Breast screening is done to determine whether lesions are present in the breast. The biopsy of found lesions is usually done under the guidance of Ultrasound (US) to track the location of the needle in the breast. Some cancer types cannot be detected using US. Therefore, Magnetic Resonance Imaging (MRI) will instead be used for these types of cancer. An MR image gives a highly detailed image of the breast and enables visualization of most cancer types. A major disadvantage of using MRI for guided biopsies is the lack of real time information, since metal objects and electronics cannot go near the scanner once it is turned on due to the magnetic field this machine produces.
This work describes the implementation of a Mixed Reality guidance system to aid the surgeon when performing an MRI-guided biopsy. The proposed solution consists of 3 components. Measurements were performed using an Optitrack camera system which tracked retro reflective markers in 3D space. Matlab was used as the processing software and filtered and matched the measured marker data to the relevant objects and updated the position and shape of the hologram accordingly. The Thin Plate Spline algorithm was used to track deformations of the breast that could occur during the biopsy and updated the hologram to account for the change of position of the lesion. The Microsoft HoloLens was used to visualize the holograms of the breast which has been created from MRI data and placed them at the correct position in 3D space.
The solution is capable of placing the hologram with an error (E) of (Ex;Ey;Ez) = (7:2; 9:6; 8:1)mm. 72 biopsies have been performed in this work, resulting in a 64% success rate when TPS was disabled and a 28% success rate when TPS was enabled. By combining those outcomes, a total success rate of 46% was achieved for the current implementation.
To join the presentation via Microsoft Teams click here
