ESTRO 2024 - Abstract Book

S6031

RTT - Treatment planning, OAR and target definitions

ESTRO 2024

Stage 2. Observers delineated GTV_MR on the MR image, this included disease visualised on MR. The GTV_MR target volume was copied over to the 3DCECT_BH structure named GTV_MR_CT and adjusted to account for anatomy visualised on the multi-modality images i.e. the resultant GTV structure including disease. Dice similarity coefficient (DSC) was calculated between GS and OBS- categorised according to score (0-0.2 very poor, 0.21-0.4 poor, 0.41-0.6 moderate, 0.61-0.8 very good, 0.81-1.0 excellent).

Volume descriptions Stage1.3DCECT_BH+4DCT delineations GTV_3D – Pancreatic tumour visible on image

GTV_inhale – On the maximum inhale bin of the 4DCT delineate GTV GTV_exhale - On the maximum exhale bin of the 4DCT delineate GTV ITV_CT-On 3DCECT_BH create the ITV using boolean operators to produce union of GTV_3D+GTV_ inhale+GTV_ exhale Once ITV_CT is created, verify involved disease is covered on all phases of 4DCT

Stage2.Registered 3DCECT+4DCT+MR delineations GTV_MR-delineate disease visualised on MR only GTV_MR_CT-GTV_3D amended using registered MR and CT scan ITV_MR –GTV_MR_CT and any motion from GTV_inhale and GTV_exhale

Table 1. Summary of volumes/delineations and datasets used (CT only, MR only, and MR/CT registered).

Results:

Eight consecutive patients were anonymised and assessed for the study, 3/8 patients were excluded due diffuse tumour appearance on RT planning images. Five patients were included in the analysis (patient, 1, 2, 3, 5 and 8).

Total volumes analysed were 175 (GTV_3D/GTV_exhale/GTV_inhale/ GTV_MR/GTV_MRCT/ITV_CT/ITV_MR for 5 patients), completed by GS team, and 4 further observers i.e. 2 CO (Obs 21 and Obs 65) and 2 COtrain (Obs 43 and 65). Mean volumes (SD) for individual patients are shown in figure 1. Mean GTV_MR structures were smaller than GTV_3D for 4/5 patients. GTV_MRCT was greater than GTV_3D by 49%, 65%, 26%, < 1% for patients 1,2,3 and 5; and 11% lesser for patient 8. ITV volumes created using CT only (ITV_CT), and MR and CT (ITV_MR) were the similar. Observer volumes were larger across all structures compared to GS (Figure 1). Highest magnitude of mean difference was observed between OBS volumes with and without MR i.e. (i) GTV_MR and GTV_3D, difference of 3cm3 (p=0.048) and (ii) GTV_3D and GTV_MRCT, -3.7cm3 (p= 0.145). Mean across all observers+GS volumes between GTV_MR and GTV_3D was 3.22cm3 p=0.019 (GTV_MR smaller volumes); and was - 3.53cm3, p=0.040 for GTV_MRCT and GTV_MR (larger volumes than 3D_GTV).