MR imaging of the brachial plexus

In this retrospective study we describe the MR imaging findings in 230 consecutive patients with suspected pathology in or near the brachial plexus. These patients were studied from 1991 through to 1996. Chapter 2 describes the anatomy and the MR imaging techniques. As the anatomy of the brachial plexus and the related structures is quite complicated, we eventually use as protocol of choice a 3D volume acquisition for the best understanding of this complex anatomy. The advantages of this 3D volume acquisition are, besides the use of thin overlapping slices, the MPR and cine-display viewing mode possibilities. The use of thin slices (2 mm) provides excellent anatomical detail. The following anatomic details can be discerned: the individual ventral rami of the nerve roots, the three trunks, the three cords and the stellate ganglion. The overlapped images can be reconstructed in any plane with the same image quality, and the use of the cine-display viewing mode affords a better insight into the continuity of the nerves and vessels. In our experience, the use of a 3D volume acquisition markedly improves the understanding of the normal anatomy of the brachial plexus on MR imaging and can possibly better delineate the pathology involving the brachial plexus. Chapter 3 mainly deals with the tumors we have found in or near the brachial plexus. In this chapter we also describe a group of patients alleged to have a tumor, but where no tumor was found. The first group consists of 66 patients where a tumor in or near the brachial plexus was found with MR imaging. We found 10 neurogenic tumors (five schwannomas, one neurofibroma, two malignant schwannomas and two without a histological diagnosis). These tumors show characteristic MR imaging findings: a low signal intensity on the T1-weighted images, an increased signal intensity on the proton-density images, a high signal intensity on the T2-weighted images, enhancement after administration of gadolinium-DTPA, a fusiform growth, a sharply defined edge, and the involved nerve can often be found entering and leaving the tumor. These specific imaging characteristics applied to nine of these 10 tumors. Besides these 10 Chapter 7 Summary and Conclusions?102 Chapter 7 neurogenic tumors, we describe 56 non-neurogenic tumors: lung tumor (n=24), metastasis of breast carcinoma (n=9), metastasis of other tumors (n=8), B-cell non-Hodgkins lymphoma (n=2), leiomyosarcoma (n=1), liposarcoma (n=1), chondrosarcoma (n=2), synoviosarcoma (n=1), aggressive fibromatosis (n=2), meningocele (n=1), lipoma (n=3), a hematoma in a neck cyst (n=1), and one tumor of unknown origin. In the preoperative evaluation of tumors near the brachial plexus, the 3D volume acquisition provides the radiologists and the surgeons with better insight into the precise extension of the tumor. MR imaging could delineate the extension of these tumors well, and determine whether or not there was brachial plexus involvement. The second group consists of 70 patients where no tumor was detected with MR imaging. Three patients had an infection, 17 patients were imaged after radiation therapy and had no signs of tumor recurrence, and 50 patients had normal MR imaging findings. Chapter 4 discusses the value of MR imaging in patients with brachial plexopathy after trauma and in patients with thoracic outlet syndromes. We found 31 abnormal MR imaging investigations in patients with a history of trauma. Ten of these patients presented with a flail arm after a severe accident, in most cases a motorcycle accident. In five of these patients an additional 3D-TSE sequence of the C-spine was done in order to visualize traumatic nerve root avulsions and traumatic meningoceles. Abnormalities we found in this group were thickening of the brachial plexus, hematoma, and a clavicle fracture with compression of the brachial plexus. We were not able to visualize a rupture of the brachial plexus, nor visualize all the nerve root avulsions. Abnormalities we found in the remaining 21 patients included clavicle fractures with suspected brachial plexus compression (n=14), traumatic meningoceles in patients with known nerve root avulsions to exclude neuroma formation (n=3), a stab wound (n=1), a coracoid process fracture (n=1), a battered child (n=1) and a shoulder luxation (n=1). In 18 patients with a history of trauma MR imaging was normal. Two of these patients were operated upon and appeared to have a rupture of a part of the brachial plexus. We conclude that MR imaging is not very good at predicting the surgical findings in cases of brachial plexus rupture or nerve root avulsions. MR imaging can be helpful in demonstrating brachial plexus compression by a hematoma or a clavicle fracture with callus formation. In this chapter we also describe a group of 23 patients with a wide variety of symptoms, which could be due to a thoracic outlet syndrome. In three patients a cervical rib was present, two patients had had a cervical rib removal previously. In one patient, who had the typical true neurogenic thoracic outlet syndrome, a slight angulation of the ventral ramus of root C8 was seen, which appeared to be due to a fibrous band at surgery. Most patients did not show any abnormalities with MR imaging. We were?Summary and Conclusions 103 not able to demonstrate a fibrous band with MR imaging. We conclude that MR imaging in patients with a thoracic outlet syndrome is not very useful, except for serving to exclude other structural abnormalities. Chapter 5 describes the MR imaging appearance of radiation-induced brachial plexopathy. MR imaging was performed in two patients with the clinical diagnosis of radiation-induced brachial plexopathy and in one with surgically proven radiation fibrosis of the brachial plexus. Three patients who had had radiation therapy to the axilla and supraclavicular region (two with breast carcinoma and one with Hodgkins lymphoma) presented with symptoms of the arm and hand. To exclude metastases or tumor recurrence MR imaging was performed. In one patient, fibrosis showing low signal intensity was found, while in two patients high signal intensity fibrosis surrounding the brachial plexus was found on the T2-weighted images. In one case gadolinium-DTPA enhancement of the fibrosis 21 years after radiation therapy was seen. We conclude that radiation-induced brachial plexopathy can have different MR imaging appearances. We found that radiation fibrosis can have both low and high signal intensities on T2-weighted images, and that fibrosis can enhance even decades after radiation therapy. In Chapter 6 we studied whether MR imaging of the brachial plexus is useful to distinguish multifocal motor neuropathy (MMN) from lower motor neuron disease (LMND) and whether abnormalities resemble those of chronic inflammatory demyelinating polyradiculoneuropathy (CIDP). MMN is a potentially treatable pure motor neuropathy which clinically resembles LMND. Both diseases are characterized by progressive asymmetric weakness of the limbs and muscular atrophy without sensory symptoms. CIDP is a symmetric polyneuropathy which affects both motor and sensory fibers. Both CIDP and MMN are probably immune-mediated neuropathies, as autoantibodies to peripheral nerve myelin have been found. We compared MR imaging scans of the brachial plexus from nine patients with MMN with scans from five patients with CIDP, eight patients with LMND, and 174 controls. In two patients with MMN, and in three patients with CIDP, the MR imaging scans showed an increased signal intensity on the T2-weighted images of the brachial plexus. Two other patients with MMN demonstrated a more focal increased signal intensity on the T2-weighted images, in one patient only in the axilla, and in one patient in the axilla and in the ventral rami of the roots. MR imaging of the brachial plexus of eight patients with LMND was normal. The distribution of the MR imaging abnormalities corresponded with the distribution of symptoms of the patients: asymmetrical in MMN and symmetrical in CIDP. These findings demonstrate that MR imaging abnormalities of the brachial plexus in patients with MMN resemble those seen in CIDP and may be useful to distinguish MMN from LMND.?104 Chapter