Date Published: October 16, 2018
Publisher: Public Library of Science
Author(s): Urte Kägebein, Frank Godenschweger, Brian S. R. Armstrong, Georg Rose, Frank K. Wacker, Oliver Speck, Bennet Hensen, Momen Abayazid.
The aim of this study was the development and evaluation of a real-time guidance support using optical Moiré Phase Tracking (MPT) for magnetic resonance (MR) guided percutaneous interventions. A gradient echo sequence, capable of real-time position updates by the MPT system, was modified to enable needle guidance based on four rigidly attached MPT markers at the back of a needle. Two perpendicular imaging planes were automatically aligned along the calibrated needle and centered at its tip. For user guidance, additional information about the needle trajectory and the tip to target distance were added as image overlay. Both, images and guiding information were displayed on the in-room monitor to facilitate MR guided interventions. The guidance support was evaluated by four experienced interventional radiologists and four novices targeting rubber O-rings embedded in a custom-made phantom on a 3T wide-bore MRI system (80 punctures). The skin to target time, user error, system error and total error were analyzed. The mean skin to target time was 146s±68s with no statistically significant difference between experts and novices. A low mean user error (0.91mm±0.43mm), system error (0.53mm±0.27mm) and total error (0.99mm±0.47mm) was reached in all directions. No statistically significant difference in user error, system error and total error could be found between experts and novices. The presented tracking and image guidance system combined with the user interface offers continuous and interactive control of the imaging plane while puncturing in the magnet enabling accurate real-time feedback for both, experienced and non-experienced users.
Magnetic resonance imaging (MRI) offers unique approaches to not only diagnose but also treat cancer [1–3]. Compared to CT or ultrasound, MRI has superior soft tissue contrast, lacks ionizing radiation and enables complex double-oblique trajectories . Furthermore, MRI has the unique ability to acquire functional information (e.g. temperature, perfusion). This can be used to accurately control the success of thermal ablation therapies . However, interventional MRI (iMRI) is still limited to specialized clinical centers due to access and workflow limitations [5, 6]. In order to make MRI guidance a standard procedure, the workflow has to be adjusted to be more time efficient. This is especially important for the so called freehand technique [2, 4]. Here an interventional device is advanced to the target using continuous imaging. A dedicated interventional platform, such as the Interactive Front End (IFE, Siemens Healthineers, Erlangen, Germany) makes the procedure more intuitive . In contrast to classical in-out procedures, real-time imaging allows for compensation of organ movement . Main disadvantages include the need for manual slice adjustment, in particular when the needle or the target is lost from the imaging plane. This process can quickly become cumbersome and time-consuming without a well-coordinated team, especially in cases of complex trajectories . In this realm, needle guidance support and streamlined interfaces might be able to improve precision and procedure time.
Once the real-time needle guidance sequence was started, the modified GRE sequence automatically aligned the two imaging planes along the needle centered at the needle tip using the pose information of the MPT marker. The images were sent in real-time with negligible displaying delay to the in-room monitor providing a constant feedback for the users about the needle trajectory, needle extension and target-tip distance (see Fig 5A). Thus, the users were able to continuously navigate the needle to the predefined targets by translating the needle within the MRI bore and utilizing the real-time feedback on the in-room monitor. The update rate of the real-time needle guidance including the processing time for image plane update amounted to 1fps, being equivalent to the acquisition time of the GRE-sequence. The update rate of the MPT system (including internal system delays) was much higher (9fps), but the GRE-sequence utilized only the latest pose information of the MPT marker. In total, the mean skin to target time was 146s±68s for the experts and 165s±55s for the novices (see Table 1). There was no significant difference between experienced and non-experienced users (p = 0.17).
We have developed and evaluated a system for MRI-guidance support for percutaneous punctures. The MPT system offered continuous and interactive control of the imaging slices in real-time directly from the inside of the MRI magnet while maintaining the freehand convenience. Using two orthogonal imaging planes in combination with a graphical overlay, both, experts and novices have successfully navigated the needle to a predefined target with no significant difference in targeting error or skin to target time. Thus, we conclude, that the developed guidance support seems to be so simple and intuitive that even novices can achieve similar results as experts in terms of speed and accuracy.