. The most common ROIs requiring revision were the GTV res (65%), CTV-IR (61%), sigmoid (49%) and bowel (63%) [Figure 1].

Conclusion This is the first study to correlate the results of summative (pass/fail) assessments for benchmark cases in cervix cancer IGABT with a systematic analysis of qualitative feedback. The first-time pass rate was low and most submissions required revision of more than one ROI. Qualitative analysis showed that while many errors were due to conceptual difficulties, there were also errors related to variation in case selection. Individualised feedback improved contouring, but the time taken to manually assess and give relevant feedback was a significant burden. Improved tools which will allow rapid contouring across a large number of cases with automated assessments are required. OC-0177 The value of kV-CBCT in adaptive HDR brachytherapy of cervical cancer patients C. Constantinescu 1 , N. Jastaniyah 2 , S. Wadi-Ramahi 1 1 King Faisal Specialist Hospital and Research Center, Biomedical Physics, Riyadh, Saudi Arabia; 2 King Faisal Specialist Hospital and Research Center, Radiation Oncology, Riyadh, Saudi Arabia Purpose or Objective To assess the intra-fraction variation of dose delivered to organs-at-risk (OARs) using a kV-CBCT scanner and estimate the potential of this technology for adaptive HDR brachytherapy (BT) of cervical cancer. Material and Methods Data of cervical cancer patients who underwent VMAT pelvic irradiation (45Gy/25fractions) followed by HDR BT (28Gy/4fractions) were retrospectively reviewed. Bladder-filling and rectum/bowel preparation protocols were performed. A CT scan with 2 mm slice thickness was acquired and images imported to TPS for planning purposes. Contouring of the HR-CTV and the OARs (bladder, rectum and sigmoid) was performed on CT images, guided by MR images previously acquired. Planning objectives were according to (GYN) GEC-ESTRO recommendations Immediately before treatment, a kV-CBCT scan was acquired with the patient in treatment position, using an in-built pelvis scanning protocol. CBCT images were manually registered with planning CT using the applicators as landmarks and the OARs were re- and respective EQD2 were recalculated and compared to the planned ones using a Wilcoxon test. The dose variations were evaluated by the relative percentage difference between the doses calculated on planning CT and verification kV-CBCT images. Possible correlation between OARs dose and volume variation was investigated by a linear regression test. Results Data of 19 patients and 57 BT fractions were analyzed. In each BT fraction, variations in OARs (as volume, shape contoured. Their D 2cc

A pair-wise comparison of individual performance for each ROI in the 2 cases was performed [Table 1]. The proportion of clinicians failing both cases ranged between 2-24% per ROI; a common conceptual difficulty was often identifiable from the qualitative feedback. For the GTV res , CTV-IR and sigmoid, performance was worse in one case compared to the other which could be attributed to differences in case difficulty. For the CTV- HR, the causes for clinicians passing only one case were not obvious.