Carpal tunnel syndrome (CTS) is the most common nerve compression disorder in the upper extremity. Therapy for carpal tunnel syndrome includes physical and occupational therapy, the use of splints and other local measures, and corticosteroid injection into the carpal tunnel. When these measures fail, open surgical release is considered the next step. Although the main disadvantage of corticosteroid injection is that symptoms are often short-lived relief and partial relief, it may not provide a permanent solution, corticosteroid injections are chosen because of lower level of invasiveness, faster recovery, and ease of the technique.

Diffusion tensor magnetic resonance imaging (DTI) reveals tissue microstructure based on random movements of water molecules. The measured diffusion-weighted images are further analyzed for parameter images that describe different characteristics of diffusion: apparent diffusion coefficient (ADC) is an absolute measure of the strength of diffusion, and fractional anisotropy (FA) describes the asymmetry of the diffusion direction due to tissue structures. Because the axonal cell membrane and the myelin sheath in nerve fibers prevent diffusion in the direction which is perpendicular to their fascicles, resulting in the isotropy of the diffusion of water molecules being lost. DTI is the only method which can give an indirect view of the microstructure of nervous tissue in addition to the pathway of the fibers.

DTI has been applied to study peripheral nerves, to demonstrate the feasibility of the method and to study nerve entrapment in carpal tunnel syndrome (CTS). The previous studies have demonstrated a decrease in FA in patients with CTS compared to healthy volunteers. The DTI parameters of the median nerve have revealed significant increase of FA and decrease in ADC with complete symptom relief 6 months after carpal tunnel release. However, Hiltunen et al. have demonstrated a significant decrease in ADC but no alter in FA in patients received open carpal tunnel release 1 year later and felt complete symptoms relief. By means of open carpal tunnel release, follow-up recordings were made at least 6 months after the operation to ensure time for post-operative tissue recovery. As a result, the investigators still do not understand the relevance between the parameters of DTI to symptoms relief in CTS patients receiving conservative treatment.

Different from carpal tunnel release, steroid injections are popular technique for CTS treatment and are believed to reduce perineural inflammation or soft tissue swelling, and may stabilize the neural membrane, thus limiting the ephaptic transmission in ischemic nerve fibers which causes symptoms. Corticosteroid injections can provide a rapid symptom relief at 2 weeks follow up. However, there is no report addressing the relation of functional change of median nerve at several anatomic locations to the symptom relief of CTS. Here the investigators monitored, by means of DTI, median nerve integrity in CTS patients before and after corticosteroid injection. This information may help to explain the hypothesis regarding the effect of corticosteroid to the median nerve, to identify which anatomic location of median nerve relevant to the symptom relief of CTS after corticosteroid injection, and be useful for the clinical follow-up of patients with nerve entrapments following conservative treatment.

Study subjects This study was approved by the institutional research ethics board. Written informed consent was prospectively obtained from all study subjects. Twelve consecutive patients (nine females, three male; mean age, 56 years; range, 38 to 76 years) suffering from bilateral CTS referred for consideration of carpal tunnel steroid injection were prospectively recruited. Inclusion criteria consisted of a clinical diagnosis of unilateral or bilateral CTS based on a standardized and validated diagnostic scale. The degree of the entrapment ranged from mild to severe according to the American Association of Electrodiagnostic Medicine ratings: mild = slowing of sensory conduction velocity (<50 m/s), moderate = slowing of sensory conduction velocity (<50 m/s) and delayed motor distal latency (>4 ms), and severe = absence of sensory response. Exclusion criteria included prior carpal tunnel release or contraindications to MRI. All patients also had the clinical diagnosis supported by electrodiagnostic investigations.

Ultrasound needle guidance:

The US-guided injection procedure was performed in a standardized manner. The one-needle two-syringe technique with US guidance was used (1) one needle is used for anesthesia, hydrodissection, and intra-carpal tunnel injection; (2) the first syringe is used to anesthetize, hydrodissect, and dilate the intra-carpal tunnel space; and (3) the second syringe is used to inject the corticosteroid therapy into the new hydrodissected space. After hydrodissection, the empty lidocaine syringe was detached from the needle while still in the carpal tunnel, and a 3-ml syringe prefilled with 10 mg triamcinolone acetonide suspension was attached to the indwelling needle, and the treatment was slowly injected into the hydrodissected neutral space. The injections were performed in a standardized fashion by a musculoskeletal radiologist (Y.C.H.) with 13 years of experience in US.

MRI protocol Magnetic resonance images were acquired at 3.0 T (Discovery MR750, GE Healthcare, Milwaukee, WI, USA) using a 8-channel wrist coil (GE Healthcare, Gainesville, FL, USA). The coil was at the center of the magnet to maximize the magnetic-field homogeneity. The subject's hand was fixed with a plastic plate and Velcro tape to the coil to restrict movements; the subjects were examined in prone position.

For DTI, the investigators recorded non-diffusion-weighted b0 image and 15 diffusion gradient directions with b = 1,200 s/mm2 from 20 axial slices of 4 mm thickness. Slices were positioned to cover the whole carpal tunnel and in part the proximal and distal nerve. The two most proximal and distal slices were excluded from the analysis because of potential bias caused by the rapid decrease in homogeneity at the edge of the coil. Thus, the total proximal-distal range examined was 80 mm. The other imaging parameters were as follows: repetition time (TR) = 10,000 ms, echo time (TE) = 101 ms, number of averages = 3 (during post-processing), field of view (FOV) = 12 cm and matrix 100 × 80 pixels. The voxel size was thus 1.46 × 1.46 ×4 mm3.

In addition to DTI, the imaging protocol comprised the following MRI sequences:

1. Axial T2-weighted fast spin echo (FSE): TR = 5,430 ms, TE = 88 ms, echo train length (ETL) = 16, FOV = 12 cm, matrix 224 × 256 pixels, phase FOV = 0.7, the same slice positions as in DTI.
2. Axial T1-weighted fast spin echo (FSE): TR = 457 ms, TE = 11~33 ms, flip angle = 110°, FOV = 12 cm, matrix 384 × 256 pixels, phase FOV = 0.7, the same slice positions as in DTI.

Data collection:

The patients' medical records were reviewed by a neurologist (F.C.Y.). Routinely, the investigators recorded clinical data (including sex, age, and side of injection, and the effect of treatment. The satisfaction scale of patients was rated using a Likert scale: 5 = greatly satisfactory; 4 = some residual symptoms and satisfaction > 50%; 3 = some residual symptoms and satisfaction = 50% ; 2 = residual symptoms and satisfaction < 50%; 1= unsatisfactory. Boston Carpal Tunnel Syndrome Questionnaire (BCTQ) is the most commonly used questionnaire for the measurement of the severity of symptoms and functional status with reproducibility, internal consistency, and validity in patients with CTS. The symptom severity subscale of BCTQ consists of 11 questions with scores from 1 point (mildest) to 5 points (most severe), and the functional status of BCTQ subscale is made up of 8 questions with scores from 1 point (no difficulty in activity) to 5 points (unable to perform the activity at all). Patients were seen in the subsequent clinical visit to determine outcomes. The investigators also recorded the decrement of BCTQ between the pre-injection and post-injection.

Data analysis Two investigators (Y.C.H., H.Y.C.) with experience in peripheral nerve imaging assessed the quality of the DTI source data. No case with visible motion artefacts was identified. Eddy current induced distortion occasionally appeared on the very proximal imaging planes, but did not affect the measurement regions. The DTI source data was post processed using the commercial post-processing workstation (GE Healthcare, ADW4.5). The investigators blinded to clinical data, side, and time point relative to steroid injection randomly performed all measurements. The same author performed a second blinded analysis of all acquired datasets after a 2-week period. Raw DTI data was initially reviewed with regard to image quality and occurrence of artefacts. After performing the GE Workspace imaging co-registration of EPI images for motion correction (through plane), DTI-based tractography (DTT) of the median nerve was performed. By defining at least two seed points, fibers passing in anterograde and retrograde direction followed a streamline algorithm and terminated if FA values were below 0.15 or if there were angle changes over 27°. The median nerve was visualized using at least three different attempts with two ROIs placed at different slice positions along the nerve (covering the entire imaged carpal tunnel). The 3D course of the nerve was tracked with DTT computed with the FACT algorithm (Fiber Assignment by Continuous Tracking).

Fractional anisotropy (FA) and the apparent diffusion coefficient (ADC) were measured bilaterally at each time point. FA and ADC were calculated from free-hand regions of interest (ROIs) placed in the center of the median nerve at three levels: at the level of the distal radioulnar joint (the most proximal slice where by the distal radio ulnar joint was visualized) and at the level of the flexor retinaculum (level of the pisiform). To avoid partial volume artifacts, care was taken to draw the ROIs slightly smaller than the cross-sectional area of the median nerve. The size of the ROIs depended on the cross-sectional area of the nerve and all ROIs were placed on one slice only. Anatomic reference images were used to identify the median nerve. The measurement of each parameter was repeated three times and the mean was calculated. Representative images are presented.

Regarding the assessment of DTT, the investigators determined the nerve fiber tractography by morphology. Four different nerve trajectory patterns were defined by Breitenseher et al.: (1) If there was no alteration of the peripheral nerve trajectories, the pattern was rated "continuous". (2) If the trajectories were continuous, but irregularly organized, peripheral nerve tractography result was rated "deranged" .(3) In cases where only some of the ulnar nerve trajectories were interrupted in one peripheral nerve segment, the finding was defined as "partially discontinuous ". (4) If there was a complete discontinuity of all trajectories and a "gap" on at least one plane, tractography was classified as "completely discontinuous". DTT images were evaluated by two readers (Y.C.H., G.S.H.) in consensus.

Inclusion Criteria:

• clinical diagnosis of unilateral or bilateral CTS based on a standardized and validated diagnostic scale

Exclusion Criteria:

• prior carpal tunnel release or contraindications to MRI

1. All patients had the clinical diagnosis supported by electrodiagnostic investigations.
2. All patients underwent US-guided corticosteroid injection of median nerve and received MR study before and after the injection.