Abstract Book

ESTRO 37

S418

3-plan library

Margin (cm)

Original plan

7-plan library

CTV / ITV

ITV

+0.7 x

x x

x x x x x x x

CTV_full

+1.5 +1.2 +1.0

CTV_empty +1.5

x

+1.2 +1.0

Results 9/10 patients required the original plan to be selected in <30% of fractions. The mean tissue volume spared from using the 3-plan library was 702 cm 3 and 1580 cm 3 from the 7-plan library (p<0.001) compared to the original plan. The V40 and V45 for bowel was significantly different between the original plan and the 7-plan library (mean 26.5% v 25.6%, p=0.003 and mean 21.5% v 20.7%, p= 0.002 respectively) and the 3-plan library and the 7- plan library (mean 26.2% v 25.6%, p=0.001 and mean 21.2% v 2-0.7%, p=0.002 respectively). The bladder V40 and the rectal V45 were also significantly reduced between the original plan and the 7-plan library (mean 72.6% v 68.6%, p=0.041 and mean 80.3% v 76.2%, p=0.006) and the 3-plan library and the 7-plan library.

PO-0804 Knowledge-based planning in image guided adaptive radiotherapy for cancer of the cervix S. Otter 1 , J. Itcovitz 1 , M. Hussein 1 , A. Franklin 1 , A. Stewart 1 1 Royal Surrey County Hospital, St Luke's Cancer Centre, Guildford, United Kingdom Purpose or Objective Locally advanced cervical cancer (LACC) is treated with chemoradiotherapy and brachytherapy. IMRT has been shown to reduce dose to OARs compared to 3D conformal radiotherapy (RT). However, with better conformity comes the risk of missing the CTV as the uterus and cervix are highly mobile. Adaptive RT using the ‘Plan of the Day’ approach allows a suitable plan to be selected based on CBCT imaging for each RT fraction based on the position of the CTV. The use of knowledge-based planning (KBP) reduces the time taken to generate each plan and therefore makes the creation of a plan library clinically feasible. Material and Methods 10 patients with LACC who had received IMRT (50.4Gy/28#) with daily CBCT imaging were identified. The original plan was created from an internal target volume (ITV) taking into account uterus and cervix positions on full and empty bladder CT scans (CTV_full and CTV_empty). The CTV was retrospectively outlined on each CBCT. 3 strategies were compared – the original plan, a 3-plan library and a 7-plan library; see Table 1 for the PTV margins for each strategy. KBP (Varian RapidPlan™) was used to generate each plan. A plan for each fraction was selected based on CTV coverage as the first criteria and the smallest PTV volume as the second criteria. The total volume of tissue spared and OAR DVH parameters were estimated over the course of treatment and compared between the 3 strategies based on the plan selection for each fraction of RT.

Conclusion The 7-plan library significantly reduced volume of tissue irradiated and improved the dose volume parameters for bladder, bowel and rectum compared to the original plan and the 3-plan library. There was no difference between the original plan and the 3-plan library in terms of dose to OAR therefore highlighting that the reduction in margins is crucial to reducing the volume of normal tissue irradiated. Using KBP, the creation of a library of plans with reducing margins is feasible and this strategy ultimately spares OAR and therefore potentially reduces long-term toxicity.