Registration of Supine MR Mammography with Breast Ultrasound for Surgical Planning of Breast Conserving Surgery: A Feasibility Study

Abstract

To report the feasibility, accuracy and initial clinical experience of the use of real-time magnetic resonance navigated ultrasound (rtMRnUS) in the surgical planning of breast-conserving surgery (BCS) via guide wire insertion. 29 participants were recruited into this prospective ethics committee approved study. The first 4 cases were utilized as a training set. Participants underwent a supine contrast-enhanced breast MR examination with external fiducials and corresponding ink marks placed on the skin of the affected breast to act as co-registration pairs. MR examinations included both functional and morphological images. A LOGIQ E9 ultrasound system (GE Healthcare, Milwaukee, WI, USA) equipped with a 6 - 15 MHz transducer was utilized for rtMRnUS. To facilitate point co-registration of the previously acquired MR dataset with the real-time ultrasound, co-registration pairs were identified on both imaging modalities. The following co-registration quality metrics were recorded: root mean square deviation (RMSD), lesion and global accuracies. Post co-registration guide wire insertion was performed. Co-registration was successfully undertaken in all participants. Results from 25 participants are presented. The median (min, max) RMSD was 3.3 mm (0.6 mm, 8.8 mm). The global accuracy was assessed as very good (8), good (12), moderate (3) and poor (2) while the median (min, max) lesion accuracy was recorded at 8.9 mm (2.1 mm, 33.2 mm). The use of rtMRnUS to facilitate guide wire insertion is a feasible technique. Generally, very good or good global registration can be expected. Lesion accuracy results indicate that a median difference, in 3 D space, of 9 mm can be expected between imaging modalities.