Adaptation of dual isocentre plans for cervix treatments on the 1.5T MR- Linac

Abstract

Purpose or Objective Cervix patients exhibit large inter-fraction anatomical changes. The MR-Linac (MRL) can image these patients with MR prior to & during treatment. Due to the improved image contrast and lack of imaging dose compared to CBCT, daily plan adaptation can be achieved. However, the MRL has a limited treatment field in the Sup/Inf direction of 22 cm. This can restrict the treatment of patients who require longer treatment fields such as cervix patients with nodes. Here we explore adaptive workflows in combination with a dual isocentre approach, to widen the range of cervix patients that can benefit from this treatment. Materials and Methods Four cervix cancer patients were retrospectively planned with a dual isocentre technique delivering 45Gy in 25 fractions. 2 node negative (NN) and 2 node positive (NP) patients were planned using the EMBRACE II study protocol on research Monaco (Elekta AB, Sweden) with an MRL beam model and 1% statistical uncertainty. A 2 cm overlap region between the two isocentres was created, positioned entirely in the nodes (Fig 1a). To create a smooth dose gradient in the junction a structure-based optimisation method was used (Figure 1b). A treatment workflow was simulated using an MR from the second week of treatment, with the plan adapted to the nodal region first using an adapt-to-position (ATP) technique. The plan for the primary disease was fully re-optimised via the adapt-to-shape (ATS) technique onto the nodal region plan using the bias dose method. Additionally, the plans were recalculated without adaptation to simulate a standard workflow. The dosimetry of the adapted and non-adapted plans were compared against the reference plan. Finally an investigation into the effect of intra-fraction motion was performed. Relative shifts between the two isocentres on the adapted plans were shifted superiorly (sup) and inferiorly (inf) by 3mm and 6mm, and recalculated. Results The mean PTV lengths of the NN and NP patients were 19.5 cm and 18.8 cm respectively. To allow room for daily adaptation to anatomical changes none of these patients were eligible for the MRL. Using a dual isocentre technique resulted in field lengths averaging 12.8 cm for isocentre 1 and 7.4 cm for isocentre 2, allowing them to be treated on the MRL. Fig 2 shows dual isocentre adapted plans give better coverage than non-adapted plans. This is true for both NN and NP patients but a larger benefit is seen for nodal boost volumes in the NP cases, with the difference to the reference plan for the D98% being 0.4 Gy and -3.5 Gy for the adapted and recalculated plans respectively. Shifting the adapted plans sup didn’t affect the percentage coverage by more than 3% but shifting them inf caused the PTV coverage on the plan to reduce by ~2.2% per mm for the NP and ~3.2% for the NN cases.Conclusion A dual isocentre technique for cervix treatments can enable an adaptive workflow that recovers the required plan quality. This illustrates the feasibility of a dual isocentre technique for the MRL