ESTRO 2020 Abstract book

S1053 ESTRO 2020

and clinical performance of the marker ,i.e., visibility on CT‐scan, CBCT and electronic portal images (EPI) and safety were evaluated. Results Seventy‐four markers were available for analysis. Marker pair distance showed horizontal or negative slope of fits during the treatment course in 96% of the markers. One out of seventy‐four markers showed clinically relevant migration. One marker was lost before start of treatment, and one marker was lost during treatment. Visibility of the marker was good on CT and CBCT and moderate on EPI (figure 2). The injection procedure was feasible and well tolerated. No severe adverse events related to marker placement were observed and no marker related toxicity occurred up to 2 months follow‐up. Conclusion The novel liquid fiducial marker showed to be stable and well visible on CBCT. Negative slope of fits are most likely the result of tumor shrinkage during treatment. Also, injection of the marker in the rectum showed to be technically feasible and no marker injection related toxicity occurred. This novel marker may act as a tumor location surrogate on daily CBCT for image guided radiotherapy in rectal cancer patients. It may provide us a tool to follow day‐to‐day variation in tumor location and thereby enable dose escalation in rectal cancer patients. PO‐1794 Real‐life experiences of an MR‐Only prostate radiotherapy pathway with MR‐CBCT soft‐tissue matching H. McCallum 1,2 , J.J. Wyatt 2,3 , J. Frew 3 , C. Walker 3 , N. Richmond 3 , M. Wilkinson 3 , S. Driver 3 , K. Pilling 3 , B. Rachel L 3 , S. West 3 , R.A. Pearson 2,3 1 The Newcastle Upon Tyne Hospitals NHS Trust, Northern Centre for Cancer Care, Newcastle upon Tyne, United Kingdom ; 2 Newcastle University, Northern Institute of Cancer Research, Newcastle upon Tyne, United Kingdom ; 3 The Newcastle Upon Tyne Hospitals NHS Foundation Trust, Northern Centre for Cancer Care, Newcastle upon Tyne, United Kingdom Purpose or Objective The role of MRI in radiotherapy is increasing due to the superior soft‐tissue contrast offered by MRI compared to X‐ray imaging. When developing an MR‐Only radiotherapy pathway with conventional treatment machine delivery, the validation of a synthetic CT (sCT) is only the start of the process. This paper presents the routine clinical implementation of MR‐Only radiotherapy for prostate cancer patients, believed to be the first using MR‐kV CBCT soft tissue matching on a conventional treatment machine. Material and Methods Patients received a planning MRI scan (1.5T Somatom Espree, Siemens, Germany) using a flat couch top, lasers and a coil bridge. In‐house manufactured silicone dots were used as skin reference marks on the patient. Two MRI sequences were acquired: a 2D T2* weighted small field‐ of‐view (FOV) image for prostate delineation and a 3D T2 weighted large FOV scan for sCT generation and OAR delineation. A sCT was produced using MriPlanner (Spectronic Medical, Sweden). A VMAT treatment plan was produced on the sCT using RayStation (RaySearch Laboratories, Sweden) according to standard local practice. RadCalc (Lifeline Software Inc. USA) was used to provide and independent MU check. The treatment plan and structure set were copied from the sCT onto the large FOV MRI which had been DICOM relabelled as a CT to permit transfer to the treatment machine. Treatment delivery was performed using an on‐line soft‐tissue match comparing the relabelled MRI to a kV‐CBCT on a TrueBeam linear accelerator (Varian, USA). Results The patient workflow is shown in Fig 1.

Early experiences allowed refinement of the procedures. Greater care was required when setting up the patient on the MRI scanner to ensure that the entire skin contour was visualised. Bladder preparation time was reduced by 5 minutes to ensure that the bladder was not over‐filled during MRI acquisition. Delineation training and guidelines were revised to improve MRI delineation accuracy and reproducibility. Treatment radiographer training provided prior to implementation included anatomy training from a physician, benchmark MR‐CBCT matching and off‐line matching to a library of example datasets. This allowed efficient and effective on‐line soft tissue matching using MRI anatomy as the reference image set, illustrated in Fig 2.

Conclusion Routine implementation of an MR‐Only pathway using a conventional treatment delivery machine with X‐ray IGRT has been successfully demonstrated. Building on the foundations of rigorous technical validation of the sCT calculation and MR based treatment verification, a crucial key to smooth implementation was multi‐disciplinary engagement and expertise‐sharing across the whole multi‐ disciplinary team. Despite treatment radiographers having no prior MRI experience, MR‐CBCT matching was found to be achievable and repeatable with a minimum of training. The workflow was enthusiastically received by all staff groups and was perceived to be straightforward and efficient. PO‐1795 Clinical Validation of an AI software to auto segment OARs on CT datasets for Radiation Oncology C. Kota 1 , M. Haas 1 , A. Smith 1 , C. Murphy 1 , H. Zhang 1 1 Reading Hospital. Tower Health, Radiation Oncology, West Reading, USA Purpose or Objective To validate the accuracy of an Artificial Intelligence (AI) based software for auto segmenting normal organs (OARs) in the Head & Neck and Abdomen for Radiation therapy treatment planning. Material and Methods