ESTRO 2020 Abstract book

S946 ESTRO 2020

The defined-filling protocol did not result in a reproducible daily gastric volume, which in consequence could alter adequate coverage of the TVs while increasing doses to OARs. In turn, adaptive radiotherapy significantly improved accuracy on dose coverage and OARs sparing. PO-1635 How do UK centres clinically use, commission, and QA deformable image registration? M. Hussein 1 , J. McClelland 2 , R. Speight 3 , C.H. Clark 1 1 National Physical Laboratory, Centre for Metrology in Medical Physics - MEMPHYS, Teddington, United Kingdom ; 2 University College London, Centre for Medical Image Computing - CMIC, London, United Kingdom ; 3 Leeds Teaching Hospitals NHS Trust, Leeds Cancer Centre, Leeds, United Kingdom Purpose or Objective The use of Deformable Image Registration (DIR) in radiotherapy is gradually being clinically adopted as most commercial TPS vendors now offer some form of DIR. For the commissioning and ongoing QA of DIR, there have until recently been limited guidelines [1]. No national information exists on UK centres are using DIR clinically, which applications are being used, and what commissioning and QA tests are performed. This study aimed to survey UK centres to gather this data and to determine whether additional national guidelines are required. Material and Methods The survey was sent to all 71 UK-based radiotherapy centres. Questions were designed to gather information on the following: which applications are DIR algorithms currently used for in clinical practice and for which anatomical sites, how DIR is being commissioned, what type of commissioning data is used, and what routine QA practice is in place and how frequent. The survey also sought to identify whether further UK-specific guidance may be required. DIR software was defined as software that explicitly performs a deformable registration, and saves the resulting transformation and/or deformed images to be further used for another application. Centres that do not use DIR clinically were encouraged to fill in the survey and were asked if they have any future plans and in what timescale. The survey was opened between January and April 2019. Results In total 50 (70%) centres responded. From these centres, 46 reported access to one of the commercial software that could perform DIR according to our definition. In total, 20 centres indicated that they already use DIR clinically, and a further 22 centres had plans to implement an application of DIR within 1 year of the survey. Table 1 shows a summary of the different applications of DIR in clinical use, including the most common clinical sites, type of commissioning data used and frequency of ongoing QA. The most common clinical application of DIR was to propagate contours from one scan to another (19 centres). Table 1 shows that in all of the applications, patient data and patient-specific QA were most frequently used for commissioning tests and ongoing QA respectively. The types of tests performed varied depending on the type of application, with qualitative tests being more common for ongoing QA. For example, Fig 1 shows the different types of tests used for commissioning and ongoing QA for contour propagation. Specific comments from respondents included the need for improved user-friendliness of QA software, difficulty implementing existing QA recommendations and determining when a deformable registration was satisfactory.

Planning and cone-beam CT data (PCT and CBCT) of a gastric lymphoma patient were used and the neoadjuvant therapy of gastric carcinoma was simulated. PTVs and OARs were defined according to the TOPGEAR study protocol (TV 45 Gy/1.8 Gy) with an additional integrated boost (GTV 50.4 Gy/2.016 Gy). For each fraction, patients needed to follow a defined gastric filling regime consisting of a fasting period of 12h followed by the intake of 200 ml (2 glasses of water) immediately before irradiation. On each daily CBCT, OARs and TVs were contoured. Using an image registration platform, non-rigid matching of the PCT and CBCT scans was performed. The treatment plan without adaptation was recalculated on each CBCT (R- CBCT). For analyzing the DVIs of TVs and OARs, each fraction was evaluated for various DVH parameters PTV, GTV and OARs (kidneys, liver, heart) and gastric volumes. Furthermore, an adapted treatment plan was created and optimized for the new anatomy (A-CBCT). The data on gastric volume and dose parameters in the PTV, GTV and OARs were compared.

Results The mean gastric volumes were 277.32 (± 54.40) cm 3 in CBCTs and 519.2 cm 3 in PCT. Mean doses to the PTV did not differ significantly, with a volume variation of 1.54% in average. In parallel, the D95 was improved in the GTV boost volume coverage by 5.26% compared to the R-CBCT plan, as well as the D min with 6.35% better coverage. The mean dose to the heart, liver and right kidney could be reduced with the A-CBTC plan by -35.74%, -10.71% and - 29.47%, respectively. No differences were found in the left kidney or overdosing (V>110%) parameters.

PTV

GTV

Dmean [Gy]

Dmin [Gy]

Dmean [Gy]

D95 [Gy]

DVH-Parameter

CBCTs recalculated (R) (Mean value) CBCTs adapted (A) (Mean value)

45.79

45.30 49.47

46.79

45.09

48.17 50.30

49.26

Difference R to A [%] -1.54

6.35 1.68

5.26

Conclusion