Myelography is a simple and economical diagnostic method. It is available in most hospitals, but involves disadvantages such as radiation exposure and adverse effects from contrast material. It enabled demonstration of the nerve root avulsions with a water soluble contrast agents. Myelography was the traditional mainstay for diagnosis of brachial plexus lesions20). The diagnostic accuracy is known to be only 37.5%, although accuracy at the level of C8 and T1 was 75%46). The diagnostic accuracy of myelography was lower at the level of C5 and C6, probably due to the narrow subarachnoid space at these levels. However, the sensitivity of myelography has been superior to which of computed tomography (CT) myelography for evaluation of the lesions at the level of C8 and T1, because shoulder artifacts could obscure a CT view at these level46,79).

Nowadays, CT myelography is the current gold standard diagnostic method for avulsion injuries33,34). Using CT myelography, ventral and dorsal nerve roots can be evaluated separately and intradural nerve defects is detected. The great advantage of CT myelography compared to myelogram is the ability to detect the partial avulsions at C5 and C6 levels, with an overall accuracy of 75%46). Carvalho et al.14) performed direct surgical examination of the intradural nerve roots through multiple hemilaminectomy and achieved an accuracy of 85% using 3 mm sections11,32,46) CT myelography at 3 to 4 weeks after injury is recommended, because the clots in the area of injury that would otherwise displace the dye might disappear at that time56). The presence of a pseudomeningocele at 3 to 4 weeks after injury highly suggests root avulsion. Traumatic meningocele is a meaningful suggestive sign of a preganglionic lesion, although it is not a specific finding. Nerve root avulsion can occur without a meningocele, and a meningocele can exist without nerve root avulsion60). Absence of roots is another important sign suggesting preganglionic lesion. Fig. 2B and C shows CT myelography representing a traumatic meningocele with nerve root avulsion.

MRI has shown less accurate imaging compared with CT myelography. In the study of Carvalho et al.14) the accuracy of MRI was only 52% and it did not provide adequate information for diagnosis in over 30% of cases. The relative advantage of MRI over CT myelography is the non-invasiveness and much higher image quality of the brachial plexus. MRI is also useful for visualization of the postganglionic brachial plexus, and for distinguishing from benign and malignant tumorous lesions81). Signal intensity changes of the spinal cord are accompanied in 20% of patients with preganglionic injuries. It provides excellent evaluation of edema in the acute phase, myelomalacia in the chronic phase, fibrosis, or neuroma formation in the post-ganglionic plexus61). Contrast-enhanced MRI is helpful to identify functionally impaired nerve roots with preserved native anatomy and avoid abortive reconstructive procedures. Abnormal enhancement of paraspinal muscles which started as early as 24 hr after injury is caused by expansion of the vascular bed and the extracellular space. It is an accurate indirect sign of root avulsion injury. Another advantage of MRI is detection of meningocele. Myelography may not demonstrate meningocele if connection between the CSF space and the meningocele is blocked by cellular proliferation. In contrast, MRI does not rely on the presence of a contrast agent within the CSF space. It can reveal fluid collection that doesn’t have connection with the CSF. Fig. 2A shows BPI with traumatic meningocele which is represented as high signal in T2-weighted MRI.

The development of MRI sequences shows even the detailed part of the plexus. The MRN is composed of T1- and T2-weighted imaging (WI), short-tau inversion recovery (STIR), 3-dimensional (3D) STIR SPACE and 3D T2 SPACE sequences. It is non-invasive and allows to identify both proximal and distal parts of the brachial plexus. Modern 3D MRN sequences are useful for identifying the extent of injury and finding the optimal treatment strategy15,75). T1-WI help to identify the anatomical structure and morphology of the plexus. Axial T2-WI using fat-suppression are useful for identifying findings related to the level of injury. Sagittal STIR images can be used to identify plexus components and their relationship to surrounding structures such as subclavian vessels. In the coronal 3D STIR SPACE sequence, the plexus appears to be bright as the background fat is homogenously suppressed. Isotropic multi-planar and curved planar reconstruction enhanced by maximum intensity projection (MIP) shows the abnormality of the injured plexus. MRN helps to classify BPI into pre-ganglionic, post-ganglionic, or mixed. Approximately 20% of pre-ganglionic injuries show altered signal intensity of the ipsilateral spinal cord due to edema, hemorrhage, and myelomalacia75). T2 hyperintense signals are associated with edema and myelomalacia, while T2 hypointense signals are evidenced by blood degradation products when hemorrhage is present. Post-ganglionic plexus injuries on MRI appear with signs that signify neuropraxia, axonotmesis or neurotmesis. When there is a stretch injury, thickened plexus components can be seen in the T2-WI. Recent modalities such as diffusion tensor imaging (DTI) continue to evolve28). Nerve tractography images may help to obtain additional information in the diagnosis and treatment of BPI.

USG of the brachial plexus has been used as guidance for brachial plexus anesthesia. Recently, it was proposed as a technique to evaluate the post-ganglionic brachial plexus47,66). The potential advantages of ultrasound include high soft tissue resolution and the ability to identify each part of the plexus as it passes in an oblique plane through the base of the neck. However, the reliability of this technique in case of postganglionic injuries should be elucidated.

In recent years, there have been attempts to utilize new imaging techniques for diagnosis and determine the effectiveness after treatment of BPI. Animal study showed ventral horn shrinkage at the level of spinal MRI by neuronal and glial changes after avulsion35). Functional MRI has also been used and demonstrated brain plasticity and compensatory mechanism by dynamic changes in resting state networks of the brain after BPI and repair11). Recent study revealed that assessment of muscle atrophy and fatty degeneration using quantitative three-point Dixon MRI sequences reflects active flexion and muscle force after BPI24).

The optimal time between injury and surgery is one of the most important factors responsible for the surgical results. Early operation might not allow sufficient time for spontaneous reinnervation; however, too late operation after 6 to 18 months of denervation can lead to failure of the motor end plate and thus failure of reinnervation78). In general, the optimal time of surgical intervention is regarded as 3 to 6 months after injury37).

The mechanism of injury determines the appropriate timing of surgical intervention. Most injuries result from vascular causes require emergency surgery. If there is complete C5 and T1 avulsion injuries, early exploration within 1 to 2 weeks is recommended44). When root avulsion is highly suspicious, early exploration and reconstruction between 3 and 6 weeks is indicated67). If there has been no functional recovery or no change in the physical examination or EMG, delayed exploration at 12 weeks after injury is recommended.

In sharp open injuries, immediate exploration and primary repair is required. In blunt open injuries, delayed exploration with tagging of the lacerated nerves might be required. By 3 to 4 weeks after injury, the nerve injury will have demarcated, and nerve repair outside the zone of injury is possible. Low-velocity gunshot wounds should be managed conservatively. On the other hand, high-velocity gunshot wounds require surgical exploration. The best operative timing for closed traction injuries is 4 to 5 months after injury, though the exact timing of surgery remains controversial.

Operative options used in BPI include nerve grafting, neurotization (nerve transfer), and other brachial plexus reconstructive techniques including the transplantation of various structures41).

Nerve grafting is a procedure which is used to make connectivity of the ruptured nerves in postganglionic injury. Nerve repair with grafting allowed the regeneration of axons down original pathways to reinnervate multiple muscles. It is theoretically the best surgical strategy; however, whether geography and structure of the regenerating fibers simulates the native sensory and motor fiber distinction remains unclear. Nerve grafting cannot be used in pre-ganglionic injuries, because a length of proximal nerve is required. In postganglionic injuries, spontaneous recovery with conservative management is usually expected if lesions in continuity are nondegenerative or fascicles are intact. However, postganglionic lesions with the nerve fiber destruction could be recovered by nerve grafting. The most commonly used donor nerve is both sural nerves; it can be as long as 30 cm5). Common examples of interpositional nerve grafting include coaptation of C5 root to suprascapular nerve or axillary nerve (for shoulder abduction), C6 to musculocutaneous nerve (for elbow flexion), and C7 to triceps or radial nerve (for elbow extension and wrist extension) (Fig. 3).

There was no relationship between the length of the nerve grafts and the surgical outcome after the brachial plexus reconstruction surgery in the study reported in 1970s63). One of the hypotheses is that the graft vascularization occurs from even the surrounding tissue, as well as the proximal nerve donor or distal nerve stump. However, Chuang et al.17) compared grafts shorter than 10 cm with those longer than 10 cm, and concluded that graft length was an important factor in functional outcome after nerve transfer for elbow flexion recovery. Patients treated with shorter grafts represented a considerably higher reinnervation rate than those with longer grafts. This finding could be explained by that the patients who required longer nerve grafts might have more extensive brachial plexus lesion and more severe proximal and/or peripheral nerve fibrosis.

Surgical technique has been demonstrated as an important factor affecting the surgical outcome6,25,64). A weak coaptation and tension at the nerve union site cause the increase of scar tissue, fibrosis and granulomatosis, which may worsen the redevelopment of growing axons64).

If the injury is preganglionic, reconstruction using nerve transfers could be helpful. The objective of the nerve transfer is not reinnervation of the brachial plexus in its original anatomy. Instead of that, lesser important nerve(donor nerve) is attached to the ruptured distal stumps; it is sacrificed to obtain a more advantageous nerve function (recipient nerve)62). Table 2 represents possible donor and recipient nerves. It represents promising surgical outcome with restoring useful limb function in the following situations; preganglionic injuries in which nerve root stumps are not available, or long-lasting injuries requiring long grafts for repair10,50,51). Theoretically, nerve transfer is useful to restore axonal input into the single denervated target. One of the advantages is the anatomic proximity of the donor and recipient nerve, and reduced operative time and the unnecessity of an intervening graft are also benefits of the nerve transfer12,70).

In general, biceps muscle function and shoulder function are regarded as to have the priority of nerve repair8,45,52). Intercostal-musculocutaneous nerve transfer is usually used to regain elbow flexion, and spinal accessory nerve-suprascapular nerve transfer is used to obtain functional recovery of the shoulder8). Various nerves which have been used as donor include the phrenic nerve, regional nerves such as C3 and C4, and other constituents of the brachial plexus for plexoplexal transfers and contralateral C7 transfers (Fig. 4, 5)30,70). Nerve transfers can also be used to innervate and power free-functioning muscle transfers.

Nerve root repair and re-implantation ventral roots into spinal cord are emerging techniques for BPI, presenting promising results13). After the exposure of the spinal cord by multiple laminectomy, a graft is inserted into a slit in the spinal cord and connected to the avulsed roots. However, further clinical study is required before using this technique as standard procedure.

Surgical results of BPI have been reported by many investigators4,9,23,34,49,72). Most patients with BPI have obtained a recovery of useful arm motor function after surgical reconstruction8). Operative repair technique was an important factor in brachial plexus reconstruction and recovering the use of an arm. In the study of Nagano et al.56), elbow flexion was recovered in more than 80% of patients under 40 year-old who underwent intercostal nerve transfer within 6 months of injury. Almost 60% of the patients obtained satisfactory outcome after brachial plexus reconstruction with nerve grafting. The outcome of neurotization or nerve transfer was also satisfactory in 40% of the patients23). Most patients were satisfied with their improved quality of life after surgical reconstruction. Only one-fifth of patients experienced permanent total disability after surgical reconstruction for BPI16).