ESTRO 2021 Abstract Book

S1060

ESTRO 2021

histological subtype (carcinosarcoma > papillary-serous carcinoma > clear-cell carcinoma) were found to be predictive of worse OS. Adjuvant EBRT and/or VBT were significant predictors of longer PFS (due mostly to reduced local recurrence). No benefit in OS or CSS was seen neither with RT or CT in our series. Conclusion The role of adjuvant RT and CT is not well established in all histological subtypes. While CT has shown improved survival in patients with papillary-serous carcinoma its effectiveness in less clear in other subtypes. Further studies are needed to assess best clinical management of non-endometrioid carcinoma patients. PO-1287 Target Deformation and Intrafraction Motion Errors in LRCTV for MR-guided Cervix Cancer Radiotherapy I.Z. Tan 1,2 , B. Tocco 1 , S. Court 3,4 , D.C.S. Tang 5 , D. McQuaid 6,7 , I. White 8,9 , S. Lalondrelle 10,9 1 Royal Marsden Hospital, Clinical Oncology, Sutton, United Kingdom; 2 Institute of Cancer Research, Radiotherapy and Imaging , Sutton, United Kingdom; 3 Royal Marsden Hospital, Radiotherapy Physics, Sutton, United Kingdom; 4 University College London Hospitals, Radiotherapy Physics, London, United Kingdom; 5 Imperial College London, School of Public Health, London, United Kingdom; 6 Royal Marsden Hospital, Joint Department of Physics, Sutton, United Kingdom; 7 Institute of Cancer Research, Joint Department of Physics, Sutton, United Kingdom; 8 Guys and St Thomas' Hospital, Gynaecology Clinical Oncology, London, United Kingdom; 9 Institute of Cancer Research, Radiotherapy and Imaging, Sutton, United Kingdom; 10 Royal Marsden Hospital, Gynaecology Clinical Oncology, Sutton, United Kingdom Purpose or Objective Typically, the cervix, whole uterus, parametrium and upper vagina are included as the low risk clinical target volume (LRCTV) for cervical cancer irradiation. Due to target motion, internal target volume and subsequent PTV volumes are large. Evidence suggests that a 5-10mm margin around the gross tumour volume (GTV) may be sufficient to encompass microscopic tumour extension towards the uterine fundus 1,2 . Direct visualisation of the GTV using magnetic resonance guided radiotherapy (MRgRT) may allow redefinition of CTVs, potentially removing the need to treat the whole uterus. We hypothesize that a reduced volume CTV (rCTV) would show less target deformation and intrafraction motion error 3 compared to a standard CTV (sCTV). Materials and Methods Repeat MRI scans were acquired during chemoradiotherapy for cervical cancer on weeks 0 (planning), 2, 3 and 4 of treatment. On each MRI, the target volumes and organs at risk were contoured on two sequences (T2-W 3D and DIXON TRA IN 3D) acquired over 4:40-36:17mins. sCTV is defined as the LRCTV as per EMBRACE2 protocol 4 . rCTV represents LRCTV but with a 1.5cm anterosuperior margin on GTV towards the lower uterus (Fig. 1). PTVs were isotropic 5 mm expansion of the corresponding CTVs. Rigid registration of MRI sequences within the same session between both sequences (mimicking intrafraction motion) and in between weeks (mimicking interfraction motion) was performed. The most extreme points of the GTV, sCTV, rCTV were measured axially in each sequence in the superior, inferior, anterior, posterior, right and left directions. The patient and population mean M, systematic Σ and random errors σ were calculated for both CTVs.

Results 18 paired MRI interfraction motion sequences in six patients were analysed for target deformation errors (Σ deform, σ deform) while 16 paired MRI intrafraction motion sequences were analysed for intrafraction motion errors (Σ treat, σ treat). Mean interfraction motion was greatest in the superior (sCTV 0.44cm, rCTV 0.27cm) and anterior directions (sCTV -0.23 cm, rCTV -0.30cm). sCTV had larger Σ deform and σ deform (0.83-1.29cm and 0.81-1.26cm respectively) compared to rCTV (0.65-1.00cm and 0.63-0.98cm respectively). However, Σ treat and σ treat were similar in both sCTV (0.18-0.55cm and 0.41-0.69 respectively) and rCTV (0.14-0.65cm and 0.35-0.70cm respectively). As the most extreme points were measured axially, only 1 extreme point was