The basic principles of MR-US fusion and potential applications to the musculoskeletal system will be discussed. Development of a simple phantom for training will be described. Examples of injections performed to date will be shown to illustrate potential applications and challenges in obtaining adequate registration.
MR-US fusion imaging for guided injections in the musculoskeletal system
Target audience: Radiologists, Clinicians, Physicists and Engineers
Objective: At the conclusion of this activity, participants will understand the basic principles of MR-US fusion and potential applications to the musculoskeletal system. Development of a simple phantom for training will be described. Examples of injections performed to date will be shown to illustrate potential applications and challenges in obtaining adequate registration.
Summary: Ultrasound guided injections have become increasing utilized over the past several years. In many instances, these derive from direct referrals following MR or CT. Correspondence of abnormalities seen on cross-sectional imaging, however, may not always be readily evident on ultrasound, Examples would include deep seated abnormalities which are not amenable to high frequency transducers or low contrast lesions on gray-scaleultrasound that are highly conspicuous on MR. MR-US co-registration provides a method to visualize the relevant soft tissue anatomy with greater confidence, while allowing for differences in the image acquisition plane. MR-US fusion has largely been applied to abdominal applications, with relatively few musculoskeletal applications reported to date. The basic requirements of registration include having a representative high-resolution MR (or CT) data set, a method to track the transducer orientation in space (usually through a “Flock of Birds technique”), and the selection of internal fiduciary points common to both data sets (minimally 3 points at different locations). Based on phantom data, accuracies of approximately 5mm or less can be achieved. A brief survey of the technical aspects of co-registration will be discussed, followed by the typical sequence employed to achieve registration with musculoskeletal applications in mind. A series of clinical examples will be presented from therapeutic injections, aspirations or biopsies in the musculoskeletlal system performed at our institution. Use of a relatively simply constructed gelatin phantom will also be described to serve as a potential learning tool to improve individual throughput in the early phase of performing these procedures. The challenges, limitations and potential benefits of using real-time co-registration techniques will be discussed.
Christopher Burke, MD
William Walter, MD
Mamadou Diallo, PhD