Clinical manifestations

Presentation and course

Patients with brachial plexus metastases typically present with pain and/or weakness. The pain is relentlessly progressive, followed after a variable interval by weakness and sensory loss in a pattern reflecting involvement of more than one nerve root (31; 33; 24; 30). Lower trunk invasion is more commonly seen than upper trunk invasion, contrasting from the upper trunk predominant damage seen in radiation-induced plexus injuries. For lung malignancies, this is due to the lymphatic drainage of apical lung lesions being in close proximity to the lower trunk of the brachial plexus (10). The pain is described as an aching pain radiating down the arm, mostly in the medial aspect of the arm and forearm (28).

As with metastasis to the brachial plexus, neuropathic pain is common with lumbosacral plexus metastasis. At least 75% of patients with lumbosacral plexus metastases present with pain, usually unilateral and affecting the low back, hip, and thigh (44; 24). Nearly all patients eventually develop local or radicular pain radiating into the buttock and lower extremity. The pain is usually unrelenting despite analgesics and characteristically worse at night. Dysesthesia is considered rare. Most patients subsequently develop weakness and sensory symptoms after a lag period of weeks to months. Sensory deficits are patchy, with progression to more severe lower motor neuron pattern weakness becoming more evident. Bladder dysfunction is uncommon unless the coccygeal plexus is involved. Neurologic examination may be normal early in the disease process, when pain and paresthesias are the only symptoms. Inspection of the supraclavicular and infraclavicular fossa on the involved side may reveal fullness. Palpation of the supraclavicular fossa or axilla may reveal adenopathy, suggestive of tumor. Later in the disease course, weakness, sensory loss, and diminished reflexes are often observed in a pattern reflecting involvement of more than one nerve root.

Some patients have selective tumor involvement at the L1 level affecting the ilioinguinal, iliohypogastric, or genitofemoral nerves while sparing motor trunks of the plexus. Patients report pain or paresthesias in the groin or lower abdominal quadrant with no motor or sensory symptoms in the legs (24). Patients may develop a "hot dry foot" from involvement of the sympathetic fibers. Another group of patients, mainly with rectal tumors, presented with perineal pain, sensory loss, and early bowel or bladder sphincter involvement, reflecting selective tumor involvement of the coccygeal plexus (38).

Prognosis and complications

The pain and neurologic deficits caused by brachial or lumbosacral plexus metastases represent a debilitating complication. Response to medical therapy depends in part on the duration and severity of symptoms. Patients with relatively recent onset of pain and minimal or no neurologic deficits have a higher likelihood of significant improvement with treatment compared with patients who have severe, longstanding pain and sensory or motor deficits. In the latter group the neuropathic or causalgic pain may prove intractable.

Clinical vignette

A 40-year-old, right-handed woman underwent left modified radical mastectomy and axillary lymph node dissection for breast carcinoma, followed by six cycles of systemic chemotherapy. Two years later, she developed pain around the left axilla, scapula, and upper arm without numbness or weakness in the setting of recurrent tumor in the left anterior chest wall. She received 5040 cGy radiation to the anterior chest wall, followed by resection of the mass. Her pain subsequently improved. Approximately 2 months after resection, she developed pain, numbness, and an “electric burning” sensation over the first two digits of her left hand. Over the next month the sensations spread to involve the entire hand with subsequent decreased grip strength, despite two cycles of paclitaxel chemotherapy. She obtained mild pain relief from oxycodone and amitriptyline.

Neurologic examination 1 month later showed normal cranial nerve function, with no Horner syndrome. Muscle bulk and tone in all extremities were normal, with no fasciculations. Strength was 5/5 in all muscle groups except in the left upper extremity: deltoid 5/5, biceps 5-/5, brachioradialis 5-/5, pronator 4+/5, supinator 4+/5, triceps 5-/5, grip 5/5, abductor pollicis brevis and opponens pollicis 4+/5, other intrinsic hand muscles 5/5. Muscle stretch reflexes were 2+/4 and symmetric except for diminished left biceps reflex. Sensory and cerebellar testing was normal. MRI scan showed an extrapleural tumor mass involving the left lung apex, extending to encompass the subclavian vessels and the brachial plexus.

The patient received another course of radiation with 5040 cGy to the region of the new tumor mass; there was some overlap with previous radiation ports. After completion of radiation treatment, the pain in the upper arm partially improved with less numbness and dysesthesias of the right thumb; however, there was no change in the neurologic examination. Repeat MRI scan 2 months after radiation showed significant tumor shrinkage, although there was residual abnormal enhancement along the brachial plexus. Six months later, she developed multiple brain metastases and a malignant pleural effusion.

Biological basis

Etiology and pathogenesis

Tumors usually spread to the brachial plexus by direct extension from axillary or supraclavicular lymph nodes, whereas apical lung carcinomas (Pancoast tumors) spread directly to the adjacent plexus (17). Rarely, breast carcinoma or other tumors spread hematogenously into the nerve trunks in the absence of tumor in the adjacent soft tissues or lymph nodes (37). Tumor spread is thought to primarily occur through lymph nodes around the brachial plexus or lumbosacral plexus (10). Brachial plexus metastases may occur as an isolated site of tumor spread, and some patients have no evidence of metastatic tumor for several years or more after the onset of metastatic plexopathy (31).

In patients with metastases to the lumbosacral plexus, the tumors invade the plexus by direct extension from a primary pelvic tumor, from metastases to nearby lymph nodes, or from bone metastases in the spine or pelvis (38). Prostate carcinoma may spread to the lumbosacral plexus along the perineural spaces (06). A complex cell-signaling role appears to occur between nerve and tumor cells and could lead to further tumor cell growth; for instance, in prostate cancer, prior observations have shown reduced tumor cell apoptosis and increased proliferation and migration through the expression of certain cell adhesion molecules (06).

The brachial or lumbosacral plexus may be the site of diffuse or multifocal infiltration by malignant lymphocytes ("neurolymphomatosis") (04; 42; 20; 23; 27).

Neoplastic invasion from lymphoma occurs more commonly with large B-cell non-Hodgkin lymphoma but can also occur with other B-cell and T-cell lymphomas (11). Sometimes it can occur exclusively in the lumbosacral plexus without evidence of systemic disease (11). Neurologic signs and symptoms are caused by direct tumor cell invasion of the plexus or by infiltration of tumor cells along the connective tissue and epineurium of nerve trunks.

Epidemiology

Carcinomas of the breast and lung together account for an overwhelming majority of brachial plexus metastases. Symptomatic brachial plexus metastases have been estimated to occur in approximately 4% of lung cancer and 2% of breast cancer patients. Primary malignancies that infiltrate the brachial plexus less frequently include melanoma, sarcomas, and gastrointestinal carcinomas (05). There are scattered reports of brachial plexus metastases from a variety of other tumors, such as head and neck cancers (39).

Metastases to the lumbosacral and coccygeal plexi most often arise from several types of pelvic malignancies, including colorectal, prostate, bladder, rectal, and gynecological cancers due to the contiguity of the plexus to these pelvic organs. Prostate carcinoma may rarely spread along the lumbosacral plexus without other evident tumor in the pelvis (06). Primary plexus involvement without a large tumor burden in the pelvis is thought to occur due to perineural spread along the autonomic nerves in the pelvis (06).

Differential diagnosis

The most common dilemma in the differential diagnosis is distinguishing radiation-induced brachial or lumbosacral plexopathy from tumor metastasis to the plexus. Breast carcinoma is the tumor most often associated with delayed radiation brachial plexopathy, accounting for 40% to 75% of patients in the literature, followed by bronchogenic carcinoma, head and neck carcinoma, and lymphoma (31). The risks of delayed radiation toxicity likely correlate, at least somewhat, with the histology-specific likelihood of long-term survival with metastatic cancer. The delayed effects are thought to be due to development of fibrosis from damaged nerve and vascular tissue (13). The latent interval for delayed radiation brachial plexopathy varies from several months up to more than 10 years. In most reported series the latent interval is at least 12 months, with a broad peak of onset from 2 to 4 years after the radiotherapy (31). Median latent intervals were shorter for metastatic than for radiation plexopathy patients in some but not all series, with considerable overlap between the two groups (31; 33).

The most common presenting symptoms of radiation brachial plexopathy are numbness and paresthesias of the upper extremity. Pain can also develop but is not as severe when compared to malignant plexopathies. Symptoms usually begin 6 months after radiation therapy but can occur 2 to 3 decades after treatment (38). In approximately two thirds of patients, the motor and sensory deficits gradually worsen over several years to a level of severe neurologic disability. In the remainder of patients, progression spontaneously ceases after 1 to 3 years (29).

Radiation-induced skin changes, lymphedema, and soft tissue induration of the axilla and supraclavicular fossa are each present in 30% to 75% of patients with radiation brachial plexopathy (31; 33). These changes are also present in a lower but still sizable proportion of patients with brachial plexus metastases. Horner syndrome is an important diagnostic clue for metastatic brachial plexopathy, which occurs as a result of extrinsic compression of the ascending sympathetic chain. Patients with Horner syndrome frequently have paraspinal and epidural tumor extension.

The distribution of neurologic signs and symptoms varies somewhat between patients with metastatic versus radiation brachial plexopathy, but the degree of overlap prevents an absolute distinction from being made. In general, metastases to the brachial plexus predominantly involve the lower trunk or C8-T1 nerve roots, whereas radiation-induced plexopathy primarily involves the upper trunk with weakness in C5-C6 innervated muscles. The different localization of radiation versus metastatic plexopathy was thought to be due to the close proximity of the lower trunk of the plexus to the lateral group of axillary lymph nodes (which drain the breast) and to the upper lobe of the lung, enabling breast and lung tumors to reach the plexus by direct extension. Lymphomas from cervical or axillary lymph nodes can also directly invade the plexus (02). The relative sparing of the lower plexus in radiation injury was believed to be due to partial shielding of these elements by the clavicle.

In several reported series, however, the clinical distinction between radiation and metastatic brachial plexopathy was less clear-cut. Many patients with radiation brachial plexopathy in these series had weakness involving mainly the muscles innervated by the C8-T1 roots or lower trunk (33). Conversely, "diffuse" involvement of the plexus in some series was equally common among patients with metastases and patients with radiation damage (45).

Although not infallible, the most reliable clinical feature to distinguish radiation brachial plexopathy from plexus metastases is pain. Metastatic brachial plexopathy is characterized by early, severe, and unrelenting pain that is difficult to manage. Pain often precedes numbness or weakness by up to several months. Pain is generally, but not always, most severe in a C8-T1 dermatomal distribution. Exceptional patients with radiation plexopathy do have early and severe pain.

If pain and sensory symptoms are absent, one should consider the possibility of motor neuron disease. Radiation-induced lumbosacral plexopathy most often follows irradiation for pelvic tumors, including carcinoma of the bladder, uterus, or cervix; testicular tumors; or lymphoma. The latent interval between irradiation and onset of neurologic symptoms in patients with lumbosacral plexus injury varies from 3 months to 30 years, with a median interval of approximately 5 years. Approximately one third of patients have early and prominent numbness and paresthesias. Pain eventually occurs in approximately one half of patients but is usually not early and is rarely as severe or unrelenting as the pain caused by lumbosacral plexus metastases. In most patients the neurologic signs and symptoms slowly progress over months to years, although a more rapid tempo of progression, stabilization of neurologic deficits after a period of progression, or even partial resolution of weakness and sensory symptoms are also possible (44).

Delayed radiation injury to the cauda equina after radiation can mimic lumbosacral radiation plexopathy. Patients develop asymmetric bilateral leg weakness and, less often, pain or sensory loss more than 10 years after radiation for lymphoma, testicular cancer, or other abdominal or pelvic tumors (15). MRI scans show patchy or multinodular enhancement along the conus medullaris and cauda equina. The neurologic deficits eventually stabilize or may slowly continue to worsen over the years.

Radiation injury to the cauda equina may also mimic the findings of lower-motor neuron predominant amyotrophic lateral sclerosis. Patients develop subacute unilateral or asymmetric bilateral leg weakness beginning 6 months to as long as 10 years after completion of irradiation. Examination shows muscle atrophy, fasciculations, normal or decreased tendon reflexes, flexor plantar responses, and no sensory or sphincter involvement. Sensory nerve action potentials on electrodiagnostic testing are usually normal. Needle EMG testing often shows active and chronic denervation changes. MR scans may show contrast enhancement along the cauda equina. In most patients, the syndrome slowly progresses over several months and then stabilizes, but it does not improve. It is currently believed that nerve roots are the primary site of injury in these patients, explaining the motor-predominant clinical manifestations. However, radiation damage to the anterior horn cells has also been postulated (21). Primary benign or malignant nerve sheath tumors can arise in the brachial or lumbosacral plexus and become symptomatic due to local compression and mass effect (12). Schwannomas, neurofibromas, and neurinomas are common benign tumors arising in the brachial plexus (02). Approximately one third of reported patients have neurofibromatosis type 1. Schwannomas tend to affect the proximal portions of the brachial plexus (02). Some neurofibromas are "dumbbell tumors" that extend into the spinal epidural space. In the setting of neurofibromatosis, neurofibromas tend to affect a sizeable portion with neurologic deficits occurring more commonly than other nerve sheath tumors (02). Malignant nerve sheath tumors arising in the brachial plexus are much less common than benign tumors.

Malignant peripheral nerve sheath tumors have been reported to occur in patients from 2 to 20 years following 2000 to 4250 cGy irradiation to the region for Hodgkin disease or breast cancer (48; 08). Postirradiation malignant fibrous histiocytoma or other sarcomas in the brachial or lumbosacral plexus have also been reported (22; 45).

There are anecdotal reports of plexopathy occurring as a complication of cancer treatment other than radiation. Brachial plexopathy with pain and weakness beginning several hours after administration of high-dose (3 g/m2) cytarabine occurred in a patient with leukemia, leaving residual weakness and atrophy (43). Reversible unilateral or bilateral brachial plexopathy has occurred within 1 week of intravenous interleukin-2 infusion for renal carcinoma or melanoma (34). Lumbosacral plexopathy with acute onset and permanent neurologic deficit has been described following infusion of cisplatin, doxorubicin (Adriamycin), 5-fluorouracil, or other chemotherapy agents into the internal iliac artery (07). Lumbosacral polyradiculopathy can occur acutely after intrathecal methotrexate administration (40).

Brachial plexopathy may rarely occur as a paraneoplastic syndrome in patients with Hodgkin lymphoma (32), manifesting as painless, bilateral asymmetric weakness and sensory symptoms that may improve with prednisone. Involvement of the anterior horns of the cervical spinal cord in patients with paraneoplastic encephalomyelitis and small cell lung carcinoma may produce bilateral upper extremity weakness mimicking brachial plexopathy (36).

Rarely, a paraneoplastic syndrome associated with Hodgkin lymphoma or thymoma can cause a subacute motor neuronopathy (18). Patients develop subacute, progressive weakness, often asymmetric and predominantly involving the legs, without pain or significant sensory loss. Examination reveals a pure lower motor neuron picture with moderate to severe weakness, fasciculations, atrophy, and diminished deep tendon reflexes. The syndrome can occur prior to the discovery of a neoplasm, as well as after attainment of complete remission. The neurologic deficits eventually spontaneously stabilize or improve in most patients after treatment of the underlying malignancy (18).

Diagnostic workup

A chest radiograph may be ordered when suspecting metastatic brachial plexopathy to evaluate the apices of the lungs and supraclavicular area. However, in the contemporary era, CT or MRI imaging would more often be used due to their superior sensitivity.

MRI neurography scanning is generally the next method of choice for detecting brachial plexus metastases as it provides simultaneous multiplanar images and better resolution of the neural elements than CT scanning. The sensitivity of MR scanning in detecting metastases in or near the brachial plexus is more than 80% (45).

Diffusion-weighted MRI neurography may provide better visualization of brachial plexus metastases than standard fat-suppressed T2-weighted and T1-gadolinium imaging and also help to distinguish between radiation versus metastatic plexus injury (03). The most specific MRI evidence for plexus metastasis is the presence of a mass, but this is not always observed. Nonspecific thickening and enhancement can be seen of patchy regions of the plexus. MRI scans may show enlarged supraclavicular lymph nodes or displacement of adjacent structures rather than a mass within the plexus itself. There is increased T2 signal intensity with metastatic infiltration, whereas radiation-induced plexopathy has low signal intensity on T2-weighted sequencing (46). Postcontrast enhancement can be seen with both radiation and metastatic plexopathies (46). Therefore, occasional patients with no obvious tumor mass and MRI changes consistent with radiation injury turn out to have plexus metastases. Follow-up studies can prove to be helpful to look for disease progression versus stabilization (46). It is important to recognize that up to 50% of patients with brachial plexus metastases have concomitant tumor spread into the epidural space (31), so CT or MR scanning of the plexus should also be extended to look at the spine.

Electrodiagnostic studies in patients with brachial plexus metastases usually show evidence of severe axonal sensorimotor abnormalities, either diffuse or predominantly affecting the lower trunk (33). Absent or reduced amplitude ulnar and medial antebrachial sensory nerve action potentials are important indications of lower trunk or medial cord plexus involvement. There are absent or reduced amplitudes of median and ulnar compound muscle action potentials, with corresponding EMG evidence for denervation.

The electrophysiologic abnormality that most reliably differentiates radiation from metastatic brachial plexopathy is myokymia, defined as spontaneous, rhythmic grouped repetitive discharges of the same motor unit (41). Myokymic discharges seen on EMG strongly suggest radiation plexopathy and are rarely seen with neoplastic invasion of the brachial or lumbosacral plexus (41).

The most frequent electrophysiologic abnormalities in patients with lumbosacral plexus metastases are asymmetric reduction of sensory nerve action potentials, reduced compound motor action potentials, and EMG findings of active and chronic denervation changes in L4-S1 innervated muscles. Prominent fibrillation potentials or positive sharp waves are often seen. Patients may have evidence of bilateral plexus involvement despite unilateral symptoms. The discharges may be widely scattered and are often present in muscles that are not overly weak.

Positron emission tomography using fluoro-deoxyglucose (FDG-PET) may be useful in identifying metastatic breast cancer in or near the brachial plexus or other rare tumors, such as malignant peripheral nerve sheath tumor (01; 08). In some patients, FDG-PET identifies tumors not clearly imaged by CT or MR scanning. Combined CT-FDG-PET imaging may identify diffuse lymphomatous infiltration of the plexus (20; 23). FDG-PET scans were negative in a few reported patients believed to have radiation plexopathy. To date there is little solid information regarding the reliability with which FDG-PET can differentiate brachial plexus metastases from radiation plexopathy.

Radionuclide bone scans in patients with lumbosacral plexus metastases show abnormal uptake in the pelvis, sacrum, or vertebrae in approximately two thirds of patients (25). CT scans of the pelvis or abdomen show a tumor mass, lymphadenopathy, or bone erosion in at least 75% of patients with lumbosacral plexus metastases (25; 44). Up to one third of patients have bilateral tumor on scans despite unilateral signs and symptoms. Up to 40% of patients have concomitant epidural tumor extension. Tumor cells that spread in a linear fashion along nerve roots may not be apparent on scans; in these patients the initial scan is normal but repeat scans usually show tumor within 6 months.

Surgical exploration or CT-guided biopsy of the brachial plexus may be done as a "last resort" in attempting to differentiate metastases from radiation damage, but surgery is technically challenging, carries morbidity, and can miss metastatic tumor (31). Proposed criteria for surgical exploration of the brachial plexus include (1) no evident tumor on MR scan, (2) definite neuroanatomic localization of the signs and symptoms (to determine the appropriate surgical approach), (3) no evidence of diffuse metastatic tumor, and (4) onset of neurologic symptoms several years after successful tumor therapy (19). FDG-PET hypermetabolic activity may be used as a guide for biopsy planning and diagnostic clarification (08).

Management

Radiotherapy produces significant pain relief in at least 40% to 70% of patients with brachial plexus metastases from lung or breast carcinoma (31). Data comparing hypofractionated radiotherapy and conventional radiotherapy show similar treatment and toxicity outcomes in postmastectomy breast cancer patients (09; 47). Fewer than one third of patients have improvement in focal motor or sensory deficits following radiotherapy. In patients who have already undergone previous radiotherapy to the region, the issue of a second course of radiation needs to be made on an individual basis. Chemotherapy is effective in some patients with brachial plexus metastases from breast cancer or lymphoma (26; 23). Treatment options for patients who have severe and persistent arm pain include dexamethasone, transcutaneous electrical nerve stimulation units, local nerve blocks, and percutaneous cervical cordotomy (31). Levetiracetam has been reported to provide pain relief in patients with plexus metastases when used as an adjunct to opioid analgesics (16). Methadone has been used as monotherapy for neuropathic pain (35).

Radiotherapy for lumbosacral plexus metastases produces significant pain relief in 30% to 80% of patients in different published series (25). As with brachial plexus, reversal of neurologic deficits is less likely and occurs in no more than one third of patients. Other treatment options for intractable pain include cordotomy, chemical rhizotomy, local anesthetic injections, or continuous epidural or intrathecal infusion of opiates (49). The first randomized, placebo-controlled clinical drug trial in the treatment of radiation-induced plexopathy failed to show a benefit with combination pentoxifylline-tocopherol and clodronate (14).

Outcomes

The pain and neurologic deficits caused by brachial or lumbosacral plexus metastases frequently represent a debilitating complication. Response to therapy depends in part on the duration and severity of symptoms. Patients with relatively recent onset of pain and minimal or no neurologic deficit have a higher likelihood of significant improvement with treatment than patients with severe, longstanding pain and sensory or motor deficit. In the latter group the neuropathic or causalgic pain may prove intractable.