Accuracy of applicator tip reconstruction in MRI-guided interstitial 192Ir-high-dose-rate brachytherapy of liver tumors

Background and purpose To evaluate the reconstruction accuracy of brachytherapy (BT) applicators tips in vitro and in vivo in MRI-guided 192Ir-high-dose-rate (HDR)-BT of inoperable liver tumors. Materials and methods Reconstruction accuracy of plastic BT applicators, visualized by nitinol inserts, was assessed in MRI phantom measurements and in MRI 192Ir-HDR-BT treatment planning datasets of 45 patients employing CT co-registration and vector decomposition. Conspicuity, short-term dislocation, and reconstruction errors were assessed in the clinical data. The clinical effect of applicator reconstruction accuracy was determined in follow-up MRI data. Results Applicator reconstruction accuracy was 1.6±0.5mm in the phantom measurements. In the clinical MRI datasets applicator conspicuity was rated good/optimal in ⩾72% of cases. 16/129 applicators showed not time dependent deviation in between MRI/CT acquisition (p>0.1). Reconstruction accuracy was 5.5±2.8mm, and the average image co-registration error was 3.1±0.9mm. Vector decomposition revealed no preferred direction of reconstruction errors. In the follow-up data deviation of planned dose distribution and irradiation effect was 6.9±3.3mm matching the mean co-registration error (6.5±2.5mm; p>0.1). Conclusion Applicator reconstruction accuracy in vitro conforms to AAPM TG 56 standard. Nitinol-inserts are feasible for applicator visualization and yield good conspicuity in MRI treatment planning data. No preferred direction of reconstruction errors were found in vivo.