n
d
a
m
Clinical and electrodi.agnostic hndin-@ in breast cancer patients with radiation .-induced brachial plexus neuropathy I
I
Mondrup K, Olsen NK, Pfeiffer P, Rose C. Clinical and electrodiagnostic findings in breast cancer patients with radiation-induced brachial plexus neuropathy. Acta Neurol Scand 1990: 81: 153-158. The clinical and neurophysiological characteristics of radiation-induced brachial plexopathy (RBP) were assessed in 79 breast cancer patients without signs of recurrent disease at least 60 months after radiotherapy (RT). Clinically, 35% (95% confidence limits: 25-47%) had RBP. Fifty percent (31-69%) had affection of the entire plexus, 18% (7-36%) of the upper trunk only, and 4% (1-18%) of the lower trunk. In 28% (14-48 %), assessment of a definite level was not possible. In most, symptoms began during or immediately after RT, thus being without significant latency. Numbness or paresthesias (71 %, 52-86%) and pain (43%, 25-62%) were the most prominent symptoms, while the most prominent objective signs were decreased or absent muscle stretch reflexes (93 %, 77-99%) closely followed by sensory loss (82%, 64-93%) and weakness (71 %, 52-86%). Neurophysiological investigations were carried out in 46 patients (58 %). The most frequent abnormalities in patients with RBP were signs of chronic partial denervation with increased mean duration of individual motor unit potentials, and decreased amplitude of compound muscle and sensory action potentials. Nerve conduction velocities were normal.
Neurological symptoms and signs of brachial dysfunction may develop in breast cancer patients treated with mastectomy and postoperative radiotherapy (RT) involving the brachial plexus. This may be due to unrelated paralytic or acute brachial neuritis, trauma to the plexus during surgery or anesthesia, postmastectomy edema, carcinomatous involvement of the plexus (NBP), or radiationinduced brachial plexopathy (RBP). Dysfunction of the plexus due to NBP or RBP are by far the most common causes. The clinical description of RBP is hampered by the heterogeneity of the patient materials and great variation in radiation doses [ 1-91. In order to purge the description of RBP for these variables and attain more reliable figures of its clinical characteristics, we conducted a neurological follow-up investigation in 79 patients who received standardized postoperative RT for breast cancer in a single department [ 101. Few reports describe the neurophysiologic abnormalities in patients with RBP, and details of electrodiagnostic findings in these patients are sparse [4, 6, 11-14]. Therefore, we report the results of detailed neurophysiologic studies in 46 of the 79 patients.
1
K. Mondrup’, N. K. Olsen’, P. Pfeiffer’, C. Rose’
’
Departments of Neurology and Clinical Neurophysiology, Oncology R, Odense University Hospital, Denmark
Key words: breast cancer; brachial plexopathy; electromyography; postoperative radiotherapy Kurt Mondrup, Department of Neurology, Odense University Hospital, DK-5000 Odense C, Denmark
I
Accepted for publication September 8, 1989
Material and methods
Since 1977, the Danish Breast Cancer Cooperative Group (DBCG) has conducted nationwide trials in early breast cancer [ 15-18]. In the DBCG 77 protocols, patients with invasive breast cancer without evidence of distant metastases were treated primarily with total mastectomy plus axillary node sampling. All patients with axillary lymph node metastases, and/or primary tumor > 5 cm, and/or proven involvement of skin, and/or deep invasion of the fascia had postoperative radiotherapy [ 16, 181. RT was administered using a standardized technique (Table 1) and the intended dosage was 1,345 ret (rad equivalent therapy, NSD) [lo, 181. The patients were irradiated with one anterior photon field (8 meV) against the supraclavicular/infraclavicular and axillary regions combined with anterior electron field against the chest wall. Starting concurrently with RT, premenopausal patients were allocated at random to receive either no further therapy, C (cyclophosphamide), or CMF (cyclophosphamide, methotrexate, and 5-fluor153
Mondrup et al.
ouracil). Postmenopausal patients received either no drug therapy or tamoxifen. Medication was to be continued for 48 weeks [ 101. After end of treatment all patients were followed at regular intervals by clinical examinations and chest x-rays. Other relevant investigations were performed if the patients were suspected for recurrent disease. Operative, histopathological and clinical data have been obtained from the DBCG data base. To ensure that any recorded neuropathy actually was induced by RT, only patients without evidence of local or distant recurrent disease at least 5 years after the primary therapy were included in the present investigation. Furthermore, patients were excluded if symptoms or objective signs preceded RT, if they had symptoms or objective signs of root lesions, of peripheral polyneuropathy or of damage confined to the territory of a single peripheral nerve, or if informed consent was not given. The study was approved by the local Ethics Committee. The investigation was limited to patients aged 70 years or less at the time of examination. In this age group, 204 patients from the county of Funen, Denmark, received postoperative RT between August 1977 and November 1982 at the Department of Oncology, Odense University Hospital. Of these, 104 had no sign or history of recurrent breast cancer. Fourteen patients were lost to follow-up. Of the remaining 90 patients, 79 fulfilled the inclusion criteria. The characteristics of these are shown in Table 1. All neurologic examinations were made by one of the authors (NKO) between May and November 1988, after a median follow-up period of 94 months (range 67-130 months). After anamnestic infor-
Table 1. Characteristics of 79 patients with breast cancer examined for radiationinduced brachial plexopathy (REP). Median values are given with ranges in brackets Age at time of radiotherapy (years) Age at follow-up (years) Disease-free survival (months) Menopausal status at time of radiotherapy (number of patients) Premonpausal Postmenopausal Radiotherapy Total dose (Gy) Peak values (Gy) Number of fractions Duration of treatment (days) Cumulative radiation effect (ret) Adjuvant therapy (number of patients) CMF Cyclophosphamide Tamoxifen No adjuvant treatment Number of lymphnodes removed at operation Size of primary tumor (mm)
47 55 94
58 21 36.60 43.36 12 40 1345 24 24 9 22 6 30
CMF: cyclophosphamide t methotrexate t 5-fluorouracil
154
(33-631 (41-70) (67-130)
(36.60-39.60) (40.80-49.56) (12-1 3) (37-49)
(0-171 (10-80)
mations, a general neurological examination was performed with special attention to symptoms and signs from the upper limbs. Patients were considered to have RBP if paresthesias, pain or weakness developed in the ipsilateral arm and if they had welldocumented hypesthesia, hypalgesia, decreased or absent muscle stretch reflexes, decreased strength or atrophy. RBP was defined as upper plexopathy, when the segments C5-C7 were involved and as lower plexopathy, when the disorder was limited to the segments C8-T1. It was considered universal when the affection included both levels. Lymphedema was defined as a difference of at least 2.0 cm between the circumference of the 2 arms measured 10cm above the medial humeral epicondyle. Forty-six patients (58 %) accepted to participate in the neurophysiological part of the study. The characteristics of these patients did not differ from the characteristics of the total group of patients (Table 1). The neurophysiological investigations were carried out by one of the authors (KM) using standard techniques of our neurophysiological laboratory [ 19, 201. The deltoideus, biceps brachii, and abductor digiti quinti muscles were examined by needle electromyography (EMG) with measurements of the mean duration of motor unit potentials (MUP’s) using a concentric needle electrode. Distal motor latencies (DML‘s) to the deltoideus and biceps brachii muscles were determined by percutaneously, supramaximally stimulation of the plexus at Erb’s point and recording of the compound muscle action potential with concentric needle electrode. DML to m. abductor digiti quinti was determined by stimulation at the wrist with near nerve monopolar needle electrodes. Furthermore, the amplitude of the compound muscle action potential (CMAP) was measured. Sensory nerve conduction velocity (SNCV) and amplitude of sensory nerve action potentials (SNAP’S) were recorded orthodromically from the ulnar nerve. The fifth finger was stimulated by surface electrodes with recording at the wrist using the above mentioned near nerve monopolar needle electrodes. Statistical evaluations were made by Fisher’s exact test and two-sample t-test. Two-sided P-values less than 0.05 were considered to be statistically significant. All median values are followed by range in brackets, and relative risks and frequencies expressed in percentages are followed by 95% confidence limits (95% CL). Results
The registered symptoms and objective clinical signs are shown in Table 2. The findings in the smaller,
RBP & breast cancer Table 2. Symptoms and objective signs given in percents with 95% confidence limits in brackets REP-patients
Symptoms Parestesias/numbness Pain Weakness Objective signs Decreased/absent MSR H ypesthesia/ hypalgesia Weakness Muscle atrophy Lymphedema Veg. dis. Horner‘s syndrome
IN = 28)
Total material (N = 79)
71 (52-861 43 (25-62) 39 (22-59)
29 (20-401 1 5 ( 9-25) i 4 ( 8-23]
93 (77-991 a2 (64-93) 71 (52-861 29 (14-48) 21 ( 9-40) 4 ( 0-18) O ( 0-12)
33 (23-44) 29 (20-40) 25 (17-36) i o i 5-18) 22 (14-32) 1 ( 0- 6) O ( 0- 41
MSR: Muscle stretch reflexes. Veg. dis.: vegetative disturbances
neurophysiologically investigated group of patients did not differ from the findings in the entire group of patients. Twenty-eight patients (35%, 25-47%) were found to have symptoms and objective signs consistent with RBP. Fifteen patients (19%, 11-29%) were physically disabled in their daily activities after RT (severe RBP). Thirteen (16%, 9-262) had RBP without disability in daily life (less severe RBP). It appears from Table 3, that 50% (31-69%) of all patients with RBP had affection of the entire plexus, while the lesion was limited to the upper plexus in 18% (7-36%), and to the lower plexus in 4% (1-18%). In 28 % (14-48 %), the clinical presentation did not allow any definite statement of the affected level. Only one patient complained of vegetative disturbances, and Horner’s syndrome was not observed in any patient. Lymphedema was present in 22% (14-32%) of all patients. There was no preponderance of lymphedema in patients with RBP compared to patients without RBP (P = 0.90). Median age at the time of RT was 47 years. Patients belonging to the younger group developed RBP more often than patients in the older age group. Thus, in patients 47 years of age or less 29% (17-45%) developed severe RBP, in contrast to patients 48 years of age or more where only 8% (2-21%) developed severe RBP (P = 0.02). When the analysis was done for patients with RBP regardTable 3. Localization of radiation-induced brachial plexopathy (RBP) given in percents with 95% confidence limits in brackets
less of severity, the difference failed to reach statistical significance (P = 0.06). In patients with severe RBP, the interval from the last dose of radiotherapy to the first symptom of plexus disorder (latency) was well-defined in all patients. The median latency was 0 months (mean 7.5 months, range 0-59 months). In patients with less severe RBP well-defined latencies were given in 7 patients (median 1 month, mean 10.7 months, range 0-44 months). In both groups, patients had persistent symptoms of brachial plexopathy for 5 years or more with the exception of 1 patient who had symptoms for 18 months after a latency period of 48 months. Forty-eight patients (62%) received chemotherapy. Higher incidence of RBP was found in patients treated with adjuvant chemotherapy. In the 15 patients with severe RBP, 12 received postoperative adjuvant chemotherapy in addition to RT, whereas in the 13 patients with less severe RBP, 8 patients received chemotherapy (P = 0.10). The same tendencies were found when patients receiving CMF were compared with patients, who did not receive chemotherapy. RBP was found in 2 of 9 (22%, 3-60%) postmenopausal patients treated with adjuvant tamoxifen. The median number of lymphnodes removed at axillary sampling was 6 and the risk of RBP was evaluated in relation to number of nodes found. In the patients who had 6 or less lymphnodes removed, Table 4. Neurophysiological findings in patients with (tRBP) and with-out (-REP) radiation-induced brachial plexopathy. Mean values and standard deviations are given
t RBP (N = 17)
- RBP N = 29)
Age at follow-up (years)
52.2 f 6.5
59.5 i.7.9 * *
MUP (msec) Deltoideus Biceps brachii Abductor dig. quinti
13.2i 1.0 13.1 f 1.4 11.3f 1.2
12.1 i.0.7 ** 11.9 1- 0.7 *** 10.3 f 0.6 *
DML (msecl Deltoideus Biceps brachii Abductor dig. quinti
4.1 i 0 . 5 4.4 0.4 2.5 f 0.3
4.1 i.0.4 4.5 i.0.5 2.6 f 0.3
CMAP (mV) Deltoideus Biceps brachii Abductor dig. quinti
5.5 f 3.7 6.8t4.i 13.2 f 6.5
5.8 i: 3.7 5.0 f 3.2 11.6 t 4 . 2
Ulnar SNCV (mlsec)
54.3 f 4.7
55.7 i. 3.4
Ulnar SNAP (microV1
17.3 i 8.8
16.5 f 7.4
~~
RBP-patients (N = 28)
Total material (N = 791 ~
Entire plexus Upper trunk Lower trunk Undefinable
50 (31-69) 18 i. 7-36) 4 ( 1-18] 28 (14-48)
~~
18 (11-27) 6 ( 3-14) 1 ( 0- 6) l o ( 5-18)
MUP: Motor unit potential. DML: Distal motor latency. CMAP: amplitude of compound muscle action potential. SNCV: sensory nerve conduction velocity. SNAP: amplitude of sensory action potential.
* p < 0.02
** p < 0.005
*** p < 0.002 155
Mondrup et al. 25% (13-40%) had RBP. The risk of RBP was as high as 48 % (3 1-64%) in patients having more than 6 lymphnodes removed (P = 0.05). We found no correlation between RBP and tumor size or number of tumor infiltrated lymphnodes. The results of the neurophysiological investigations are shown in Table 4. Significant abnormalities were found in patients with clinical symptoms and objective signs of RBP. Thirteen patients (76%, 51-93%) had increased mean duration of MUP’s in one or more of the investigated muscles, whereas no significant changes were found in DML’s, SNCV’s, and SNAP’S,with the exception of 2 patients (12%, 2-36%) who had decreased amplitude of SNAP’S. Decreased amplitude of CMAP’s were found in 9 patients (53%, 28-77%) in one or more of the investigated muscles. The findings in patients without clinical symptoms and objective signs of RBP did not differ from the normal values used in our laboratory with the exception of 4 patients (14%, 4-31 %) who had slightly increased mean duration of MUP’s, possibly consistent with subclinical RBP. We did not register any fasciculations or grouped repetitive motor unit discharges (myok ymia). Discussion
In the present homogeneous group of 79 breast cancer patients treated with simple mastectomy, axillary lymph node sampling and postoperative standardized RT 35% (25-47%) developed clinical symptoms and objective signs consistent with RBP. Stoll & Andrews [7] found a 15% incidence in breast cancer patients, who received a minimum of 51.00 Gy (1950 ret) at the plexus level [ 171, but a 73% incidence when 55.00 Gy (2100 ret) was given. Notter et al. [21] found an incidence of 17% in breast cancer patients treated with at least 45.00 Gy in 8-10 fractions. Compared with these figures, the incidence of RBP in our study, where only 1345 ret was applied, was surprisingly high. Regrettably, in these studies, the criteria for patient selection were not stated. All patients in the above mentioned studies received large single doses of RT (4.00 to 6.00 Gy) and RT was often given three times a week, but still the incidences were equal to or lower than the figures in the present study. So, large single doses in a short time could not explain our high figures, as we used daily dose of 3.05 Gy twice a week. A major difference between our study and the others was that our patients had adjuvant chemotherapy as part of the postoperative treatment. As in the development of lymphedema and other complications [ 18,221 adjuvant chemotherapy might be a contributing factor in the development of RBP. 156
However, this could not be proved from our material (P = 0.10). Nevertheless, chemotherapy, when given in combination with RT, might have a superimposed deteriorating influence on peripheral nerve function. This is in accordance with Salner et al. [23], who found a significantly increased incidence of reversible brachial plexopathy in patients treated with adjuvant chemotherapy. In our study, the incidence of severe RBP was significantly higher in the younger age group. This could be due to the fact that adjuvant chemotherapy was given to the younger patients. Numbness or paresthesias was the most prominent symptom in our RBP patients confirming the findings of others [2, 4-6, 81. The most prominent objective sign was found to be decreased or absent muscle stretch reflexes closely followed by sensory loss and weakness. This is in accordance with the findings of Lederman & Wilbourn [ 61 and Thomas & Colby [ 81. Numerous reports in the literature have discussed differential diagnosis between RBP and NBP [ 1, 4-6, 81. Among the neurologic symptoms only one seems potentially useful in the differential diagnosis. Thus, Harper et al. [4], Kori et al. [5] and Thomas & Colby [ 81 concluded that early, severe pain suggests tumor infiltration of the plexus. However, pain may be prominent in RBP [ 1, 2, 4, 6-9, 241. Most patients in our material had diffuse affection of the brachial plexus. This observation is in accordance with Harper et al. [4], Lederman & Wilbourn [6], Stoll & Andrews [7], and Thomas & Colby [8], who concluded that the level of the brachial lesion did not allow any distinction between RBP and NBP. Kori et al. [5], however, did observe such a distinction and concluded “that painless upper trunk lesion with lymphedema suggests radiation injury, and painful lower trunk lesion with Horner’s syndrome implies tumor infiltration”. Some of these differences might be explained by the great numbers of tumors other than breast cancer in the series of Kori et al. Our study was restricted to patients clinically known to be without recurrent disease, and we can therefore not make any distinction between RBP and NBP. Horner’s syndrome was not observed in any patient in our material, consistent with the experience of others [ 5 , 61. Another side effect of RT is the development of lymphedema. The frequency of lymphedema in our material (22%, 14-322) agrees well with that found by Harper et al. [4] and by Ryttov et al. [25]. The latency period of RBP found in our material was shorter than in previous reports [ 1, 2, 5, 7, 8,]. Thus, Kori et al. [ 51 and Thomas & Colby [ 81 found average latencies of 3.5 years (3 months to 14 years) and 6 years ( 5 months to 20 years), respectively. However, Bagley et al. [ I ] concluded that when
RBP & breast cancer neurologic symptoms appear within one year after RT, the diagnosis is probably RBP. In contrast to this, Basso-Ricci et al. [2] concluded that postirradiation lesions almost always appear more than one year after RT. The neurophysiological investigations revealed significant abnormalities in our patients. The most frequent abnormalities were signs of chronic partial denervation with changes in MUP’s with increased mean duration of individual MUP’s, and decreased amplitude of CMAP’s and SNAP’S. Nerve conduction velocities were normal with normal DML‘s and normal SNCV’s. These changes are in accordance with the findings revealed by other investigators [4, 6, 11-14]. In contrast to several of these investigators, we were not able to demonstrate the presence of fasciculations or myokymic discharges in patients with RBP. Harper et al. [4] found that myokymic discharges were the only electromyographic finding useful in differentiating RBP from NBP. They emphasized, however, the need for caution in using this as an absolute criterion. The findings in our study strongly support this caution. Furthermore, Harper et al. [4] found that nerve conduction studies were not helpful in distinguishing RBP from NBP. The most helpful electrodiagnostic test complementing the clinical findings in RBP patients are, in our experience, needle EMG with measurements of mean duration of individual MUP’s, whereas conventional nerve conduction velocity measurements are of little clinical relevance. In the experience of Harper et al. [4], abnormalities noted on extensive needle examination are more widespread than predicted by the clinical examination. This agrees well with our demonstration of patients with possible subclinical RBP and emphasizes the need for needle EMG in patients with suspected RBP. Lower incidences of lesions of the brachial plexus induced by RT may be expected in the future. Daily treatment with a maximum of 2.00 Gy is current practice in RT, thus being in accordance with the experience obtained by Basso-Ricci et a1 [2] and by Notter et al. [21]. Maybe more widespread use of needle EMG examination in breast cancer patients treated with RT could indicate future advisable maximum dose of RT. In addition, adjuvant RT might not be indicated in patients treated with radical mastectomy including axillary dissection [ 181. In conclusion, the damaging effect of RT on the brachial plexus and the value of needle EMG in patients suspected of RBP are documented. Because treatment of RBP is unsuccessful, prevention is warranted. The indication for RT must be considered thoroughly, and RT and chemotherapy must be administered sequentially and not concomitantly.
Acknowledgement
We thank the Danish Breast Cancer Group for access to the operative and histopathological data. References 1. BAGLEYFH, WALSH JW, CADY B, SALZMANFA, OBERFIELDRA, PAZIONOSAG. Carcinomatous versus radiation-induced brachial plexus neuropathy in breast cancer. Cancer 1978: 41: 2154-7. 2. BASSO-RICCI S, DELLA COSTA C, VIGANOTTI G, VENTAFRI~ V,AZANOLLA R. Report on 42 cases of postirradiation lesions of the brachial plexus and their treatment. Tumori 1980: 66: 117-22. 3. CASCINOTL, KORI S, KROLG, FOLEYKM. CT of the brachial plexus in patient with cancer. Neurology 1983: 33: 1553-7 4. HARPERCM, THOMASJE, CASCINOTL, LITCHYWJ. Distinction between neoplastic and radiation-induced brachial plexopathy, with emphasis on the role of EMG. Neurology 1989: 39: 502-6. KM, POSNERJB. Brachial plexus lesions 5. KORISH, FOLEY in patients with cancer: 100 cases. Neurology 1981: 31: 45-50. 6. LEDERMAN RJ,WILBOURN AJ. Brachial plexopathy: recurrent cancer of radiation? Neurology 1984: 34: 1331-5. 7. STOLLBA, ANDREWSJT. Radiation-induced peripheral neuropathy. Br Med J 1966: 11: 834-7. JE, COLBYMY. Radiation-induced or metastatic 8. THOMAS brachial plexopathy? A diagnostic dilemma. JAMA 1972: 222: 1392-5. 9. WESTLINGP, NORDING , HELEP. Cervikalplexuslzsioner efter postoperativ strgleterapi ved cancer mammae. Nord Med 1968: 80: 1636-7. 10. OLSENNK, PFEIFFERP, MONDRUP K, ROSEC. Radiationinduced brachial plexus neuropathy in breast cancer patients. Acta Oncol (in press). 1 1 . ALBERSJW, ALLENAA, BASTRONJA, DAUBEJR. Limb myokymia. Muscle Nerve 1981: 4: 494-504. 12. ALLENAA, ALBERSJW, BASTRONJA, DAUBEJR. Myokymic discharges following radiotherapy for malignancy. Electroenceph Clin Neurophysiol 1977: 43: 148. F, MIGNONB, ECOFFETM, OLLATH, JOUNEAU 13. CONTAMIN P. Les atteintes plexiques post-radiotherapiques. Sem Hop Paris 1978: 54: 1225-9. 14. STREIBEW, SUNSF, LEIBROCKL. Brachial plexopathy in patients with breast cancer: unusual electromyographic findings in two patients. Eur Neurol 1982: 21: 256-63. 15. ANDERSENKW, MOURIDSENHT, CASTBERGT, et al. Organization of the Danish adjuvant trials in breast cancer. Dan Med Bull 1981: 28: 102-6. 16. MOURIDSENHT, ROSE C, BRINCKERH, THORPESM, K, ANDERSEN KW. Adjuvant sysRANKF, FISCHERMAN temic therapy in high risk breast cancer: the Danish Breast Cancer Cooperative Group’s trials of cyclophosphamide or CMF in premenopausal and tamoxifen in postmenopausal patients. Rec Results Cancer Res 1984: 96: 117-128. M, BENTZENSM, JUUL CHRISTENSEN J, 17. OVERGAARD HJBLLUND MADSENE. The value of NSD formula in equation of acute and late radiation complications in normal tissue following 2 and 5 fractions per week in breast cancer patients treated with postmastectomy irradiation. Radiother Oncol 1987: 9: 1. M, CHRISTENSEN JJ, JOHANSENH, et al. 18. OVERGAARD Postmastectomy irradiation in high-risk breast cancer patients: present status of the Danish Breast Cancer Cooperative Group trials. Acta Oncol 1988: 27: 707-14.
157
Mondrup et al. 19. LUDINH-P. Electromyography in practice. New York: Thieme, 1980. 20. Electromyography, sensory and motor conduction. Findings in normal subjects. Laboratoriet for klinisk neurofysiologi, Rigshospitalet. Kerbenhavn 1975. 21. NOTTERG, HALLBERG 0,VIKTERLOF KJ. Strahlenschaden am Plexus brachidis bei Patienten mit Mammakarzinom. Strahlenther 1970: 139: 538-43. 22. DANOFFBF, GOODMAN RL, GLICK JH, HALLERDG, PAJAKTF. The effect of adjuvant chemotherapy on cosmesis and complications in patients with breast cancer treated by definite irradiation. Int J Radiat Oncol Biol Phys 1983: 9: 1625-30.
158
23. SALNERAL, BOTNICKLE, HERZOGAG, et al. Reversible brachial plexopathy following radiation therapy for breast cancer. Cancer Treat Rep 1981: 65: 798-802. 24. MUMENTHALER M. Armplexusparesen in Anschluss an Rontgenbestrahlung: Mitteilung von 8 eigenen Beobachtungen. Schweiz Med Wochenschr 1964: 94: 1069-15. 25. RYTTOVN, HOLM NV, QVIST N, BLICHER-TOFTM. Influence of adjuvant irradiation on the development of late arm lymphedema and impaired shoulder mobility after mastectomy for carcinoma of the breast. Acta Oncol 1988: 21: 667-70.
d
a
m
Clinical and electrodi.agnostic hndin-@ in breast cancer patients with radiation .-induced brachial plexus neuropathy I
I
Mondrup K, Olsen NK, Pfeiffer P, Rose C. Clinical and electrodiagnostic findings in breast cancer patients with radiation-induced brachial plexus neuropathy. Acta Neurol Scand 1990: 81: 153-158. The clinical and neurophysiological characteristics of radiation-induced brachial plexopathy (RBP) were assessed in 79 breast cancer patients without signs of recurrent disease at least 60 months after radiotherapy (RT). Clinically, 35% (95% confidence limits: 25-47%) had RBP. Fifty percent (31-69%) had affection of the entire plexus, 18% (7-36%) of the upper trunk only, and 4% (1-18%) of the lower trunk. In 28% (14-48 %), assessment of a definite level was not possible. In most, symptoms began during or immediately after RT, thus being without significant latency. Numbness or paresthesias (71 %, 52-86%) and pain (43%, 25-62%) were the most prominent symptoms, while the most prominent objective signs were decreased or absent muscle stretch reflexes (93 %, 77-99%) closely followed by sensory loss (82%, 64-93%) and weakness (71 %, 52-86%). Neurophysiological investigations were carried out in 46 patients (58 %). The most frequent abnormalities in patients with RBP were signs of chronic partial denervation with increased mean duration of individual motor unit potentials, and decreased amplitude of compound muscle and sensory action potentials. Nerve conduction velocities were normal.
Neurological symptoms and signs of brachial dysfunction may develop in breast cancer patients treated with mastectomy and postoperative radiotherapy (RT) involving the brachial plexus. This may be due to unrelated paralytic or acute brachial neuritis, trauma to the plexus during surgery or anesthesia, postmastectomy edema, carcinomatous involvement of the plexus (NBP), or radiationinduced brachial plexopathy (RBP). Dysfunction of the plexus due to NBP or RBP are by far the most common causes. The clinical description of RBP is hampered by the heterogeneity of the patient materials and great variation in radiation doses [ 1-91. In order to purge the description of RBP for these variables and attain more reliable figures of its clinical characteristics, we conducted a neurological follow-up investigation in 79 patients who received standardized postoperative RT for breast cancer in a single department [ 101. Few reports describe the neurophysiologic abnormalities in patients with RBP, and details of electrodiagnostic findings in these patients are sparse [4, 6, 11-14]. Therefore, we report the results of detailed neurophysiologic studies in 46 of the 79 patients.
1
K. Mondrup’, N. K. Olsen’, P. Pfeiffer’, C. Rose’
’
Departments of Neurology and Clinical Neurophysiology, Oncology R, Odense University Hospital, Denmark
Key words: breast cancer; brachial plexopathy; electromyography; postoperative radiotherapy Kurt Mondrup, Department of Neurology, Odense University Hospital, DK-5000 Odense C, Denmark
I
Accepted for publication September 8, 1989
Material and methods
Since 1977, the Danish Breast Cancer Cooperative Group (DBCG) has conducted nationwide trials in early breast cancer [ 15-18]. In the DBCG 77 protocols, patients with invasive breast cancer without evidence of distant metastases were treated primarily with total mastectomy plus axillary node sampling. All patients with axillary lymph node metastases, and/or primary tumor > 5 cm, and/or proven involvement of skin, and/or deep invasion of the fascia had postoperative radiotherapy [ 16, 181. RT was administered using a standardized technique (Table 1) and the intended dosage was 1,345 ret (rad equivalent therapy, NSD) [lo, 181. The patients were irradiated with one anterior photon field (8 meV) against the supraclavicular/infraclavicular and axillary regions combined with anterior electron field against the chest wall. Starting concurrently with RT, premenopausal patients were allocated at random to receive either no further therapy, C (cyclophosphamide), or CMF (cyclophosphamide, methotrexate, and 5-fluor153
Mondrup et al.
ouracil). Postmenopausal patients received either no drug therapy or tamoxifen. Medication was to be continued for 48 weeks [ 101. After end of treatment all patients were followed at regular intervals by clinical examinations and chest x-rays. Other relevant investigations were performed if the patients were suspected for recurrent disease. Operative, histopathological and clinical data have been obtained from the DBCG data base. To ensure that any recorded neuropathy actually was induced by RT, only patients without evidence of local or distant recurrent disease at least 5 years after the primary therapy were included in the present investigation. Furthermore, patients were excluded if symptoms or objective signs preceded RT, if they had symptoms or objective signs of root lesions, of peripheral polyneuropathy or of damage confined to the territory of a single peripheral nerve, or if informed consent was not given. The study was approved by the local Ethics Committee. The investigation was limited to patients aged 70 years or less at the time of examination. In this age group, 204 patients from the county of Funen, Denmark, received postoperative RT between August 1977 and November 1982 at the Department of Oncology, Odense University Hospital. Of these, 104 had no sign or history of recurrent breast cancer. Fourteen patients were lost to follow-up. Of the remaining 90 patients, 79 fulfilled the inclusion criteria. The characteristics of these are shown in Table 1. All neurologic examinations were made by one of the authors (NKO) between May and November 1988, after a median follow-up period of 94 months (range 67-130 months). After anamnestic infor-
Table 1. Characteristics of 79 patients with breast cancer examined for radiationinduced brachial plexopathy (REP). Median values are given with ranges in brackets Age at time of radiotherapy (years) Age at follow-up (years) Disease-free survival (months) Menopausal status at time of radiotherapy (number of patients) Premonpausal Postmenopausal Radiotherapy Total dose (Gy) Peak values (Gy) Number of fractions Duration of treatment (days) Cumulative radiation effect (ret) Adjuvant therapy (number of patients) CMF Cyclophosphamide Tamoxifen No adjuvant treatment Number of lymphnodes removed at operation Size of primary tumor (mm)
47 55 94
58 21 36.60 43.36 12 40 1345 24 24 9 22 6 30
CMF: cyclophosphamide t methotrexate t 5-fluorouracil
154
(33-631 (41-70) (67-130)
(36.60-39.60) (40.80-49.56) (12-1 3) (37-49)
(0-171 (10-80)
mations, a general neurological examination was performed with special attention to symptoms and signs from the upper limbs. Patients were considered to have RBP if paresthesias, pain or weakness developed in the ipsilateral arm and if they had welldocumented hypesthesia, hypalgesia, decreased or absent muscle stretch reflexes, decreased strength or atrophy. RBP was defined as upper plexopathy, when the segments C5-C7 were involved and as lower plexopathy, when the disorder was limited to the segments C8-T1. It was considered universal when the affection included both levels. Lymphedema was defined as a difference of at least 2.0 cm between the circumference of the 2 arms measured 10cm above the medial humeral epicondyle. Forty-six patients (58 %) accepted to participate in the neurophysiological part of the study. The characteristics of these patients did not differ from the characteristics of the total group of patients (Table 1). The neurophysiological investigations were carried out by one of the authors (KM) using standard techniques of our neurophysiological laboratory [ 19, 201. The deltoideus, biceps brachii, and abductor digiti quinti muscles were examined by needle electromyography (EMG) with measurements of the mean duration of motor unit potentials (MUP’s) using a concentric needle electrode. Distal motor latencies (DML‘s) to the deltoideus and biceps brachii muscles were determined by percutaneously, supramaximally stimulation of the plexus at Erb’s point and recording of the compound muscle action potential with concentric needle electrode. DML to m. abductor digiti quinti was determined by stimulation at the wrist with near nerve monopolar needle electrodes. Furthermore, the amplitude of the compound muscle action potential (CMAP) was measured. Sensory nerve conduction velocity (SNCV) and amplitude of sensory nerve action potentials (SNAP’S) were recorded orthodromically from the ulnar nerve. The fifth finger was stimulated by surface electrodes with recording at the wrist using the above mentioned near nerve monopolar needle electrodes. Statistical evaluations were made by Fisher’s exact test and two-sample t-test. Two-sided P-values less than 0.05 were considered to be statistically significant. All median values are followed by range in brackets, and relative risks and frequencies expressed in percentages are followed by 95% confidence limits (95% CL). Results
The registered symptoms and objective clinical signs are shown in Table 2. The findings in the smaller,
RBP & breast cancer Table 2. Symptoms and objective signs given in percents with 95% confidence limits in brackets REP-patients
Symptoms Parestesias/numbness Pain Weakness Objective signs Decreased/absent MSR H ypesthesia/ hypalgesia Weakness Muscle atrophy Lymphedema Veg. dis. Horner‘s syndrome
IN = 28)
Total material (N = 79)
71 (52-861 43 (25-62) 39 (22-59)
29 (20-401 1 5 ( 9-25) i 4 ( 8-23]
93 (77-991 a2 (64-93) 71 (52-861 29 (14-48) 21 ( 9-40) 4 ( 0-18) O ( 0-12)
33 (23-44) 29 (20-40) 25 (17-36) i o i 5-18) 22 (14-32) 1 ( 0- 6) O ( 0- 41
MSR: Muscle stretch reflexes. Veg. dis.: vegetative disturbances
neurophysiologically investigated group of patients did not differ from the findings in the entire group of patients. Twenty-eight patients (35%, 25-47%) were found to have symptoms and objective signs consistent with RBP. Fifteen patients (19%, 11-29%) were physically disabled in their daily activities after RT (severe RBP). Thirteen (16%, 9-262) had RBP without disability in daily life (less severe RBP). It appears from Table 3, that 50% (31-69%) of all patients with RBP had affection of the entire plexus, while the lesion was limited to the upper plexus in 18% (7-36%), and to the lower plexus in 4% (1-18%). In 28 % (14-48 %), the clinical presentation did not allow any definite statement of the affected level. Only one patient complained of vegetative disturbances, and Horner’s syndrome was not observed in any patient. Lymphedema was present in 22% (14-32%) of all patients. There was no preponderance of lymphedema in patients with RBP compared to patients without RBP (P = 0.90). Median age at the time of RT was 47 years. Patients belonging to the younger group developed RBP more often than patients in the older age group. Thus, in patients 47 years of age or less 29% (17-45%) developed severe RBP, in contrast to patients 48 years of age or more where only 8% (2-21%) developed severe RBP (P = 0.02). When the analysis was done for patients with RBP regardTable 3. Localization of radiation-induced brachial plexopathy (RBP) given in percents with 95% confidence limits in brackets
less of severity, the difference failed to reach statistical significance (P = 0.06). In patients with severe RBP, the interval from the last dose of radiotherapy to the first symptom of plexus disorder (latency) was well-defined in all patients. The median latency was 0 months (mean 7.5 months, range 0-59 months). In patients with less severe RBP well-defined latencies were given in 7 patients (median 1 month, mean 10.7 months, range 0-44 months). In both groups, patients had persistent symptoms of brachial plexopathy for 5 years or more with the exception of 1 patient who had symptoms for 18 months after a latency period of 48 months. Forty-eight patients (62%) received chemotherapy. Higher incidence of RBP was found in patients treated with adjuvant chemotherapy. In the 15 patients with severe RBP, 12 received postoperative adjuvant chemotherapy in addition to RT, whereas in the 13 patients with less severe RBP, 8 patients received chemotherapy (P = 0.10). The same tendencies were found when patients receiving CMF were compared with patients, who did not receive chemotherapy. RBP was found in 2 of 9 (22%, 3-60%) postmenopausal patients treated with adjuvant tamoxifen. The median number of lymphnodes removed at axillary sampling was 6 and the risk of RBP was evaluated in relation to number of nodes found. In the patients who had 6 or less lymphnodes removed, Table 4. Neurophysiological findings in patients with (tRBP) and with-out (-REP) radiation-induced brachial plexopathy. Mean values and standard deviations are given
t RBP (N = 17)
- RBP N = 29)
Age at follow-up (years)
52.2 f 6.5
59.5 i.7.9 * *
MUP (msec) Deltoideus Biceps brachii Abductor dig. quinti
13.2i 1.0 13.1 f 1.4 11.3f 1.2
12.1 i.0.7 ** 11.9 1- 0.7 *** 10.3 f 0.6 *
DML (msecl Deltoideus Biceps brachii Abductor dig. quinti
4.1 i 0 . 5 4.4 0.4 2.5 f 0.3
4.1 i.0.4 4.5 i.0.5 2.6 f 0.3
CMAP (mV) Deltoideus Biceps brachii Abductor dig. quinti
5.5 f 3.7 6.8t4.i 13.2 f 6.5
5.8 i: 3.7 5.0 f 3.2 11.6 t 4 . 2
Ulnar SNCV (mlsec)
54.3 f 4.7
55.7 i. 3.4
Ulnar SNAP (microV1
17.3 i 8.8
16.5 f 7.4
~~
RBP-patients (N = 28)
Total material (N = 791 ~
Entire plexus Upper trunk Lower trunk Undefinable
50 (31-69) 18 i. 7-36) 4 ( 1-18] 28 (14-48)
~~
18 (11-27) 6 ( 3-14) 1 ( 0- 6) l o ( 5-18)
MUP: Motor unit potential. DML: Distal motor latency. CMAP: amplitude of compound muscle action potential. SNCV: sensory nerve conduction velocity. SNAP: amplitude of sensory action potential.
* p < 0.02
** p < 0.005
*** p < 0.002 155
Mondrup et al. 25% (13-40%) had RBP. The risk of RBP was as high as 48 % (3 1-64%) in patients having more than 6 lymphnodes removed (P = 0.05). We found no correlation between RBP and tumor size or number of tumor infiltrated lymphnodes. The results of the neurophysiological investigations are shown in Table 4. Significant abnormalities were found in patients with clinical symptoms and objective signs of RBP. Thirteen patients (76%, 51-93%) had increased mean duration of MUP’s in one or more of the investigated muscles, whereas no significant changes were found in DML’s, SNCV’s, and SNAP’S,with the exception of 2 patients (12%, 2-36%) who had decreased amplitude of SNAP’S. Decreased amplitude of CMAP’s were found in 9 patients (53%, 28-77%) in one or more of the investigated muscles. The findings in patients without clinical symptoms and objective signs of RBP did not differ from the normal values used in our laboratory with the exception of 4 patients (14%, 4-31 %) who had slightly increased mean duration of MUP’s, possibly consistent with subclinical RBP. We did not register any fasciculations or grouped repetitive motor unit discharges (myok ymia). Discussion
In the present homogeneous group of 79 breast cancer patients treated with simple mastectomy, axillary lymph node sampling and postoperative standardized RT 35% (25-47%) developed clinical symptoms and objective signs consistent with RBP. Stoll & Andrews [7] found a 15% incidence in breast cancer patients, who received a minimum of 51.00 Gy (1950 ret) at the plexus level [ 171, but a 73% incidence when 55.00 Gy (2100 ret) was given. Notter et al. [21] found an incidence of 17% in breast cancer patients treated with at least 45.00 Gy in 8-10 fractions. Compared with these figures, the incidence of RBP in our study, where only 1345 ret was applied, was surprisingly high. Regrettably, in these studies, the criteria for patient selection were not stated. All patients in the above mentioned studies received large single doses of RT (4.00 to 6.00 Gy) and RT was often given three times a week, but still the incidences were equal to or lower than the figures in the present study. So, large single doses in a short time could not explain our high figures, as we used daily dose of 3.05 Gy twice a week. A major difference between our study and the others was that our patients had adjuvant chemotherapy as part of the postoperative treatment. As in the development of lymphedema and other complications [ 18,221 adjuvant chemotherapy might be a contributing factor in the development of RBP. 156
However, this could not be proved from our material (P = 0.10). Nevertheless, chemotherapy, when given in combination with RT, might have a superimposed deteriorating influence on peripheral nerve function. This is in accordance with Salner et al. [23], who found a significantly increased incidence of reversible brachial plexopathy in patients treated with adjuvant chemotherapy. In our study, the incidence of severe RBP was significantly higher in the younger age group. This could be due to the fact that adjuvant chemotherapy was given to the younger patients. Numbness or paresthesias was the most prominent symptom in our RBP patients confirming the findings of others [2, 4-6, 81. The most prominent objective sign was found to be decreased or absent muscle stretch reflexes closely followed by sensory loss and weakness. This is in accordance with the findings of Lederman & Wilbourn [ 61 and Thomas & Colby [ 81. Numerous reports in the literature have discussed differential diagnosis between RBP and NBP [ 1, 4-6, 81. Among the neurologic symptoms only one seems potentially useful in the differential diagnosis. Thus, Harper et al. [4], Kori et al. [5] and Thomas & Colby [ 81 concluded that early, severe pain suggests tumor infiltration of the plexus. However, pain may be prominent in RBP [ 1, 2, 4, 6-9, 241. Most patients in our material had diffuse affection of the brachial plexus. This observation is in accordance with Harper et al. [4], Lederman & Wilbourn [6], Stoll & Andrews [7], and Thomas & Colby [8], who concluded that the level of the brachial lesion did not allow any distinction between RBP and NBP. Kori et al. [5], however, did observe such a distinction and concluded “that painless upper trunk lesion with lymphedema suggests radiation injury, and painful lower trunk lesion with Horner’s syndrome implies tumor infiltration”. Some of these differences might be explained by the great numbers of tumors other than breast cancer in the series of Kori et al. Our study was restricted to patients clinically known to be without recurrent disease, and we can therefore not make any distinction between RBP and NBP. Horner’s syndrome was not observed in any patient in our material, consistent with the experience of others [ 5 , 61. Another side effect of RT is the development of lymphedema. The frequency of lymphedema in our material (22%, 14-322) agrees well with that found by Harper et al. [4] and by Ryttov et al. [25]. The latency period of RBP found in our material was shorter than in previous reports [ 1, 2, 5, 7, 8,]. Thus, Kori et al. [ 51 and Thomas & Colby [ 81 found average latencies of 3.5 years (3 months to 14 years) and 6 years ( 5 months to 20 years), respectively. However, Bagley et al. [ I ] concluded that when
RBP & breast cancer neurologic symptoms appear within one year after RT, the diagnosis is probably RBP. In contrast to this, Basso-Ricci et al. [2] concluded that postirradiation lesions almost always appear more than one year after RT. The neurophysiological investigations revealed significant abnormalities in our patients. The most frequent abnormalities were signs of chronic partial denervation with changes in MUP’s with increased mean duration of individual MUP’s, and decreased amplitude of CMAP’s and SNAP’S. Nerve conduction velocities were normal with normal DML‘s and normal SNCV’s. These changes are in accordance with the findings revealed by other investigators [4, 6, 11-14]. In contrast to several of these investigators, we were not able to demonstrate the presence of fasciculations or myokymic discharges in patients with RBP. Harper et al. [4] found that myokymic discharges were the only electromyographic finding useful in differentiating RBP from NBP. They emphasized, however, the need for caution in using this as an absolute criterion. The findings in our study strongly support this caution. Furthermore, Harper et al. [4] found that nerve conduction studies were not helpful in distinguishing RBP from NBP. The most helpful electrodiagnostic test complementing the clinical findings in RBP patients are, in our experience, needle EMG with measurements of mean duration of individual MUP’s, whereas conventional nerve conduction velocity measurements are of little clinical relevance. In the experience of Harper et al. [4], abnormalities noted on extensive needle examination are more widespread than predicted by the clinical examination. This agrees well with our demonstration of patients with possible subclinical RBP and emphasizes the need for needle EMG in patients with suspected RBP. Lower incidences of lesions of the brachial plexus induced by RT may be expected in the future. Daily treatment with a maximum of 2.00 Gy is current practice in RT, thus being in accordance with the experience obtained by Basso-Ricci et a1 [2] and by Notter et al. [21]. Maybe more widespread use of needle EMG examination in breast cancer patients treated with RT could indicate future advisable maximum dose of RT. In addition, adjuvant RT might not be indicated in patients treated with radical mastectomy including axillary dissection [ 181. In conclusion, the damaging effect of RT on the brachial plexus and the value of needle EMG in patients suspected of RBP are documented. Because treatment of RBP is unsuccessful, prevention is warranted. The indication for RT must be considered thoroughly, and RT and chemotherapy must be administered sequentially and not concomitantly.
Acknowledgement
We thank the Danish Breast Cancer Group for access to the operative and histopathological data. References 1. BAGLEYFH, WALSH JW, CADY B, SALZMANFA, OBERFIELDRA, PAZIONOSAG. Carcinomatous versus radiation-induced brachial plexus neuropathy in breast cancer. Cancer 1978: 41: 2154-7. 2. BASSO-RICCI S, DELLA COSTA C, VIGANOTTI G, VENTAFRI~ V,AZANOLLA R. Report on 42 cases of postirradiation lesions of the brachial plexus and their treatment. Tumori 1980: 66: 117-22. 3. CASCINOTL, KORI S, KROLG, FOLEYKM. CT of the brachial plexus in patient with cancer. Neurology 1983: 33: 1553-7 4. HARPERCM, THOMASJE, CASCINOTL, LITCHYWJ. Distinction between neoplastic and radiation-induced brachial plexopathy, with emphasis on the role of EMG. Neurology 1989: 39: 502-6. KM, POSNERJB. Brachial plexus lesions 5. KORISH, FOLEY in patients with cancer: 100 cases. Neurology 1981: 31: 45-50. 6. LEDERMAN RJ,WILBOURN AJ. Brachial plexopathy: recurrent cancer of radiation? Neurology 1984: 34: 1331-5. 7. STOLLBA, ANDREWSJT. Radiation-induced peripheral neuropathy. Br Med J 1966: 11: 834-7. JE, COLBYMY. Radiation-induced or metastatic 8. THOMAS brachial plexopathy? A diagnostic dilemma. JAMA 1972: 222: 1392-5. 9. WESTLINGP, NORDING , HELEP. Cervikalplexuslzsioner efter postoperativ strgleterapi ved cancer mammae. Nord Med 1968: 80: 1636-7. 10. OLSENNK, PFEIFFERP, MONDRUP K, ROSEC. Radiationinduced brachial plexus neuropathy in breast cancer patients. Acta Oncol (in press). 1 1 . ALBERSJW, ALLENAA, BASTRONJA, DAUBEJR. Limb myokymia. Muscle Nerve 1981: 4: 494-504. 12. ALLENAA, ALBERSJW, BASTRONJA, DAUBEJR. Myokymic discharges following radiotherapy for malignancy. Electroenceph Clin Neurophysiol 1977: 43: 148. F, MIGNONB, ECOFFETM, OLLATH, JOUNEAU 13. CONTAMIN P. Les atteintes plexiques post-radiotherapiques. Sem Hop Paris 1978: 54: 1225-9. 14. STREIBEW, SUNSF, LEIBROCKL. Brachial plexopathy in patients with breast cancer: unusual electromyographic findings in two patients. Eur Neurol 1982: 21: 256-63. 15. ANDERSENKW, MOURIDSENHT, CASTBERGT, et al. Organization of the Danish adjuvant trials in breast cancer. Dan Med Bull 1981: 28: 102-6. 16. MOURIDSENHT, ROSE C, BRINCKERH, THORPESM, K, ANDERSEN KW. Adjuvant sysRANKF, FISCHERMAN temic therapy in high risk breast cancer: the Danish Breast Cancer Cooperative Group’s trials of cyclophosphamide or CMF in premenopausal and tamoxifen in postmenopausal patients. Rec Results Cancer Res 1984: 96: 117-128. M, BENTZENSM, JUUL CHRISTENSEN J, 17. OVERGAARD HJBLLUND MADSENE. The value of NSD formula in equation of acute and late radiation complications in normal tissue following 2 and 5 fractions per week in breast cancer patients treated with postmastectomy irradiation. Radiother Oncol 1987: 9: 1. M, CHRISTENSEN JJ, JOHANSENH, et al. 18. OVERGAARD Postmastectomy irradiation in high-risk breast cancer patients: present status of the Danish Breast Cancer Cooperative Group trials. Acta Oncol 1988: 27: 707-14.
157
Mondrup et al. 19. LUDINH-P. Electromyography in practice. New York: Thieme, 1980. 20. Electromyography, sensory and motor conduction. Findings in normal subjects. Laboratoriet for klinisk neurofysiologi, Rigshospitalet. Kerbenhavn 1975. 21. NOTTERG, HALLBERG 0,VIKTERLOF KJ. Strahlenschaden am Plexus brachidis bei Patienten mit Mammakarzinom. Strahlenther 1970: 139: 538-43. 22. DANOFFBF, GOODMAN RL, GLICK JH, HALLERDG, PAJAKTF. The effect of adjuvant chemotherapy on cosmesis and complications in patients with breast cancer treated by definite irradiation. Int J Radiat Oncol Biol Phys 1983: 9: 1625-30.
158
23. SALNERAL, BOTNICKLE, HERZOGAG, et al. Reversible brachial plexopathy following radiation therapy for breast cancer. Cancer Treat Rep 1981: 65: 798-802. 24. MUMENTHALER M. Armplexusparesen in Anschluss an Rontgenbestrahlung: Mitteilung von 8 eigenen Beobachtungen. Schweiz Med Wochenschr 1964: 94: 1069-15. 25. RYTTOVN, HOLM NV, QVIST N, BLICHER-TOFTM. Influence of adjuvant irradiation on the development of late arm lymphedema and impaired shoulder mobility after mastectomy for carcinoma of the breast. Acta Oncol 1988: 21: 667-70.
