Initial experience with 3D isotropic high-resolution 3 T MR arthrography of the wrist

The wrist, given its relative small size compared to other joints and complex arrangement of ligaments, requires high-resolution, high signal-to-noise, and small field of view MR imaging for accurate assessment of anatomy and pathology. When attention is required at the ulnar side of the wrist for evaluation of triangular fibrocartilage complex (TFCC) or capsular injury, direct MR arthrography has become an increasingly utilized modality, allowing for improved visualization of the closely grouped ligamentous structures following distention of the joint with injected contrast [1–3]. In addition to joint distension, intra-articular contrast allows for visualization of contrast leakage/extravasation and has been shown to be particularly useful in evaluating partial ligamentous tears including of the TFCC [4]. With increasingly widespread use of 3 T MR in musculoskeletal imaging, the wrist included, 3 T MR has been shown to be sensitive and specific for wrist ligament tears in addition to improved visualization of normal anatomy [5, 6]. The addition of 3D isotropic imaging has also shown utility in musculoskeletal imaging, including of the knee [7, 8]. More recent literature involving 3 T 3D isotropic imaging of the wrist in healthy volunteers has been carried out, highlighting the advantages of multi-planar reconstructions and additional post-processing capabilities afforded by isotropic imaging which prove useful in evaluating the intricate anatomy of the wrist [9].

Our study was performed to evaluate image quality of 3 T MR arthrograms which include 3D isotropic proton density weighted fat suppressed sequence (PDFS). Comparison was carried out with standard 1.5 T MR arthrogram studies and additional quantitative analysis of relative signal intensity and relative contrast of the individual 3 T sequences was also performed.

All studies of both 1.5 T and 3 T groups were obtained following injection of 3–4 ml of dilute gadolinium contrast mixture (15 mL sterile saline, 5 mL iodinated contrast, and 0.1 mL of gadolinium contrast) into the radiocarpal joint under fluoroscopy using standard department protocol. The patients were imaged within 30 min of injection.

Quantitative analysis

Quantitative analysis was performed for the 11 3 T studies. For each of the 3 T wrist sequences (2D T1FS, 2D PDFS, isotropic 3D PDFS), ROIs for each patient were drawn and relative signal intensity (SI) and relative contrast were measured using. ROI’s were identical in size and placed in identical positions on matching sections. Fluid was measured from the space adjacent to TFCC, prestyloid recess, or intercarpal space (Fig. 1). The mean and standard deviation (SD) of SI were determined for the disc of the TFCC, fluid, cartilage, and bone marrow. Then, relative SI and relative contrast were used for direct comparison of image quality between the 2D FSE and 3D isotropic FSE MR images both with and without fat suppression, because all sequences in this study were obtained with a parallel imaging technique. Relative SI of each structure was calculated as SI/SD, and relative contrast of structure A (a) to structure B (b) was calculated as (SI_a-SI_b) / (SD_a ² + SD_b ²)^1/2 [9, 11].

Evaluation of the ulnar side of the wrist remains a challenge in everyday medical practice, particularly due to the complex nature of triangular fibrocartilage complex. This is complicated by discussion involving anatomic nature of ulnar wrist structures including ulnomeniscal homologue [13–17] and ulnar collateral ligament complex [18–21], both of which have been debated in the literature. Despite the nuances in the anatomy, direct MR arthrography with injected intra-articular gadolinium contrast has become the chosen modality for evaluation of the ulnar side of the wrist, with several prior studies demonstrating improved diagnostic ability compared routine 2D MRI [22–27] and previously supplanting conventional arthrography due to conventional arthrography’s low specificity and accuracy [28].

Given the shape and complex geometry of the TFCC it had been proposed as a candidate for 3D imaging [29]. Additional studies using 3D images of the wrist have been done including a study in healthy volunteers [9], and a study using 3D T1 sequence examining scapholunate, lunotriquietral, and TFCC tears [30]. These studies also showed improved diagnostic potential with 3D imaging. A consensus among these studies as a factor in improved diagnostic potential using 3D isotropic sequence include ability to construct multi-planar reformatted images along an arbitrary obliquity to follow the TFCC and additional ulnar sided structures (or any ligament). Another reason for improved diagnostic potential is the smaller slice thickness of 3D isotropic sequence (0.35 mm versus 2–3 mm) with subsequently decreased partial volume averaging. With smaller slices there is potential for improved identification of subtle abnormalities of the TFCC and capsular complex, which may not bee seen on 2D sequences. Image blur was not a major factor in image quality of the 3 T images, although it has been implicated in the past as a disadvantage to 3 T imaging using long echo train length FSE sequences [31]. Improvement in the imaging of the TFCC is important for many reasons including defining MR anatomy, patient selection for surgical versus non surgical treatment, pre-operative planning, and ultimately leading to improved patient outcomes.

In our study we evaluated 3 T MR wrist arthrogram studies with attention to ulnar wrist structures including TFCC and ulnar capsular complex and compared the image quality of the 3 T sequences to one another and to that of 1.5 T MR wrist arthrogram, along with quantitative analysis amongst the 3 T images using relative SI and relative contrast. The results showed 3 T image quality to be significantly superior to that of 1.5 T and that the 3D isotropic PDFS sequence had the highest scores for image quality among the 3 T sequences. Inter-rater reliability was high for 3 T 3D sequence. In addition, as previous studies, the capability of multi-planar reformations in any obliquity planes proved usefulness for 3D isotropic MR arthrograms at 3 T for improved visualization of TFCC and capsular injuries, especially at the ulnar styloid attachment.

As part of the qualitative grading we looked at the presence of contrast leakage outside the joint capsule at the ulnar wrist, and graded it on confidence of leak and whether it was major or minor, because contrast leakage is frequently seen and is a specific finding with MR arthrogram. The clinical significance of leak at the ulnar side of the wrist, and furthermore whether major or minor plays a role, remains to be seen and is a source of further investigation. Prior studies characterizing leakage on conventional arthrography identified leakage at the ulnar capsular complex as pathologic, supported by improvement in symptoms following repair [32]. However, it has also been found that TFC “communicating defects” on conventional arthrography happen frequently in asymptomatic patients and have also been shown in a significant percentage of cadaveric studies and may in fact represent changes related to degeneration [33–36]. This discussion of contrast extravasation at the ulnar side of the wrist is similar to that found with MR arthrogram of the shoulder, where capsular leakage is also not uncommon, having both pathologic causes including a number of labral tears and conditions such as adhesive capsulitis, but also identified through the subscapular recess along subscapularis and adjacent to axillary recess, both unrelated to needle path and without adjacent related pathologic findings [37]. Regardless, future studies with focus on clinical outcomes will be needed to clarify the significance of contrast material outside the capsule at the ulnar side of the wrist on MR arthrography.

Limitations of our study include the relatively small size, with 11 3 T studies and 10 of which included 3D isotropic sequence. Additionally, the 1.5 T studies were collected over a number of years and the imaging protocols had changed, leaving comparison with similar but not identical imaging parameters for both 3 T and 1.5 T studies. Another limitation is related to the relative longer acquisition time of 3D sequence, which adds approximately 5–6 min to total scan time. The ability to quickly and seamlessly create MPR images is also a limitation, usually requiring additional functional PACS capabilities or additional software, which may not be widely available. The fact that this 3D sequence is PD weighted rather than T1, differentiation between subtle edema and minor contrast leakage occasionally can be difficult. This may suggest 3D isotropic PDFS could potentially replace 2D PDFS, but not 2D T1FS. Finally, selection bias also plays a role given that the majority of patient’s had chronic ulnar wrist pain and had been evaluated by a hand surgeon prior to imaging and initially referred for MR arthrography.

The results of our initial study investigating the image quality of 3 T MR arthrogram with 3D isotropic PDFS sequence adds to some existing evidence that 3D isotropic PDFS sequences may be useful when imaging the wrist at 3 T. Given the inherent advantages of direct MR arthrography combined with isotropic high-resolution images, this technique shows promise for the evaluation of ulnar wrist pathology including TFCC and capsular injuries.