Session Item

There is increasing interest in dose escalation to the GTV in patients receiving radiotherapy for localised prostate cancer. The MR-Linac (MRL) presents a unique opportunity to facilitate this approach through improved target definition and adaptive radiotherapy delivery. We aimed to assess intra-fraction GTV motion during MRL treatment delivery and to evaluate the efficacy of a PTV derived based on internal target volume (ITV).  

Five patients with an identifiable GTV who received whole gland prostate radiotherapy on the MRL were selected for this retrospective planning study. During each treatment session, three T2 MR images were obtained – localization for adaptation (MR_Loc), verification pre-treatment (MR_Ver), and during beam-on (MR_BeamOn). Average session time was 38 minutes. MR_Loc and MR_BeamOn from 5 treatment sessions for each patient were retrieved. To assess for intra-fraction motion, displacement of the GTV on MR_BeamOn was calculated using coordinates of the centre of mass. An ITV was generated summing GTV_Loc and GTV_BeamOn. Two PTVs were defined to compare dosimetry: GTV_Loc + 2 mm (GTV-PTV), and ITV + 2 mm (ITV-PTV). A 9-beam IMRT plan was generated to deliver 15 Gy in a single fraction to each PTV on the MR_Loc scans. Rectum, urethra and bladder were included in OAR assessment. 

A total of 50 GTVs and 25 ITVs were analysed. Median GTV was 1.05 cm³ (range: 0.46 – 2.77). Intra-fraction GTV motion was greatest in the transverse direction (mean = 2.2 mm), followed by cranio-caudal (mean = 1.4 mm), and lateral (mean = 0.7 mm). Three of the 5 patients had consistent GTV displacements of < 3mm. In the remaining two patients, shifts of 3.1 – 10.0 mm were demonstrated. Median GTV_BeamOn D95% was 16.5 Gy for GTV-PTV (range: 5.5 – 18.9). For ITV-PTV, this was 17.7 Gy (range: 16.3 – 19.5). Median OAR doses were comparable between GTV-PTV and ITV-PTV. Urethra D10% was 10.6 Gy and 10.0 Gy, respectively; rectum D0.5cc 8.2 Gy and 9.2 Gy; and bladder D0.5cc 3.3 Gy and 3.5 Gy. 

Variable intra-fraction shifts in GTV were found between patients, with resultant discrepancies in target coverage. The results suggest that generation of an ITV may allow for improved accuracy in GTV dose escalation on the MRL, whilst maintaining acceptable OAR dose constraints.