Geometric distortion caused by metallic femoral head prosthesis in prostate cancer imaging on an MR linac: in-vivo measurements of spatial deformation

Abstract

Objectives Metallic implants cause artefacts and distortion on MRI. To ensure accurate dose delivery and plan adaptation on an MR Linac, there is a need to evaluate distortion caused. Methods Participants were imaged on an MR Linac (Elekta Unity (TM)). Three sequences were evaluated. Two vendor supplied (T2W TSE 3D), and one T2W TSE 3D optimised to reduce metal artefact distortions. Images were rigidly registered to CT images by a single observer, using bony anatomy. Three coronal and three axial images were selected, and six paired, adjacent, bony landmarks were identified on each slice. Images bisecting treatment isocentre were included. Difference between landmark coordinates was taken to be measure of distortion. Results Five observers participated. 36 pairs of bony landmarks were identified. Median difference in position of landmarks was <= 3 mm (range 0.3-4.4 mm). One-way ANOVA between observer means showed no significant variation between sequences or patients (p = 1.26 in plane, p = 0.11 through plane). Interobserver ICC was 0.70 in-plane and 0.78 through-plane. Intra-observer ICC for three observers was 0.76, 0.81, 0.83, showing moderate to good reliability on this small cohort. Conclusions This in-vivo feasibility study suggests distortion due to metallic hip prosthesis is not an obstacle for pelvic radiotherapy on an MR Linac. Research on the impact on plan quality is warranted. Advances in knowledge This work supports feasibility of treating patients with metallic hip prosthesis on an MR Linac.