ESTRO 2020 Abstract book

S1123 ESTRO 2020

the MV image with that of the DRR generated after treatment planning. Results The new application of Medusa for patient monitoring during radiation therapy asked for solutions for camera positioning and zoom function, room illumination and interference of images due to scattered radiation. We included all this in a project aimed at implementing respiratory control for breast cancer patients avoiding the need for an active system like spirometry. Analysis of the visible pulmonary margin in the tangential beam shows no significant difference between "medusa" patients and "spirometry" patients (unpaired student test p = 0.35).

• •

12 training cases

12 test cases The PoD selections for the 12 test cases were compared to the gold-standard consensus selection. The on-trial QA programme initially consisted of retrospective data collection of CBCTs and PoD selections. The PoD selections for the on-trial patients were compared to the offline gold-standard expert reviewer selection. For each patient, i.e. submission, a guideline compliance score was calculated = number of compliant PoD selections/total number of PoD selections. Differences in PoD selection compliance scores between pre-trial and on- trial submissions, trials, PoD experience, centre size and recruitment rate were tested using Mann-Whitney U and Kruskal-Wallis tests. Results 8,286 PoD selections from 35 centres were reviewed, equating to 1,339 PoD submissions, Table 1.

Conclusion In a series of patients treated using spirometry, we confirmed the utility of Medusa to check both the amplitude and the reproducibility of the inspiration. The RTT’s, using skin-drawn marks visualised by the Medusa screen, can guide the patient to the required inspiration during the pre-treatment CT-scan and subsequently throughout the entire treatment series to verify and maintain a stable extent of apnoea. PO-1917 Pre-trial plan selection training - is it sufficient for optimal plan selections online? A. Webster 1 , S. Hafeez 2,3 , H. McNair 3,4 , V.N. Hansen 5 , C. Griffin 6 , R. Lewis 6 , E. Hall 6 , R. Huddart 3,7 1 National Radiotherapy Trials Quality Assurance Group, Mount Vernon Cancer Centre, London, United Kingdom ; 2 Institute of Cancer Research, Department of Radiotherapy and Imaging, London, United Kingdom ; 3 The Royal Marsden NHS Foundation Trust, Department of Radiotherapy, London, United Kingdom ; 4 Institute of Cancer Research, Translational Therapeutic Radiography Team, London, United Kingdom ; 5 Odense University Hospital, Department of Radiotherapy Physics, Odense, Denmark ; 6 Institute of Cancer Research, Clinical Trials and Statistics Unit, London, United Kingdom ; 7 Institute of Cancer Research, Clinical Academic Radiotherapy Team, London, United Kingdom Purpose or Objective HYBRID (CRUK/12/055) and RAIDER (CRUK/14/016) are two randomised phase II multicentre bladder trials evaluating plan of the day (PoD) radiotherapy. Previously we reported the introduction of a pre-trial PoD Quality Assurance (QA) programme (ESTRO 36 OC-0351) and the initial guideline compliance in on-trial PoD selections for RAIDER (ESTRO 38 OC-0634). Here we aim to compare the difference in compliance between pre-trial and on-trial PoD selections, reported and unreported, to determine QA requirements. Material and Methods The pre-trial QA programme initially included: • A video overview of PoD • A PoD selection guidance document

The compliance score of PoD selections pre-trial was significantly higher than PoD selections on-trial ( P <.001), Table 2. Thus, the PoD selections online were less compliant than PoD selections offline. Compliance scores differed by centre size (p=.01) and recruitment rate (p=.03) with median and mean scores both lowest in small centres and in centres with low recruitment.

Therefore the support for all centres in on-trial PoD selections was reviewed. Our approach was to engage with the centres and collaborate together. PoD selection documentation was updated and formal on-trial PoD reports were produced. For HYBRID we conducted centre visits, reviewing the PoD selections for each centre’s first patient. For the larger RAIDER trial, remote access interactive workshops were hosted with centres, including a PoD presentation and Q&A. With this additional support, the PoD selection on-trial in RAIDER has improved from 72.5% to 100% in the median compliance score (P<0.001) (and 70% to 95% in the mean compliance score). Conclusion This work highlights there is a difference between the pre- trial and on-trial PoD selections. Building upon previous work it is clear that pre-trial training alone is not sufficient to maintain compliance in online PoD selections. For optimal PoD selections online, continuous on-trial support