The authors have declared that no competing interests exist.
Conceived and designed the experiments: AJC RDB MHB II RMS. Performed the experiments: MHB AJC ZA II GJS. Analyzed the data: RDB ZA II AJC. Contributed reagents/materials/analysis tools: RMS AJC MHB. Wrote the paper: AJC RDB MHB II RMS.
Non-invasive imaging techniques such as magnetic resonance imaging (MRI) provide the ability to evaluate the complex anatomy of bone and soft tissues of the wrist without the use of ionizing radiation. Dynamic instability of wrist – occurring during joint motion – is a complex condition that has assumed increased importance in musculoskeletal medicine. The objective of this study was to develop an MRI protocol for evaluating the wrist during continuous active motion, to show that dynamic imaging of the wrist is realizable, and to demonstrate that the resulting anatomical images enable the measurement of metrics commonly evaluated for dynamic wrist instability.
A 3-Tesla “active-MRI” protocol was developed using a bSSFP sequence with 475 ms temporal resolution for continuous imaging of the moving wrist. Fifteen wrists of 10 asymptomatic volunteers were scanned during active supination/pronation, radial/ulnar deviation, “clenched-fist”, and volarflexion/dorsiflexion maneuvers. Two physicians evaluated distal radioulnar joint (DRUJ) congruity, extensor carpi ulnaris (ECU) tendon translation, the scapholunate (SL) interval, and the SL, radiolunate (RL) and capitolunate (CL) angles from the resulting images.
The mean DRUJ subluxation ratio was 0.04 in supination, 0.10 in neutral, and 0.14 in pronation. The ECU tendon was subluxated or translated out of its groove in 3 wrists in pronation, 9 wrists in neutral, and 11 wrists in supination. The mean SL interval was 1.43 mm for neutral, ulnar deviation, radial deviation positions, and increased to 1.64 mm during the clenched-fist maneuver. Measurement of SL, RL and CL angles in neutral and dorsiflexion was also accomplished.
This study demonstrates the initial performance of active-MRI, which may be useful in the investigation of dynamic wrist instability
The wrist is considered unstable clinically if it exhibits symptomatic dysfunction, is not able to bear loads, and does not exhibit normal kinematics during any portion of the wrist’s arc of motion
Magnetic resonance imaging (MRI) has the advantage of producing improved soft tissue contrast compared to other imaging modalities, without the concerns of ionizing radiation. Conventional non-contrast MRI studies of wrist instability performing static imaging have focused largely on the diagnosis of ligament derangements in the neutral position. These investigations however have reported inconsistent results. Using arthroscopic and arthrographic techniques as reference standards, studies have found that conventional MRI does not consistently diagnose all tears of the SL ligament, lunotriquetral ligament, or triangular fibrocartilage complex
Fast gradient-echo MRI pulse sequences termed “balanced steady-state free precession” (bSSFP) are capable of generating images rapidly (e.g., <600 ms per image) with a matrix size of 128×128, and with high ratios of both signal-to-noise and contrast-to-noise
An MRI protocol, which we term “active-MRI”, was developed for evaluating the wrist during active motion. We use the “active-MRI” terminology here instead of “dynamic-MRI” to avoid potential confusion with dynamic (contrast-enhanced) MRI, which refers to scanning conducted to measure MRI contrast agent dynamics, or with “cine-MRI” where static MR images are played in a loop. Our objectives in this work were [i] to evaluate the technical feasibility and reliability of acquiring bSSFP images during active wrist motions (i.e., during active supination/pronation, radial/ulnar deviation, a dynamic clenched fist maneuver, and volarflexion/dorsiflexion), and [ii] to show that the resulting images enable the standardized derivation of metrics typically associated with wrist instability (i.e., DRUJ subluxation ratio, ECU tendon translation, SL interval, ulnar variance, and the SL, RL and CL angles).
Ten asymptomatic volunteers (7 men, 3 women; average age 36 years; age range 27–58 years) were recruited for this prospective, HIPAA-compliant study. The study had approval from the University of California, Davis Institutional Review Board (IRB) and was conducted at the University of California, Davis, USA. Written informed consent was obtained for each volunteer based on approved IRB documentation prior to study initiation. Inclusion criteria for this study included asymptomatic wrists, age <60 years, and the ability to follow directions to perform wrist motions while in the MRI scanner. Exclusion criteria were contraindications to MRI (including claustrophobia) and history of wrist derangements (including trauma and arthritis).
MRI was performed on a 3-T system (VB17A Magnetom Trio, a Total Imaging Matrix System; Siemens Healthcare, Erlangen, Germany) equipped with an 8–channel radiofrequency (RF) head coil (Invivo Inc., Gainesville, FL, USA). Prior to commencing the study, initial imaging sessions were conducted in two healthy volunteers in order to develop the active-MRI protocol. The following three steps were undertaken. First, we confirmed that susceptibility artifacts, specifically banding artifacts that are unique to bSSFP sequences, partially obscured the wrist during motion. These artifacts were reliably minimized in the region of interest by using dielectric pads containing a perfluorocarbon liquid (Sat-Pad, Image Engineering Laboratories, Basking Ridge, NJ). In particular, these pads reduced magnetic field inhomogeneity across the volume of the wrist joint, and resulted in the banding artifact appearing only in the periphery of the imaged slice, outside the anatomic structures of interest. The pads eliminated the need for real-time dynamic shimming, which was not available on our MRI system. Pads were attached to the patient using medical grade tape so that they maintained contact with the wrist (
The harness has been slid outwards for better visualization of both the harness and the MR coil.
Second, a positioning harness was designed to allow for the wrist to undergo maximal ROM along the principal axes (
Third, imaging experiments were performed to determine pulse sequences and scanning parameters to provide satisfactory contrast resolution and a satisfactory trade-off between temporal and spatial resolution. Initial images acquired using the spin-echo echo-planar imaging pulse sequence had substantial spatial distortion and were unsuitable for this study, while images acquired using a single shot turbo spin echo sequence had an insufficient signal to noise ratio (SNR), due to single slice acquisition using 90° and 180° pulses and short TR in the 300–500 ms range). Gradient recalled echo (GRE) sequences provided images with significant tissue susceptibility-related signal loss, and also relatively low SNR, and were deemed unsatisfactory. The bSSFP sequence, referred to by the vendor as true fast imaging with steady state precession (true-FISP) provided very high SNR images with essentially no geometric distortion (since the sequence is spin-echo based). The true-FISP sequence was evaluated using 3-, 6-, and 10-mm section thicknesses. For each of the orthogonal slice orientations, optimal pulse sequence parameters were selected by consensus based on image review by four co-investigators (MHB, AJC, GJS, RDB), two of whom have greater than 10 years of experience in pulse sequence optimization.
The parameters for the true-FISP bSSFP sequence leading to acceptable spatial and temporal resolution converged to those providing 60 images with a 0.94 mm in plane spatial resolution, a 6 mm section thickness and 475 ms (coronal and axial orientations) or 562 ms (sagittal orientation) acquisition time per image (detailed sequence parameters are listed in
Parameter | Coronal/Axial | Sagittal |
TR (repetition time) | 3.98 ms | 4.71 ms |
TE(echo time) | 1.99 ms | 2.36 ms |
Flip angle | 47 degrees | 47 degrees |
Averages | 1 | 1 |
Field of view | 120 mm×120 mm | 120 mm×120 mm |
Section thickness | 6 mm | 6 mm |
Base resolution | 128 | 128 |
Phase resolution | 100% | 100% |
Phase partial Fourier | 4/8 | 4/8 |
Phase encode direction | R>>L/A >>P | A>>P |
Phase oversampling | 44% | 44% |
Acquisition ordering | Centric | Centric |
Asymmetric Echo | Off | Off |
Bandwidth | 781 Hz/Px | 454 Hz/Px |
Parallel imaging | Off | Off |
Scan time per image | 475 ms | 562 ms |
Final image resolution | 0.94 mm×0.94 mm | 0.94 mm×0.94 mm |
Each volunteer was placed in the “superman position” (with one arm out-stretched above the head into the RF coil) while outside the scanner bore. The wrist, surrounded by dielectric pads that limit susceptibility artifacts, was placed into a wrist harness and each volunteer was trained to perform the following four wrist maneuvers, utilizing his/her full, active ROM (absent pain), at a comfortable speed, i.e., continuously between the “start” and “stop” instruction interval of 35 s: [i] radial/ulnar deviation; [ii] the clenched fist maneuver with the wrist in the neutral position and in ulnar deviation; [iii] supination/pronation; [iv] volarflexion/dorsiflexion. The volunteer was then moved into the scanner bore. Automated high-order B0 field shimming was performed on the wrist while it was motionless in the neutral position. This was followed by each of the maneuvers, with 2D image acquisitions in three planes. Five volunteers in the prospective study had bilateral exams performed successfully on a single MRI scanner during July 2012, and five additional subjects had a single wrist imaged. Each acquisition took ∼ 35 s. The time required for detaching and re-attaching the dielectric pads was ∼ 2 min. The entire exam took ∼10 min per wrist once the patient was positioned on the system. This time included shimming with the wrist in the neutral position, localization and slice selection.
Measurements were performed by a blinded fellowship-trained musculoskeletal radiologist (RDB) and a blinded orthopedist (II) exclusively subspecializing in hand/wrist surgery, by consensus. All measurements were performed using the digital ruler and angle measurement tools available in the clinical viewing software (iSite PACS, Ver. 3.6, Philips Healthcare, Andover, MA). Specific anatomic measurements were performed in neutral and at the maximal endpoints of the ROM in each of the three planes as follows: [i] DRUJ congruity (
(a)
The relationship of the ECU tendon (arrow) to its groove as the forearm is rotated from (a) pronation, through (b) neutral to (c) supination – on axial images of the DRUJ using the active-MRI scan. In this volunteer, the ECU tendon was located within its groove in pronation, while in the neutral position, the tendon is subluxated eccentrically at the margin of the ulnar groove. In supination, the tendon is dislocated. Also visualized is the trajectory of the ulnar styloid process (white star) during the supination/pronation maneuver. Lister's tubercle (white triangle) at the dorsal aspect of the radius is shown as an anatomical reference point.
On axial images, the DRUJ congruity was assessed using the “DRUJ subluxation ratio method”
On coronal images, the SL interval was measured through the middle of the SL articulation. The distance between the cortices of the scaphoid and lunate was recorded, at the half-way point between the Gilula lines of the midcarpal and radiocarpal joints
On sagittal images, the CL and RL angles were measured through the center of the lunate. A line was drawn connecting the palmar and dorsal corners of the lunate, and a line perpendicular to it was considered the lunate axis. The angles between lunate axis and the longitudinal axes of the capitate and radius were measured to establish the CL and RL angles, respectively. Sagittal images performed through both the scaphoid and the lunate during the volarflexion/dorsiflexion movement enabled the determination of the SL angle, according to the method modified from Maizlin and Vos
Representative snapshots obtained from the active-MRI protocol are shown in
Snapshots of the coronal images of the wrist in the ulnar deviation (a), neutral (b) and radial deviation (c) positions during the continuous radial-ulnar deviation maneuver. SL interval and ulnar variance (see
Detailed measurement results are recorded in
DRUJ subluxation ratio in dorsal direction [mean (range)] |
0.14 (0.06 to 0.27) | 0.10 (0.06 to 0.2) | 0.04 (−0.06 to 0.17) | |
ECU tendon location relative to its groove | Dislocated | 1/14 (7%) | 2/14 (14%) | 6/14 (43%) |
Perched | 2/14 (14%) | 7/14 (50%) | 5/14 (36%) |
One volunteer unable to complete the pronation/supination motion protocol.
DRUJ subluxation ratio, as described by
SL gap (mm) [mean (range)] | 1.43 (1 to 2) | 1.43 (1 to 2) | 1.43 (1 to 2) |
Ulnar variance (mm) [mean (range)] | −0.93 (0 to −2) | −0.93 (0 to −2) | −0.92 (0 to −2) |
SL gap (mm) [mean (range)] | 1.36 (1 to 2) | 1.64 (1 to 3) |
RL angle (°) in dorsal direction [mean (range)] | 3.7 (0 to 17) | 29.5 (12 to 49) |
CL angle (°) in dorsal direction [mean (range)] | 4.9 (−8 to 17) | 26.7 (7 to 46) |
SL angle (°)[mean (range)] | 59 (34 to 84) | 48.5 (29 to 69) |
The wrist harness limited the ability for achieving the full range of volarflexion in a subset of volunteers therefore measurements for volarflexion are not reported.
Wrist instability is a complex phenomenon that has been studied extensively. There are potentially discordant findings during static and dynamic examinations, and there is substantial interest in diagnostic imaging performed during motion or loading
The bSSFP family of fast gradient echo pulse sequence is best known for their application in real-time dynamic cardiovascular imaging. Vendor-specific names for bSSFP pulse sequences include true Fast Imaging with Steady Precession (true-FISP, Siemens), Fast Imaging Employing Steady sTate Acquisition (FIESTA, General Electric), and balanced Fast Field Echo (bFFE, Philips). Tissue image intensity is determined by each tissue's T2/T1 ratio, and consequently fat and fluid usually appear with greater image intensity, i.e. brighter, than other tissues
Active-MRI enabled the measurement of metrics typically associated with dynamic wrist instability in all subjects, including those associated with some soft tissue components like the tendons. The measurements obtained in our study are in agreement with values reported for normal subjects as we describe below. The ECU tendon had greatest displacement from its groove in supination, followed by neutral, and was least displaced in the pronated position as has been reported earlier
A limitation of the proposed active-MRI protocol is that the spatial resolution had been traded-off for temporal resolution. We are exploring three approaches to improve upon the spatial and temporal resolution obtainable with bSSFP. First, we are exploring the use of a higher peak gradient strength, since this would allow higher spatial resolution to be achieved without compromising acquisition time or the SNR. Currently, the manufacturer imposes 27 mT/m peak gradient strength in all of its commercial pulse sequences, yet the rated and advertised maximum for orthogonal imaging is 40 mT/m. Second, an approach to increasing spatial resolution, without increasing acquisition time, is the use of parallel imaging in one or two dimensions, or the use multi-band acquisition. A multi-channel RF coil custom-fitted for the wrist would be needed for such techniques, replacing the large 8-channel head coil. Third, the current true-FISP sequence allows acquiring 2D images only at a single slice location. 3D data acquisition with true-FISP
Our study had limitations. In order to facilitate standardized, reproducible measurements in the axial, coronal and sagittal planes, we constructed a harness to constrain wrist motion in the typical orthogonal planes during active motion. Unfortunately, this custom-designed harness presented a mechanical obstruction that limited full range of volarflexion in three of our early subjects, and therefore we do not report metrics for this position. As we learned with experience, full ROM is not difficult to achieve when the wrist was positioned to allow movement freely within the coil. For this study, we used a head coil to allow space for the performance of the maneuvers during scanning. This choice led to a poorer SNR compared to dedicated wrist coils. We are exploring designs of flexible wrist coils that could be used for this purpose in the future.
Another limitation of this study is that it is only a “proof of concept” investigation carried out in a small number of asymptomatic volunteers, without any evaluation of symptomatic patients or comparison to a reference standard. Additional studies are necessary to establish the usefulness active-MRI in patients with suspected dynamic wrist instability. Further, correlative studies of active-MRI with arthroscopic or surgical findings
The focus of our investigation was on active-MRI, however, there are other imaging techniques that ultimately may prove clinically useful. Of particular recent interest is four-dimensional CT (4D-CT), in which high temporal resolution images can be acquired during wrist motion on the latest generation of CT scanners such as multirow detector CT (MDCT)
This pilot study documents the development of a protocol for active-MRI using a bSSFP pulse sequence that reliably provides real-time image acquisition during movements and consequently, enables measurements of acceptable anatomical parameters associated with dynamic wrist instability. Given the short acquisition time (<35 s), additional work is warranted to assess active-MRI that is optimized for temporal resolution as a possible supplement to conventional static images acquired during routine MRI exams that are typically optimized spatial and contrast resolution. Additional studies are also needed for correlating active-MRI metrics with arthroscopic or surgical findings of dynamic wrist instability, and with techniques such as MR arthrography and dynamic CT.
(AVI)
(AVI)
(AVI)
The authors would like the acknowledge Drs Ramsey D. Badawi, John C. Hunter, Eva Escobedo, Nancy E. Lane, Barton L. Wise, Ellen B. Gold and Lars Berglund for fruitful discussions regarding the content of this manuscript.