Real-Time MRI of Joint Movement With TrueFISP

Transcription

1 JOURNAL OF MAGNETIC RESONANCE IMAGING 15: (2002) Technical Note Real-Time MRI of Joint Movement With TrueFISP Harald H. Quick, MSc, 1 * Mark E. Ladd, PhD, 1 Matthias Hoevel, MD, 2 Silke Bosk, RT, 1 Joerg F. Debatin, MD, 1 Gerhard Laub, PhD, 3 and Tobias Schroeder, MD 1 Purpose: To develop a technique for dynamic magnetic resonance imaging (MRI) of joint motion based on a combination of real-time TrueFISP (fast imaging with steady state precession) imaging with surface radiofrequency (RF) coils. Materials and Methods: The metacarpal, elbow, tarsal, and knee joint of five volunteers and the knees of four patients were examined with a real-time TrueFISP sequence during movement of the joints. Results: All examined joints could be assessed under dynamic conditions with high image contrast and high temporal resolution. Conclusion: Dynamic MRI of joints with TrueFISP is feasible and can provide information supplemental to static joint examinations. Key Words: joints; dynamic examination; TrueFISP; RF surface coils; real-time MRI J. Magn. Reson. Imaging 2002;15: Wiley-Liss, Inc. 1 MR Center, Department of Diagnostic Radiology, University Hospital Essen, Essen, Germany. 2 Department of Orthopedic Surgery, University Hospital Essen, Essen, Germany. 3 Siemens Medial Systems, Chicago, Illinois. Presented at the 9th Annual Meeting of ISMRM, Glasgow, April p *Address reprint requests to: H.H.Q., MSc, University Hospital Essen, Department of Radiology, MR-Center, OZ II, Hufelandstr. 55, D Essen, Germany. HHQuick@uni-essen.de Received November 6, 2001; Accepted February 15, DOI /jmri Published online in Wiley InterScience ( DYNAMIC MAGNETIC RESONANCE IMAGING (MRI) techniques have been developed to provide diagnostic information related to the functional aspects of joints. Dynamic MRI evaluation of joints is used to assess the various interactions of the soft tissues and bony anatomic structures that comprise the joint, and to evaluate the relative alignment of these structures through a specific range of motion (1 4). The temporomandibular joint, patellofemoral joint, wrist, shoulder, cervical spine, and ankle have been studied with dynamic MRI techniques (4). The evaluation of joint movement under dynamic conditions can provide important diagnostic information for detailed assessment of joint abnormalities and malformations, as has been shown for the patellofemoral joint (1,4,5). Diagnostic evaluation of moving joints in MRI can be performed using basically two different approaches: with kinematic MRI, using motion-triggered cine MRI techniques (3,5 7), or alternatively, with dynamic MRI, using ultrafast real-time imaging techniques (8) that do not require any motion triggering. Because of its short examination time and good image quality, dynamic MRI with the use of ultrafast sequences has been shown to be less dependent on patient compliance and thus superior to motion-triggered and static MRI, especially in the evaluation of uncooperative patients and in instances of severe joint pain (5). Thanks to the availability of very short repetition times, MR scanners equipped with the latest generation of powerful gradient systems permit real-time gradientecho imaging based on TrueFISP (fast imaging with steady state precession). Short repetition times are a precondition for TrueFISP imaging to avoid banding artifacts associated with field inhomogeneities. Realtime TrueFISP MRI allows the acquisition of high-contrast images with high temporal resolution (9). This study evaluates the diagnostic potential of an approach based on real-time TrueFISP for dynamic MRI of joint pathologies. MATERIALS AND METHODS Coil Support Design Imaging was performed on a 1.5-T Siemens Sonata system equipped with high performance gradients (40 mt/m maximum amplitude, 200 mt/m/msec slew rate). Two different U-shaped coil supports (Fig. 1) were designed out of Plexiglas to hold the flexible circularly polarized (CP) surface radiofrequency (RF) receive coils of the scanner manufacturer (Siemens CP SmallFlex and CP LargeFlex). The holders allow free movement of the hand, ellbow, knee, or ankle joint in a single plane. The coil supports were fixated via a snap mechanism to a multi-purpose base plate on the table of the scanner. Imaging Sequence A single-slice, real-time TrueFISP sequence was employed for dynamic joint evaluation (Fig. 2). TrueFISP is a steady-state precession gradient-echo sequence Wiley-Liss, Inc. 710

2 Real-Time MRI of Joint Movement 711 Figure 1. Plexiglas holder for the flexible surface coil attached to the multi-purpose base plate on the table of the 1.5-T MR system. Compared to FISP, where only one or two of the gradient axes are balanced, the TrueFISP sequence is characterized by completely balanced gradients in all three directions (9). At the end of the repetition time (TR), the transverse magnetization is refocused and the next excitation ( pulse) can be started without further prepa- Figure 2. Pulse diagram of a TrueFISP sequence. All three gradients (Gslice, Gread, Gphase) are refocused in order to retain maximal transverse magnetization before the next excitation pulse. ration. The use of very short TRs ( 4 msec) is mandatory to avoid T2* effects. The real-time TrueFISP sequence (TR/TE 2.2/1.1 msec, flip angle 50, field of view (FOV) cm, slice thickness 6 mm) collects one line every 2.2 msec. The in-plane data acquisition matrix was pixels. The acquisition time per image amounted to 297 msec ( msec). Echosharing was employed to improve temporal resolution (Fig. 3). For this purpose, 15 additional lines of central k-space were collected between image one and image three. Image two was subsequently reconstructed from k of image one, k of image three, and the 15 additional central k-lines inbetween. Echo-sharing improved the temporal resolution to 165 msec, thus enabling the acquisition of 60 images during 10 seconds Figure 3. Real-time TrueFISP sequence. Fifteen additional lines (black bar between k and k ) are collected to improve temporal resolution using echo sharing.

3 712 Quick et al. Figure 4. Selection of joints that can be dynamically assessed with the proposed technique inside a closed-bore magnet. Metacarpal joint (a), elbow joint in a sagittal (b) and same joint in a coronal (c) plane, tarsal joint (d), knee in a sagittal plane (e), and another knee in a coronal (f) plane. of joint movement. Because no triggering of the joint movement was necessary for imaging purposes, synchronization of joint motion and imaging was uncritical. For the experiments, movement was provoked shortly after the imaging sequence was started. The real-time images were reconstructed on-line and displayed with no detectable delay on an in-room console that was placed next to the scanner. The investigator provoking the joint motion was thus provided with instant real-time feedback. Patient Study Five healthy volunteers and four patients were enrolled in this study in accordance with the regulations of the local institutional review board. The dynamic joint motion study was performed on four joints (elbow, knee, metacarpal, tarsal) of five volunteers (mean age 30 3 years), and on the knees of four patients (mean age 9 3 years). The patients had a history of muscular contracture of the knee joint bilaterally (N 1), lateral instability of the knee joint (N 1), lateral subluxation of the tibia against the femur (N 1), and a residual hemangioma at the ligamentum patellae (N 1). For all motion studies, the volunteers and patients were placed inside the scanner in a prone position with the joint under investigation placed in the middle of the coil. Sagittal planes were acquired to assess flexion and extension of the elbow, the knee, and the tarsal joint. Coronal slices were acquired to additionally assess supination and pronation of the elbow, as well as lateral instability of the knee in one patient. The metacarpal joint was assessed in a coronal plane. RESULTS Coil Support Design The RF coil supports, in combination with the flexible surface coils, enabled optimization of the signal-tonoise ratio (SNR) and signal homogeneity over the volume of the joint under investigation. The open design of the receive coil supports allowed for flexion and extension, as well as for supination and pronation, of the evaluated joints without moving the coils relative to the scanner. Despite the spatial constraints of a closedbore whole body scanner, all evaluated joints of the volunteers and patients could be assessed under dy-

4 Real-Time MRI of Joint Movement 713 Figure 5. Selection of six out of 60 images that were acquired in a saggittal plane during 10 seconds of flexion and extension of the foot. Flexion and extension of the tibiotalar part of the ankle joint could be depicted in detail. namic conditions (Fig. 4). For adult patients, however, the flexion of knees was limited to approximately 45. Imaging Sequence The TrueFISP sequence was characterized by high image contrast, which allowed good depiction of muscles (hypointense), bones (hyperintense), ligaments (hypointense), and menisci (intermediate signal). The soft tissue contrast was thus, to a large extent, comparable to that of a conventional proton density (PD)-weighted spin-echo sequence. The imaging sequence was furthermore characterized by high temporal resolution, which allowed for detailed dynamic analysis of active joint motion (Fig. 5). With six reconstructed images per second, temporal resolution of the sequence was sufficient to resolve joint motion without the appearance of motion artifacts (Fig. 5). In some of the joint motion studies, however, image quality was degraded by the appearance of band artifacts. Due to the strong dependency of TrueFISP image quality on field homogeneity, a three-dimensional shim routine turned out to be a mandatory prerequisite for the removal of band artifacts in the subsequently acquired dynamic images. Patient Study Joint instability and muscular contracture could be assessed in the knees of the three patients. In one patient, a residual 20 lateral instability following surgical correction of the knee could be depicted. In another patient, a discontinuous lateral translation of the tibia, which could be provoked at will, could be observed. In a patient with a known bilateral muscular contracture, the static MR exam showed that the tibial head was positioned slightly posterior to the femoral condylus. The dynamic real-time exam additionally showed a limited 20 extension due to the muscular contracture. There was no detectable mechanical cause for the limited range of movement. In the fourth patient investigated, a limited 20 extension was found to be caused by residual hemangioma at the ligamentum patellae. Dynamic real-time imaging showed compression

5 714 Quick et al. Figure 6. High-resolution static proton density-weighted turbo spin-echo (a), T1-weighted spin-echo (b), and dynamic real-time TrueFISP image in the left knee of a 14-year-old male patient with residual hemangioma at the ligamentum patellae (arrows). Dynamic examination with TrueFISP (c) revealed compression of the hemangioma between the patella and the head of the tibia being responsible for an extension deficit of 20. of the residual hemangioma, which was pushing up from beneath the patella (Fig. 6). DISCUSSION By combining flexible RF receive coils with dedicated coil supports and a real-time, motion-insensitive imaging sequence, high-quality images of joints under motion could be achieved. The spatial resolution and contrast achieved with real-time TrueFISP allowed good depiction of muscles, bones, and ligaments, therefore enabling detailed analysis of joint movement. In comparison to other fluoroscopic MRI techniques, such as single-shot fast spin-echo (SSFSE), echo-planar imaging (EPI), or spiral acquisitions, the SNR of TrueFISP sequences is largely independent of the TR due to refocussing of the transverse magnetization in all three directions. This allows acquisition with very short TRs while simultaneously maintaining high SNR, which often is a conflicting goal for the aforementioned fast acquisition techniques. TrueFISP imaging requires high homogeneity of the static magnetic field within the FOV; otherwise image quality can be severely degraded by stripe artifacts originating from field inhomogeneities. These stripe artifacts are especially predominant and problematic when the volume under investigation undergoes movement during imaging, as in this real-time motion study. Performing a three-dimensional shim routine before starting TrueFISP imaging turned out to be a mandatory and effective prerequisite. Additionally, the shim and the RF prescan routine should be performed with the extremity in the middle of the range of motion under investigation, i.e., if the range of elbow movement allows for a 60 angle, shimming and the acquisition of the scout image should be performed with that joint angulated in the 30 position. This strategy ensured an almost artifactfree acquisition of the subsequent TrueFISP real-time images over the full range of motion. Despite the spatial restrictions of a closed-bore whole body scanner, all evaluated joints could be assessed under dynamic conditions. For adult patients, however, the flexion of knees was limited to approximately 45. Compared to the closed bore design of the MR scanner employed in this study, open magnet design MR scanners, such as the C-shape, double doughnut, and various dedicated extremity scanners, would potentially offer more room for investigation over the full range of joint motion. Additionally, those scanners could potentially enable the hip and the shoulder joints to be imaged under dynamic conditions. However, open MR scanner designs in general do not offer the gradient performance, static magnetic field strength, or field homogeneity necessary to perform high-resolution realtime TrueFISP imaging with the same quality. Because the additional time required to perform dynamic MRI of a joint is relatively short, this procedure can be included with routine static MRI of the joint, since both procedures can be completed within an acceptable time in the clinical setting. This provides a more thorough overall examination of the joint than would be achieved with static MRI alone. In conclusion, the proposed approach for dynamic MRI of joints renders detailed dynamic analysis of joint motion. Therefore, the technique appears well suited for the assessment of joint abnormalities that express themselves only under dynamic conditions, such as neuromuscular contractures and joint instabilities and malformations. Furthermore, dynamic studies of joint motion may be used for assessment of the therapeutic response after traumatic injuries. Clearly, the clinical potential can only be determined based on larger studies. REFERENCES 1. Shellock FG, Mink JH, Fox JM. Patellofemoral joint: kinematic MR imaging to assess tracking abnormalities. Radiology 1988;168: Shellock FG, Mink JH, Deutsch AL, Pressmann BD. Kinematic magnetic resonance imaging of the joints: techniques and clinical applications. Magn Reson Q 1991;7: Melchert UH, Schröder C, Brossmann J, Muhle C. Motion-triggered cine MR imaging of active joint movement. Magn Reson Imag 1992; 10:

6 Real-Time MRI of Joint Movement Shellock FG, Deutsch AL, Mink JH. Kinematic MR imaging of the patellofemoral joint: basic and advanced techniques. In: Syllabus: a categorial course in muscoskeletal radiology. 79th Scientific Assembly and Annual Meeting of the Radiological Society of North America. Chicago, IL, November 28 December p Muhle C, Brossmann J, Heller M. Kinematic CT and MR imaging of the patellofemoral joint. Eur Radiol 1999;9: Brossmann J, Muhle C, Schroder C, et al. Patellar tracking patterns during active and passive knee extension: evaluation with motiontriggered cine MR imaging. Radiology 1993;187: Muhle C, Brinkmann G, Brossmann J, Wesner F, Heller M. Kinematic MR imaging of the ankle: initial results with ultra-fast sequence imaging. Acta Radiol 1997;38: Rasche V, Holz D, Proksa R. MR fluoroscopy using projection reconstruction multi-gradient-echo (prmge) MRI. Magn Reson Med 1999; 42: Barkhausen J, Ruehm SG, Goyen M, Buck T, Laub G, Debatin JF. MR evaluation of ventricular function: true fast imaging with steadystate precession versus fast low-angle shot cine MR imaging: feasibility study. Radiology 2001;219: