Imaging of the Carpal Tunnel and Median Nerve

Fig. 8.1

Carpal tunnel view radiograph —normal anatomy; P pisiform, H hamate, C capitate, S scaphoid, T trapezium

Fig. 8.2

Carpal tunnel view radiograph —fracture at the base of the hook of the hamate (yellow arrow)

Computed tomography (CT) has increased sensitivity for detecting soft tissue mineralization and fracture. The contours of mass lesions are more evident by cross-sectional imaging, but soft tissue contrast remains limited by the modality. The thin slice acquisition of multidetector array CT acquires isotropic data sets which allow multiplanar reformats to be created. These reformats increase the sensitivity for fracture (Fig. 8.3) and for space-occupying lesions within the carpal tunnel.

Fig. 8.3

Computed tomography axial image—fracture at the base of the hook of the hamate (yellow arrow)

Magnetic Resonance Imaging

MRI depicts the carpal tunnel and potential pathology with high spatial resolution and soft tissue contrast. While imaging sequence protocols vary between institutions, the axial plane is frequently performed and highly diagnostic. Generally, a fluid-sensitive sequence such as a fast spin-echo proton density (PD) or T2-weighted sequence is utilized, depicting edema patterns as a bright or hyperintense signal. Because the nerve is small, edema signal can often be obscured by the surrounding fat; such sequences are thus often performed with fat suppression, making the signal more conspicuous. Commonly used methods of fat suppression are frequency-selective saturation of the fat resonance or a short tau recovery (STIR) [9]. Nonfat-saturated T1-weighted or fast spin-echo PD sequences are often included for anatomic detail of the tendons, median nerve, bone contour, and flexor retinaculum. While there is no particular standard matrix for carpal tunnel imaging, sequences on routine wrist evaluations can range from a matrix of 256 × 256 to as high as 512 × 256. Regardless of the matrix, the images must be able to display the overall morphology of the nerve and its cross-sectional area, both of which are of particular relevance in CTS imaging. The field of view is usually set at 8 cm, and slice thickness is acquired between 2 and 3 mm [7, 9] (Figs. 8.4, 8.5, 8.6, and 8.7).

Fig. 8.4

Axial MRI proton density (a) and T2 fat saturation (b)—normal anatomy at the level of the DRUJ; R radius, U ulna, Mn median nerve, Un ulnar nerve, Fdp flexor digitorum profundus, Fds flexor digitorum superficialis, Fpl flexor pollicis longus, Fcr flexor carpi radialis, Fcu flexor carpi ulnaris

Fig. 8.5

Axial MRI proton density (a) and T2 fat saturation (b)—normal anatomy at the level of the pisiform; S scaphoid, C capitate, H hamate, T triquetrum, P pisiform, Mn median nerve, Un ulnar nerve, Fdp flexor digitorum profundus, Fds flexor digitorum superficialis, Fpl flexor pollicis longus, Fcr flexor carpi radialis, Fcu flexor carpi ulnaris, Fr flexor retinaculum

Fig. 8.6

Axial MRI proton density (a) and T2 fat saturation (b)—normal anatomy at the level of the hamate; Tm trapezium, Td trapezoid, C capitate, H hamate, Mn median nerve, Un ulnar nerve, Fdp flexor digitorum profundus, Fds flexor digitorum superficialis, Fpl flexor pollicis longus, Fcr flexor carpi radialis, Fr flexor retinaculum

Fig. 8.7

Axial MRI proton density (a) and T2 fat saturation (b) at the level of the pisiform—median nerve enlargement and high signal with variant vascular anatomy; persistent median artery (yellow arrow)

The median nerve contours seen on MRI are defined by its outermost connective tissue sheath, called the epineurium. A surrounding rim of epineural fat may be visible, appearing as a high-signal rim on PD- or T1-weighted images and low signal on fat-suppressed imaging. High-resolution imaging may also depict the individual nerve fascicles, which should demonstrate intermediate signal slightly higher than that of the normal muscle. The loss of the normal internal fascicular architecture of the nerve coupled with high signal within the nerve is an abnormal finding, suggesting edema or neuritis. Segmental or diffuse enlargement of the nerve is also commonly seen with neuritis. When the median nerve demonstrates focal or fusiform thickening, one should assess for a possible neoplasm or posttraumatic neuroma [9, 10].

On fast spin-echo PD- and T2-weighted sequences , the normal muscle is intermediate-signal intensity. Acutely denervated, muscle tissue will demonstrate abnormal MRI signal that precedes actual muscle volume loss and atrophy. This abnormal signal or prolonged T2 relaxation time is hyperintense (bright) on PD- and T2-weighted sequences. This change is thought to be due to fluid movement from the intracellular space into the extracellular space as well as capsular engorgement and increased muscle blood volume [7, 11, 12]. The chronic findings of muscle atrophy and fat replacement are also well depicted on fast spin-echo PD- or even better on T1-weighted sequences, as fat is bright or hyperintense on these sequences.

On a conventional wrist MRI, there are four general categories of imaging findings in carpal tunnel syndrome: increased median nerve size, median nerve flattening, median nerve signal change, and flexor retinaculum bowing. Each of these criteria has varied sensitivities and specificities for CTS when evaluated individually. However, when combined as an overall impression, MRI has a sensitivity as high as 96% but a low specificity (33%) [13].

Median Nerve Enlargement

Patients with carpal tunnel syndrome can often have median nerve enlargement (Fig. 8.8). This is most objectively assessed by measuring the cross-sectional area of the nerve on an axial image. In the literature, three levels within the wrist have been most commonly used as landmarks for measurement: (1) distal radioulnar joint (DRUJ) , (2) pisiform, and (3) hamate. Absolute cutoff values for the normal median cross-sectional area have varied between studies, some with less than desirable reproducibility [14]. A more useful and reproducible reference standard for CTS is the ratio of the surface area of the median nerve at the level of the pisiform relative to the median nerve surface area at the DRUJ. Quantitative analysis has shown that in normal asymptomatic subjects, the mean pisiform/DRUJ ratio is 1.1. In patients with CTS, the mean pisiform/DRUJ ratio is as high as 2.4 [15]. While the mean cross-sectional area ratio can also be taken at the level of the hamate, the difference between normal and CTS is not as large. Careful measurements of the median nerve cross-sectional area can be time-consuming, and an effective method for determining median nerve enlargement is finding a nerve that is 2–3 times larger at the pisiform than at the DRUJ.

Fig. 8.8

Axial MRI T2 fat saturation —median nerve enlargement with high internal signal (yellow arrow) and bowing of the flexor retinaculum

Median Nerve Flattening

The median nerve normally undergoes a small degree of flattening within the carpal tunnel; however, excessive nerve flattening can indicate CTS. As this observation can be subjective, a flattening ratio can be a more effective means of quantitatively determining the severity. The ratio is made by measuring the major and minor axes of the nerve at the level of the DRUJ as well as within the carpal tunnel. In patients with CTS, the mean flattening ratio at the DRUJ is 1.8 at the distal radius, but increases up to 3.8 at the level of the hamate [12, 15]. Thus, a median nerve that is 3–4 times wide as it is thick is associated with CTS (Fig. 8.9).

Fig. 8.9

Axial MRI PD —flattening of the median nerve (yellow arrow) at the level of the pisiform

Flexor Retinaculum Bowing

Normally, the flexor retinaculum is flat or convex at the level of the hamate, where thickness is the greatest. In CTS, a bowed flexor retinaculum implies median nerve compression. The degree of bowing can be quantified by dividing the distance of palmar displacement of the retinaculum by the distance between the hook of the hamate and the tubercle of the trapezium. In normal patients, the ratio ranges from 0 to 0.15. In patients with carpal tunnel syndrome, this ratio is between 0.14 and 0.26 [15] (Fig. 8.10).

Fig. 8.10

Flexor retinacular bowing ratio —distance of palmar displacement (blue arrow) divided by the distance between the hamate hook and trapezium (yellow arrow)

Hyperintense Median Nerve Signal

Normally, the median nerve will have an internal signal similar to or slightly higher than that of muscle tissue on fast spin-echo PD- and T2-weighted sequences . The signal is generally uniform throughout the length of the nerve within the field of view. When the median nerve is bright or high signal on these sequences, the finding is highly sensitive for CTS (88%). This finding however should be treated with caution, as it is unfortunately low in specificity (39%) [13]. The hypothesized pathophysiology for abnormal signal of the nerve may be related to localized edema signal or fluid accumulation within the endoneural spaces [9]. On routine MRI sequencing of the wrist, this finding is made by visual inspection; no reliable quantitative measuring system is accepted. Despite the potential for subjectivity, studies have shown at least substantial inter-reader agreement (weighted kappa = 0.71) when judging median nerve signal abnormality [13] (Fig. 8.11).

Fig. 8.11

Axial MRI proton density (a) and T2 fat saturation (b) at the level of the pisiform—median nerve enlargement and high-signal intensity (yellow arrow)

Additional MRI Findings

Thenar muscle denervation effect on MRI is indicated by high signal on either PD- or T2-weighted fat-suppressed imaging. This finding is not particularly sensitive for CTS (10%) but is highly specific (96%) (Fig. 8.12) and considered a late-stage finding, correlating with clinically severe CTS and high-grade denervation by EDT [10].

Fig. 8.12

Axial MRI proton density (a) and T2 fat saturation (b)—muscle atrophy and denervation high signal within the thenar muscles (opponens pollicis and abductor pollicis brevis, yellow arrow). The median nerve is flattened within the carpal tunnel (arrowhead)

When mechanical compression is the cause of CTS, there are many pathologic lesions that can be discovered by MRI. Flexor tenosynovitis due to infectious or inflammatory arthropathy (such as rheumatoid arthritis or gout) or nonspecific repetitive motion can be indicated by the high fluid signal and distension outlining the flexor tendons on either PD- or T2-weighted sequences (Fig. 8.13). Space-occupying lesions such as ganglion cysts or neoplastic lesions can also be implicated in nerve compression. If a neurogenic tumor or sarcomatous lesion is suspected, MRI protocols that are further optimized for soft tissue imaging may be needed for characterization. This includes the use of pre- and post-intravenous contrast T1-weighted fat-suppressed sequences which can further grade the lesion margins and its overall vascularity and enhancement [16] (Fig. 8.14).

Fig. 8.13

Axial MRI T1 (a) and T2 fat saturation (b)—tenosynovitis of the flexor tendons in rheumatoid arthritis

Fig. 8.14

Axial MRI T1 (a), T2 fat saturation (b), and T1 fat saturation post IV contrast (c)—synovial sarcoma within the carpal tunnel; T2 hyperintense, enhancing mass (arrowhead) that is displacing and flattening the median nerve (yellow arrow)

Postsurgical MR Imaging

MRI can be helpful in evaluating the causes of recurrent CTS after surgical release. Similar to pretreatment imaging, the findings have varied sensitivities and specificities and should be combined with EDT findings. In addition to the routine axial wrist sequences, pre- and post-intravenous contrast T1-weighted fat-suppressed sequences are recommended to evaluate the postoperative changes. The aforementioned routine MRI criteria of CTS remain helpful in the assessment of the postoperative wrist. Of these findings, an increased flattening ratio is the most statistically significant difference between EDT-confirmed recurrent CTS and postoperative controls [17].

Additional statistically significant changes in the postoperative patient with recurrent CTS include the presence of perineural fibrosis, median nerve enhancement, and insufficient release of the tunnel. Fibrosis appears as extensive low-signal intensity with an ill-defined nerve margin on either nonfat-suppressed PD- or T1-weighted fast spin-echo sequences . Fibrosis is 60% sensitive, but 83% specific for recurrent CTS. Median nerve enhancement after intravenous contrast administration is considered present if the signal is higher than the level of enhancement of the thenar muscle. While not highly sensitive (40%), this is highly specific (92%) for recurrent CTS [17].

Only gold members can continue reading. Log In or Register to continue