Electrophysiological findings in patients who received radiation therapy over the brachial plexus: a magnetic stimulation study

ELSEVIER

Electroencephalography and clinical Neurophysiology 101 (1996) 483-490

Q

Electrophysiological findings in patients who received radiation therapy over the brachial plexus: a magnetic stimulation study Ari Boyaclyan a'*, A. Emre (~ge a, Jale Yazlcl a, I~lk Aslay b, Aynur Baslo a aDepartment of Neurology, lstanbul University, lstanbul Faculty of Medicine, 34390 ~apa lstanbul, Turkey bDepartment of Radiation Oncology, lstanbul University, lstanbul Faculty of Medicine, 34390 ~apa lstanbul, Turkey Accepted for publication: 21 August 1996

Abstract Clinical and electrophysiological findings of 47 asymptomatic females who received radiation therapy (RT) over their brachial plexus region are presented and compared with 8 radiation-induced brachial plexopathy (RBP) and 4 neoplastic brachial plexopathy (NBP) patients. In the asymptomatic group, abnormal findings were more frequent in patients whose post-RT period was longer than 1 year. Flexor carpi radialis H reflex was delayed or absent in 19 patients (52%) in this subgroup of asymptomatic cases, as compared to only 2 (18%) of the patients with post-RT periods of less than 1 year. Magnetic cervical nerve root stimulation was performed in 16 asymptomatic cases, with the conclusion that there was no significant difference between the irradiated and non-irradiated sides with regard to latencies, amplitudes and areas of the muscle responses. In spite of this, muscle response amplitudes and areas on both sides were significantly lower than those obtained from healthy controls. It was postulated that this finding resulted from hypoexcitability to magnetic stimulation produced by slight nerve root damage. Any part of the brachial plexus could be affected in RBP and NBP patients. Myokymic discharges were found at a high rate (87.5%) in RBP group. Cervical magnetic nerve root stimulation may have a diagnostic value in these patients in localizing the nerve lesion over the brachial plexus. Copyright © 1996 Elsevier Science Ireland Ltd. Keywords: Radiation therapy; Magnetic stimulation; Nerve conduction studies; Brachial plexus; Plexopathy

1. Introduction Differential diagnosis is usually difficult in a patient who presents with the symptoms and signs of brachial plexopathy developing after radiation therapy (RT). Electrophysiological evaluation in these patients can give important clues which help to distinguish tumoral invasion from radiation injury. M y o k y m i c discharges, seen frequently in radiation-induced neuropathies, are useful in this regard (St6hr, 1982; Lederman and Wilbourn, 1984; Gutman, 1991; Esteban and Traba, 1993). Roth et al. attracted attention to the presence of prolonged conduction block in radiation plexopathy and its association with m y o k y m i c discharges (Roth et al., 1986; Roth and Magistris, 1987; Roth et al., 1988). The causal relationship between the conduction block and the m y o k y m i c discharges was further emphasized by Esteban and Traba * Corresponding author. Fax: +90 212 5334393.

who found conduction block in all of their radiation plexopathy cases (Esteban and Traba, 1993). This led us to think that if conduction block is one of the main electrophysiological findings in radiation plexopathy, its presence in an asymptomatic patient who received RT might be an early indicator of the plexus disease. This study was planned to investigate electrophysiological findings in post-RT patients with special attention to the clues which can indicate the presence of conduction block in the nerve fibers constituting the brachial plexus. 2. M e t h o d s 2.1. Cases 2.1.1. Radiation therapy patients Fifty-nine patients who had received RT to their brachial plexus region were included in this study. They were sent to the Department of Neurology from the outpatient

0924-980X/96/$15.00 Copyright © 1996 Elsevier Science Ireland Ltd. All rights reserved Pit S092!-884X(96)95630- 1

EEM 95630

484

A. Boyactyan et al. / Electroencephalography and clinical Neurophysiology 101 (1996) 483-490

clinic of the Institute of Oncology as a regular part of their periodic examinations and were examined clinically by one of the authors (A. Boyaclyan). Those who had any metabolic or systemic disorder, other than their primary cancer, which could affect nerve conduction studies were excluded. Patients with clinical and electrophysiological findings indicating common nerve entrapments in the upper extremities (e.g. median nerve at wrist, ulnar nerve at elbow and neurogenic thoracic outlet syndrome) were also not included. In 27 asymptomatic patients (see below), cervical computed tomography did not show any structural lesion which can cause a cervical radiculopathy. Forty-seven neurologically asymptomatic female breast cancer patients and 12 cases with brachial plexus involvement were evaluated. (1) Asymptomatic cases. Patients in this group were treated with a teletherapy 6°Co device. RT was applied to their peripheral lymphatic regions (including supraclavicular lymphatic groups), and to mid-axillary line at an approximate total dose of 5350 cGy given in 25 fractions. Twenty-five of these cases were taken into an additional chemotherapy program with the following combination (total doses are given in parentheses): cyclophosphamide (600 mg]m2), fluorouracyl (600 mg/m 2) and methotrexate (40 mg/m 2) or adriblastine(40 mg[m2). Asymptomatic patients were divided into two subgroups according to the time interval between the completion of RT and the neurological evaluation. Group la (36 cases): post-RT time interval was more than 1 year ( 1 - 1 4 years, mean 4.7 years). The mean age of this subgroup was 51.3 (range 4 7 - 5 8 ) years. Group lb (11 cases): post-RT time interval was less than 1 year (2 months to 1 year, mean 7.3 months). The mean age was 49.7 (range 3 0 - 7 5 ) years. (2) Symptomatic cases. This group consisted of 8 (6 females and two males) radiation-induced brachial plexopathy (RBP) and 4 female neoplastic brachial plexopathy (NBP) cases. Mean ages in the first and second groups were 53.3 (49-60) and 57.5 (42-66) years, respectively. Group 2a: RBP patients. There were 5 breast, one parotid, one nasopharynx and one lung cancer patients in this group. Patients were examined clinically and with direct radiography, blood chemistry and hematological tests as well as with bone scan and CT (or MRI) to show that the plexopathy was not related to the neoplastic infiltration. Local RT of 5000-6000 cGy to the areas described above had been completed 1 - 9 years before the neurological examination. Group 2b: NBP patients. Four female breast cancer patients formed this group. Every NBP patient was accepted to have neoplastic brachial plexus infiltration if there was biopsy-proven involvement of the local lymph nodes and extensive organ and bone metastases. All of these patients had RT (3600-6000 cGy) to the brachial plexus region which had been completed 1 - 6 months earlier.

2.1.2. Controls

Nerve conduction and late response studies were conducted on 12 female and 7 male volunteers whose ages ranged between 36 and 80 years (mean 55.8 years). Thirteen females and 7 males volunteered for the magnetic stimulation studies (ages 2 7 - 6 2 years, mean 37.4 years). 2.2. Electrophysiological methods

Electrophysiological studies, performed with Medelec MS 25 and MS 92-B equipment, are listed in Table 1. Nerve conduction velocity (NCV) studies were conducted by means of standard methods (Kimura, 1989), and values obtained outside _+2.5 SD of the control group means (or an amplitude side difference of more than 50% for sensory NCV studies) were considered as abnormal (Wilbourn, 1985). Special attention was paid to deliver supramaximal stimulation in motor conduction studies, using up to 350 V, 200 txs duration square waves which were delivered by a hand-held surface stimulator with a cathode-anode distance of 2.5 cm. Fifty percent reduction in the amplitude of compound muscle action potential (CMAP) elicited by stimulation at the proximal end of a nerve segment as compared to that of the CMAP produced by distal end stimulation was accepted as partial conduction block over this nerve trunk (Lewis et al., 1982; Rhees et al., 1990). Supramaximal stimulation of the brachial plexus at Erb's point was sometimes very difficult, especially in radiotherapy patients with a swollen and endurated supraclavicular region. Therefore, if a low amplitude muscle response was recorded by using the above mentioned stimulator, Erb's point stimulation was repeated with the monopolar technique in which the anode was taped under the scapular region and the stimulator cathode was applied by hand under pressure on the supraclavicular fossa (Roth et al., 1986). In late response studies, minimum latency of 20 F responses was measured (Fisher, 1982). Flexor carpi radialis H reflex (FCR-H) was recorded according to the method of Ongerboer de Visser et al., and either the absence of the response or a side difference exceeding 0.85 ms was considered as abnormal (Ongerboer de Visser et al., 1984). In needle EMG, insertional and spontaneous activities as well as the shapes and recruitment patterns of motor unit potentials were examined in the muscles indicated in Table 1. Additional upper extremity muscles were examined when needed. 2.2.1. Magnetic stimulation of lower cervical roots

This study was performed on 16 asymptomatic radiotherapy patients (all of whom belonged to group la), 3 RBP and 4 NBP cases. Magnetic stimulation studies

485

A. Boyactyan et al. / Electroencephalography and clinical Neurophysiology 101 (1996) 483-490 Table 1 Electrophysiological study protocols and cases with abnormal findings Electrophysiologicalstudies

Sensory NCV Ulnar n. Median n. Median n. Motor NCV Ulnar n.

Patient groups

Recording

Stimulation

Examineds i d e

Wrist Wrist Wrist

5th digit 1st digit 3rd digit

Bilateral Bilateral Bilateral

Hypothenar

Wrist Elbow Axilla Erb's point Wrist Elbow

Symptomatic sidea

Wrist AF

Bilateral Bilateral

Median n.

Thenar

Late responses F wave H reflex

Hypothenar FCR

Symptomatic side

Asymptomatic lb (n = 11)

RBP (n = 8)

NBPb (n = 4)

4 0 0

1 0 0

8 4 4

2 2 1

0 0 0 12 0 0

0 0 0 2 0 0

1 l 1 1 1 1

3 3 3 3 2 2

0 19

0 2

8 8

3 4

Asymptomatic la (n = 36)

Numbers of pathological results obtained in NCV studies (rows) are presented according to the case groups (columns). Needle EMG: deltoid, biceps, triceps, flexor carpi radialis (FCR), extensorcarpi ulnaris,abductordigiti minimi and abductorpollicisbrevis muscles on the symptomaticside (see text). RBP, radiation-inducedbrachial plexopathy; NBP, neoplasticbrachial plexopathy; AF, antecubitalfossa aBilateral in 16 patients in whom magnetic stimulationstudies were performed, bOne patient who belonged to NBP group had bilateral involvementof brachial plexus.

were performed with the Medelec MS 25 system using a Novametrix Magstim 200 equipment and a standard coil with the outside diameter of 12 cm. Maximum output at the center of the coil was 1.5 T. Nine m m Ag-AgCI cup electrodes were placed over the tendon and bellies of the abductor digiti minimi (ADM) muscles bilaterally. Prior to the magnetic stimulation studies, ulnar nerves on both sides were stimulated supramaximally at wrist, elbow and axilla with square wave electrical stimuli of 100200 tzs. Magnetic coil was applied tangentially over the midline of fully flexed neck of the sitting subject. Stimulator output was 100% during the study. The center of the coil was initially at C5-C6 level in a clockwise direction of current flow (when viewed from the outside) for recording from the right A D M muscle, and in the opposite direction for recording from the left side. Then, the coil was moved in a rostro-caudal direction until the muscle response with the shortest latency and highest amplitude was elicited (maximum muscle response). Recordings from right and left ADM muscles were made separately and at least 3 m a x i m u m muscle responses from each side were stored for latency and amplitude measurements. Bandpass was 2 H z - 1 0 kHz, sweep speed was 5 ms/div and sensitivity was 1 - 1 0 mV/div. Muscle responses to electrical and magnetic stimulations were stored in a PC by 'Datamed' program and were transmitted again to MS 25 system for measurements. Latencies were measured to the beginning of the first negative deflection. Amplitude and area calculations of the negative peaks were made by the EMG machine. Student's t test was used for statistical analysis.

3. Results 3.1. Asymptomatic group 3.1.1. Clinical findings Group l a (patients with the post-radiotherapy period longer than 1 year): decreased or unobtainable tendon reflexes and abnormal sensory findings were found in the upper extremity on the irradiated side in 8 (22%) patients. Sensory abnormalities consisted of hypoesthesia and hypoalgesia in the medial aspect of the arm in 5 of the cases. RT-related induration of supraclavicular region and edema of the affected upper extremity was found in 10 patients (27.7%). Group l b (patients with post-RT period less than 1 year): there was only one patient (9%) with an abnormal sensory finding (hypoesthesia in the medial aspect of the arm). Muscle weakness and abnormal tendon reflexes was not found in this group. Supraclavicular induration and arm edema was recorded in one patient. 3.1.2. N C V and EMG findings Group la: FCR-H reflex was absent or delayed in 19 patients (52%). Unelicitable or low amplitude sensory nerve action potentials were recorded in 4 cases. In 12 (33%) patients, more than 50% reduction in muscle response amplitude with Erb's point stimulation as compared to the CMAP elicited with axillary stimulation in ulnar NCV studies, was found. Abnormal needle EMG findings, consisting of positive sharp waves at rest, were obtained in one patient.

A. Boyactyan et al. / Electroencephalography and clinical Neurophysiology lO1 (1996) 483-490

486 Table 2

Results of the cervical magnetic stimulation studies in asymptomatic patients and normal controls

Controlsa AS-normal side AS-irradiated side

Latency (ms)

Amplitude (mV)

Area (/zVs)

NACT (ms)

Amplitude loss (%)b

13.3 _+ 0.96 13.3 -+ 0.95 13.3 ___1.08

8.49 _+ 2.88 6.8 ± 2.16 6.1 _+ 2.42

31.4 -+ 11.9 23.9 ± 9.1 20.1 ± 8.22

3.58 ± 0.45 3.4 + 0.39 3.4 __+0.64

16.53 ± 21 24.62 _+ 19.35 29.86 _+ 22.59

0.25 ± 0.22 0.01 _+0.65

1.48 _+0.96 0.66 ± 2.3

4.23 _+4.02 3.52 ± 9.31

0.42 ± 0.28 0.009 ± 0.83

-

NS NS NS

NS P = 0.006 P = 0.04

NS P = 0.02 P = 0.001

NS NS NS

NS NS NS

Side difference Controls AS

Statistical significance AS-normal side/AS-irradiated side Controls/AS-normal side Controls/AS-irradiated side

NACT, neck-axilla conduction time; AS, asymptomatic patients; NS, not significant. "Cumulative values of both sides. ~I'he reduction rate (%) of the amplitudes of the maximum muscle responses to cervical magnetic stimulation as compared to the CMAPs elicited by axillary electrical stimulation.

Group lb: in this group, FCR-H reflex was absent in two (18%) patients. Low amplitude ulnar nerve sensory action potential was recorded in one patient (9%). CMAP amplitude reduction of more than 50% with Erb's point stimulation was found in two cases (18%). Needle EMG findings were normal in all of the Group lb patients. Cervical nerve root stimulation (Table 2): In asymptomatic patients, all of whom belonged to group la, there was no significant difference between the irradiated and non-irradiated sides with regard to latencies, amplitudes and areas of the maximum muscle responses (Table 2). In 100

spite of this, maximum muscle response amplitudes and areas on both sides of these patients were significantly lower than those obtained from normal healthy controls. Fig. 1 shows the percentage amplitude loss of the muscle responses elicited by electrical stimulation at Erb's point and cervical magnetic stimulation (as compared to the CMAPs evoked by axillary electrical stimulation) in 16 patients who were included in the magnetic stimulation studies. In 5 cases, amplitude reduction rates were nearly the same with the two stimulation modes. Smaller reduction rates were found with cervical magnetic stimulation in

--

90 80 7O

~ 60 '~ ~

50-4O

--

30

-

20

-

I0

-

E

I

,

r~

Patient name

• Cervical magnetic stimulation [] Electrical stimulation at Erb's point.

/

)

Fig. 1. The percentage amplitude loss of the muscle responses elicited by electrical stimulation at Erb's point and cervical magnetic stimulation (as compared to the CMAPs evoked by axillary electrical stimulation) in 16 patients who were included in the magnetic stimulation studies.

487

A. Boyactyan et al. / Electroencephalography and clinical Neurophysiology 101 (1996) 483-490

the other 11 cases. Interestingly, more than 50% amplitude loss (100% in 4 cases, 92% in one, and 50% in another) was recorded with E r b ' s point electrical stimulation in 6 cases, which can be misinterpreted to indicate partial or complete conduction blocks. With magnetic stimulation, amplitude reduction rates were between 0 - 4 7 % in these patients, proving the absence of total conduction block in 4 of them. 3.2. Symptomatic group 3.2.1. RBP patients The clinical and electrophysiological findings were consistent with a lower brachial plexus involvement in 4 patients and whole plexus lesion in the remaining 4. Needle E M G examination of 7 cases (87.5%) revealed myokymic discharges and grouped fasciculations in one or more muscles of the symptomatic upper extremity (Fig. 2). Cervical magnetic nerve root stimulation was performed in 3 RBP patients. On the symptomatic side, maximum muscle response latencies were abnormally prolonged in two cases (in one of them, magnetically and electrically elicited muscle responses had very low amplitudes on the symptomatic side as compared to the non-irradiated extremity) and no muscle response was evoked by cervical magnetic nerve root stimulation in the last. 3.2.2. N B P patients Clinical examination revealed lower plexus involvement in two cases and upper plexus affection in one. Both sides were affected in the last patient, with lower brachial plexus affection on the left and whole plexus findings on the right side. In NCV studies, low amplitude or unrecordable sensory nerve action potentials were found corresponding to the affected part of brachial plexus. Needle E M G findings consistent with partial denervation were seen in the muscles innervated by the involved brachial plexus region. M y o k y m i c discharges and grouped fasciculations were not observed in any muscle in these patients. Cervical magnetic nerve root stimulation was performed in 3 patients in this group. M a x i m u m muscle response amplitude, area and latency values were within normal limits in two. In the last case in whom brachial plexus was bilaterally involved, interside differences of maximum muscle response latency, area and amplitudes were abnormal due to the more serious involvement on the right side (longer latency and lower amplitude maximum muscle responses were recorded on this side).

4. Discussion Biological mechanisms underlying RBP are still not completely understood (Wilbourn, 1993). Tissue fibrosis surrounding the brachial plexus has been considered to be

A

B

° C

9 h

C

,On',Vl 5r =

Fig. 2. (A,B). Myokymic discharges recorded from the flexor digitorum sublimis (A) and flexor carpi radialis (B) muscles on the symptomatic side of a patient with RBP; vertical bars, 0.1 mV; horizontal bars, 1 s. (C). Cervical magnetic nerve root stimulation study on the same patient. Electrical stimulation at wrist (a and e), elbow (b and f) and axilla (c and g) and cervical magnetic nerve root stimulation (d and h) on the asymptomatic (first 4 traces) and symptomatic (last 4 traces) sides. Note the 31% and 74% amplitude reduction in the neck-axilla segments on the asymptomatic and symptomatic sides, respectively (area reduction ratios were 33% and 75%). the main pathological process leading to progressive entrapment of the nerve fibers (Stoll and Andrews, 1966; Thomas and Colby, 1972; Cascino et al., 1983; Thomas et al., 1985). The resulting peripheral nerve demyelination is assumed to be one of the major sources of the clinical and electrophysiological features of radiation neuropathies, such as m y o k y m i c discharges and grouped fasciculations (Allen et al., 1977; Albers et al., 1981; St6hr, 1982; Lederman and Wilbourn, 1984; Harper et al., 1989; Esteban and Traba, 1993). Therefore, conduction block or slowing across a nerve segment, which is involved by segmental demyelination, m a y be the first electrophysiological indi-

488

A. Boyactyan et al. / Electroencephalography and clinical Neurophysiology 101 (1996) 483-490

cator of a radiation-induced neuropathy. The main purpose of this study was to investigate if there were any electrophysiological findings pointing to the presence of early changes of the demyelination process, in a symptom-free group of patients who had been given RT over their brachial plexus. To our knowledge, there is no electrophysiological study in the literature analyzing this preclinical stage of radiation neuropathy. The other patient groups were included in the study for comparison of their electrophysiological findings with those of the asymptomatic cases. In the asymptomatic group, the considerable amount of sensory and reflex abnormalities, and their increase in frequency in patients evaluated after more than 1 year (group la), are noteworthy. A 'time-related' increase of the neurological signs is well known in RBP patients (Stoll and Andrews, 1966; Thomas and Colby, 1972). Therefore, the clinical findings in asymptomatic cases can be considered as early signs of the impending RBP. However, it seems reasonable to think that the majority of these patients will stay within the asymptomatic domain and a few of them will be 'struck' by RBP, since it is reported that only 3% of the cases who received RT doses over 6000 cGy developed clinically overt RBP after 5 years (Basso-Ricci et al., 1980). In fact, a lower RBP incidence should be expected in our asymptomatic group who had received a total dose of 5350 cGy. NCV and late response studies performed in the asymptomatic group showed a high rate of prolongation or absence of FCR-H reflex, and CMAP amplitude loss with Erb's point electrical stimulation. The usefulness of the FCR-H reflex in the evaluation of RBP was formerly reported (Ongerboer de Visser et al., 1984). Abnormalities of this reflex can be thought to result from dispersion of conduction due to segmental demyelination in the brachial plexus fibers carrying the afferent volley to the spinal cord (Ongerboer de Visser et al., 1984). Abnormal FCR-H reflexes in our patients could also result from cervical radiculopathies which are frequently seen in people of the same age group. This possibility was excluded in only half of the asymptomatic cases by cervical computed tomography. However, FCR-H abnormalities were more frequent in our group 1a patients than the cases in group 1b. These two groups showed similar patient characteristics (concerning their ages and their therapeutic protocols) except the length of post-RT period. Therefore, it is reasonable to think that the post-RT changes in the brachial plexus, which became most pronounced in patients examined more than 1 year after RT, were the cause of prolonged or unobtainable FCR-H reflexes in our asymptomatic cases. Another abnormal electrophysiological finding which showed a time-related intensification was the substantial amount of CMAP amplitude reduction (more than 50%) with electrical stimulation at Erb's point in 6 of the group la asymptomatic patients. At first sight, this finding may

suggest the presence of some degree of conduction block in the brachial plexus nerve fibers. However, Erb's point stimulation comprises some technical difficulties especially in a patient with an endurated and edematous supraclavicular region (Cornblath et al., 1991). Other electrophysiological methods in the evaluation of the brachial plexus are nerve root stimulation, which can be performed by needle electrodes (Berger et al., 1987), percutaneous electrical (King and Chiappa, 1990) or magnetic stimulators. As a non-invasive and relatively painless procedure, magnetic stimulation is more practical than the others. Magnetic excitation of cervical nerve roots occurs at a fixed point located a few centimeters distal to the anterior horn cell in the vicinity of the intervertebral foramen (Ugawa et al., 1989; Britton et al., 1990; Claus, 1990; Cros et al., 1990; Schmid et al., 1990). This excitation takes place proximal to the plexus and therefore is not affected by factors which interfere with Erb's point electrical stimulation. In all of our asymptomatic patients in whom partial or complete conduction block was implicated with Erb's point electrical stimulation, maximum muscle responses evoked by magnetic nerve root stimulation showed lower rates of amplitude loss (Fig. l). In 4 of these patients, the possibility of a complete conduction block was excluded by the addition of magnetic nerve root stimulation to the electrophysiological battery, which showed that at least some brachial plexus fibers under the irradiated area were able to conduct. This represents the inadequacy of Erb's point electrical stimulation in some patients and shows the superiority of the painless magnetic stimulation for the demonstration of brachial plexus continuity in patients with endurated skins secondary to RT. In cervical magnetic nerve root stimulation studies in the asymptomatic group, there was no significant difference between the values on the irradiated and healthy extremities with regard to the maximum muscle response latencies, neck-axilla conduction times, maximum muscle response amplitudes and amplitude reduction rates (comparing CMAPs evoked by axillary electrical stimulation). Also, there was not any significant difference in the maximum muscle response latencies and neck axilla conduction times between the irradiated extremities and healthy controls (Table 2). Magnetic stimulation of nerve roots gives reproducible muscle responses which are reliable in their latencies (Britton et al., 1990; Claus, 1990; Cros et al., 1990; Schmid et al., 1990; 0ge et al., 1994); however, it is generally submaximal even at maximum output levels which we used in this study (Evans et al., 1990). Thus, it is thought to be an inadequate method in the investigation of conduction block in the nerve roots or plexi when the current electrophysiologic criteria on conduction block are used (Feasby et al., 1985; Ugawa et al., 1989; Britton et al., 1990; Claus, 1990; Schmid et al., 1990; Cornblath et al., 1991; Kaji and Kimura, 1991). Highly variable amplitudes of the maximum muscle

A. Boyactyan et al. / Electroencephalography and clinical Neurophysiology 10l (1996) 483-490

responses elicited by cervical magnetic stimulation, which are submaximal on most occasions, could readily preclude the detection of a mild segmental demyelination. However, the high frequency of abnormal FCR-H responses, in the absence of any indication of partial denervation in needle EMG, suggests that there was this kind of demyelinating lesions in the brachial plexi of asymptomatic patients. These lesions must be slight in degree and were sufficient to produce FCR-H reflex abnormalities only (by causing dispersion of the impulse volleys), without reaching a degree capable of affecting the conduction velocities and maximal muscle response amplitudes. Maximum muscle responses obtained on both irradiated and healthy sides of the asymptomatic group had significantly lower amplitudes and areas compared to healthy control values (Table 2). This observation might be related to the hypoexcitability of the cervical nerve roots bilaterally. It is known that mild nerve lesions can cause elevation of the magnetic stimulation threshold (Schriefer et al., 1988; R6sler et al., 1989; Oge et al., 1993). In our patients, this kind of hypoexcitability could result either from irradiation of the nerve roots on both sides of the spinal cord and/or from the effects of chemotherapy on these structures. Drugs with a known potential for producing neuropathy had not been administered to our asymptomatic cases. However, the presence of mild nerve lesions due to the cytostatic agents used in their treatment, which induced hypoexcitabilty to magnetic nerve root stimulation without producing other clinical and electrophysiological signs, could not be excluded. In the RBP group, clinical and electrophysiological examinations disclosed involvement of the whole brachial plexus in 4 cases and the lower part in the rest. These findings are in agreement with the opinions insisting that the upper part of the brachial plexus is not exclusively involved in RBP (Lederman and Wilbourn, 1984; Mondrup et al., 1990). The presence of myokymic discharges and grouped fasciculations in one or more muscles in 7 cases indicate their high frequency in RBP (Allen et al., 1977; Albers et al., 1981; StOhr, 1982; Lederman and Wilbourn, 1984; Thomas et al., 1985; Harper et al., 1989; Gutman, 1991). In a recent study, Esteban and Traba pointed out that 'better preservation of the muscle should have an indirect relationship with the amount of fasciculation myokymic activity, facilitating its expression, but only when a nerve conduction block of sufficient degree and duration exists' (Esteban and Traba, 1993). It was interesting to see that our patient without myokymic discharges was the most severely affected case within the whole group, showing an important amount of axon loss in his EMG study. In the other patients of this group, electrophysiological findings demonstrated partially denervated muscles in needle EMG and low amplitude CMAPs with distal nerve stimulation showing the presence of partial axonal loss, but not as pronounced as in the above mentioned case. The F responses which were

489

elicited in all patients in the asymptomatic group were totally absent on the symptomatic sides of all patients in this group. This finding can be related to the severity of brachial plexus lesions in RBP group as compared to asymptomatic cases. Cervical magnetic nerve root stimulation had some value in the localization of brachial plexus lesions in RBP group, giving at least one abnormality in the maximum muscle response latency or amplitude values. In NBP patients, predominant brachial plexus involvement occurred either in the lower or upper regions of this structure and showed that the site of damage could not help in distinguishing NBP from RBP. However, the absence of myokymic activity in the clinical and the electrophysiological examinations in the NBP group indicated once again the importance of spontaneous activity of this type (Albers et al., 1981; Kori et al., 1981; Cascino et al., 1983). The failure of cervical magnetic nerve root stimulation in localizing the lesion in this group seems to be due to severe axonal loss. However, it is impossible to be more certain about the role of cervical magnetic nerve root stimulation in NBP, because of the small number of examined cases. In conclusion, there is some degree of segmental demyelination in brachial plexus fibers of asymptomatic patients after a sufficient time following RT. Motor and sensory NCV, magnetic stimulation and F wave studies are generally unable to demonstrate these mild demyelinating lesions, whereas H reflex studies can disclose abnormalities of this kind. Magnetic nerve root stimulation is useful in the demonstration of brachial plexus continuity in patients with RT-induced skin indurations which poses difficulties in the electrical stimulation of Erb's point. One of the most important problems, from the clinician's point of view, is to know which of the RT-receiving asymptomatic patients will develop RBP. Our results did not answer this question and further studies are needed in this field.

Acknowledgements We thank Dr. Feza Deymeer for her invaluable help. Some of the equipment in this study was provided by the Research Fund of the University of Istanbul (Grant no. 643/210994).

References Albers~ J.W., Allen, II, A.A., Bastron, J.A. and Daube, J.R. Limb myokymia. Muscle Nerve, 1981, 4: 494-504. Allen, A.A., Albers, J.W., Bastron, J.A. and Daube, J.R. Myokymic discharges following radiotherapy for malignancy. Electroenceph. clin. Neurophysiol., 1977, 43: 148. Basso-Ricci, S., della Costa, C., Viganotti, G., Ventafridda, V. and Zanolla, R. Report on 42 cases of postirradiation lesions of the brachial plexus and their treatment. Tumori, 1980, 66: 117-122. Berger, A.R., Busis, N.A., Logigian, E.L., Wierzbicka, M. and Shahani,

490

A. Boyactyan et al. / Electroencephalography and clinical Neurophysiology 101 (1996) 483-490

B.T. Cervical root stimulation in the diagnosis of radiculopathy. Neurology, 1987, 37: 329-332. Britton, T.C., Meyer, B.U., Herdmann, J. and Benecke, R. Clinical use of the magnetic stimulator in the investigation of peripheral conduction time. Muscle Nerve, 1990, 13: 396-406. Cascino, T.L., Kori, S., Krol, G. and Foley, K.M. CT of the brachial plexus in patients with cancer. Neurology, 1983, 33: 1553-1557. Claus, D. Central motor conduction: method and normal results. Muscle Nerve, 1990, 13: 1125-1132. Cornblath, D.R., Sumner, A.J., Daube, J., Gilliat, R.W., Brown, W.F., Parry, G.J., Albers, J.W. et al. Conduction block in clinical practice. Muscle Nerve, 1991, 14: 869-871. Cros, D., Chiappa, K.H., Gominak, S., Fang, J., Santamaria, J., King, P.J. and Shahani, B.T. Cervical magnetic stimulation. Neurology, 1990, 40: 1751-1756. Esteban, A. and Traba, A. Fasciculation-myokymic activity and prolonged nerve conduction block. A physiopathological relationship in radiation-induced brachial plexopathy. Electroenceph. clin. Neurophysiol., 1993, 89: 382-391. Evans, B.A., Daube, J.R. and Litchy, W.J. A comparison of magnetic and electrical stimulation of spinal nerves. Muscle Nerve, 1990, 13: 414-420. Feasby, T.E., Brown, W.F., Gilbert, J.J. and Hahn, A.F. The pathological basis of conduction block in human neuropathies. J Neurol. Neurosurg. Psychiatry, 1985, 48: 239-244. Fisher, M.A. F response latency determination. Muscle Nerve, 1982, 5: 730-734. Gutman, L. Facial and limb myokymia. AAEE Minimonograph 37. Muscle Nerve, 1991, 14: 1043-1049. Harper, C.M., Thomas, J.E., Cascino, T.L. and Litchy, W.J. Distinction between neoplastic and radiation-induced brachial plexopathy, with emphasis on the role of EMG. Neurology, 1989, 39: 502-506. Kaji, R. and Kimura, J. Nerve conduction block. Curt. Opin. Neurol. Neurosurg., 1991, 4: 744-748. Kimura, J. Electrodiagnosis in Disease of Nerve and Muscle: Principles and Practice, 2nd edn. Davis, Philadelphia, 1989. King, P.J.L. and Chiappa, K.H. Motor evoked potentials. In: K.H. Chiappa (Ed.), Evoked Potentials in Clinical Medicine, 2nd edn. Raven Press, New York, 1990, pp. 509-561. Kori, S.H., Foley, K.M. and Posner, J.B. Brachial plexus lesions in patients with cancer: 100 cases. Neurology, 1981, 31: 45-50. Lederman, R.J. and Wilbourn, A.J. Brachial plexopathy: recurrent cancer or radiation? Neurology, 1984, 34: 1331-1335. Lewis, R.A., Sumner, A.J., Brown, M.J. and Asbury, A.K. Multifocal demyelinating neuropathy with persistent conduction block. Neurology, 1982, 32: 958-966. Mondrup, K., Olsen, N.K., Pfeiffer, P. and Rose, C. Clinical and electrodiagnostic findings in breast cancer patients with radiation-induced brachial plexus neuropathy. Acta Neurol. Scand., 1990, 81: 153-158. Ongerboer de Visser, B.W., Shimsheimer, R.J. and Hart, A.A. The H reflex of the flexor carpi radialis muscle; a study in controls and

radiation induced brachial plexus lesions. J. Neurol. Neurosurg. Psychiatry, 1984, 47: 1098-1101. C)ge, A.E., Yazici, J., Boyaclyan, A., ~elik, M., Konyalio~lu, R. and Baslo, A. Magnetic stimulation in hemifacial spasm and post facial palsy synkinesis. Muscle Nerve, 1993, 16:1154-1160. Oge, A.E., Yazici, J., Boyaclyan, A., Eryildiz, D., (~rnek, |., Konyalio~lu, R., Cengiz, S., Ok~ak, O.Z., Asar, S. and Baslo, A. Peripheral and central conduction in n-hexane polyneuropathy. Muscle Nerve, 1994, 17: 1416-1430. Rhees, E.K., England, J.D. and Sumner, A.J. A computer simulation of conduction block: effects produced by actual block versus interphase cancellation. Ann. Neurol., 1990, 28: 146-156. Rrsler, K.M., Hess, C.W. and Schmid, V.D. Investigation of facial motor pathways by electrical and magnetic stimulation: sites and mechanisms of excitation. J. Neurol. Neurosurg. Psychiatry, 1989, 52:11491156. Roth, G. and Magistris, M.R. Neuropathies with prolonged conduction block, single and grouped fasciculations, localized limb myokymia. Electroenceph. clin. Neurophysiol., 1987, 67: 428-438. Roth, G., Rohr, J., Magistris, M.R. and Ochsner, F. Motor neuropathy with proximal multifocal persistent conduction block, fasciculations and myokymia. Eur. Neurol., 1986, 25: 416-423. Roth, G., Magistris, M.R., Le Fort, D., Desjacques, P. and Della Santa, D. Plexopathie brachiale post-radique. Blocs de conduction persistants. Ddscharges myokymiques et crampes. Rev. Neurol., 1988, 144(3): 173-180. Schmid, V.D., Walker, G., Hers, C.W. and Schmid, J. Magnetic and electrical stimulation of cervical motor roots: technique, site and mechanisms of excitation. J. Neurol. Neurosurg. Psychiatry, 1990, 53: 770-777. Schriefer, T.N., Mills, K.R., Murray, N.M.F. and Hers, C.W. Evaluation of proximal facial nerve conduction by transcranial magnetic stimulation. J. Neurol. Neurosurg. Psychiatry, 1988, 51: 60-66. Strhr, M. Special types of spontaneous electrical activity in radiogenic nerve injuries. Muscle Nerve, 1982, 5: 578-583. Stoll, B.A. and Andrews, J.T. Radiation-induced peripheral neuropathy. Br. Med. J., 1966, 1: 834-837. Thomas, J.E. and Colby, Jr., M.Y. Radiation-induced or metastatic brachial plexopathy? A diagnostic dilemma. J. Am. Med. Assoc., 1972, 222: 1392-1395. Thomas, J.E., Cascino, T.L. and Earle, J.D. Differential diagnosis between radiation and tumor plexopathy of the pelvis. Neurology, 1985, 35: 1-7. Ugawa, Y., Rothwell, J.C., Day, B.L., Thompson, P.D. and Marsden, C.D. Magnetic stimulation over the spinal enlargements. J. Neurol. Neurosurg. Psychiatry, 1989, 52: 1025-1032. Wilbourn, A.J. Electrodiagnosis of plexopathies. Neurologic Clinics, 1985, 3: 511-529. Wilbourn, A.J. Brachial plexus disorders. In: P.J. Dyck and P.K. Thomas (Eds.), Peripheral Neuropathy, 3rd edn. Saunders, Philadelphia, 1993, Vol. 2, pp. 911-950.