Scalf et al.

MRI of Patients with Parsonage-Turner Syndrome

M u s c ul o s kel et a l I m ag i n g C l i n i c a l O b s e r va t i o n s

MRI Findings of 26 Patients with Parsonage-Turner Syndrome

Richard E. Scalf1 Doris E. Wenger1 Matthew A. Frick1 Jayawant N. Mandrekar2 Mark C. Adkins1
Scalf RE, Wenger DE, Frick MA, Mandrekar JN, Adkins MC OBJECTIVE. The objective of our study was to describe the MRI features of patients with Parsonage-Turner syndrome. Familiarity with the MRI features associated with this entity is important because radiologists may be the first to suggest the diagnosis. Twenty-six patients with Parsonage-Turner syndrome were treated at our institution between 1997 and 2005. We retrospectively reviewed the MR images of patients with clinical or electromyographic evidence (or both) of acute brachial neuritis without a definable cause. CONCLUSION. MRI of the brachial plexus and shoulder in patients with Parsonage-Turner syndrome showed intramuscular denervation changes involving one or more muscle groups of the shoulder girdle. The supraspinatus and infraspinatus muscles were the most commonly involved. MRI is sensitive for detecting signal abnormalities in the muscles of the shoulder girdle of patients with Parsonage-Turner syndrome. MRI may be instrumental in accurately diagnosing the syndrome. arsonage-Turner syndrome, also known as acute idiopathic brachial neuritis, is a painful nontraumatic disorder involving the shoulder girdle. Patients with Parsonage-Turner syndrome typically present with a sudden onset of shoulder pain or weakness (or both) of the shoulder girdle musculature [1]. Clinically, establishing the diagnosis may be challenging because symptoms are nonspecific and may mimic other shoulder girdle disorders such as labral tear with associated paralabral cyst, rotator cuff tear, impingement, and adhesive capsulitis [14]. Evaluation of patients with shoulder pain and weakness typically includes a medical history, physical examination, imaging studies, and possibly electrophysiologic evaluation. MRI is the imaging technique of choice in patients with shoulder pain and weakness and provides the most comprehensive evaluation of the shoulder girdle because of its multiplanar imaging capability and superior soft-tissue contrast. Familiarity with the MRI features of Parsonage-Turner syndrome is critical for radiologists because they may be the first to suggest the diagnosis. In this article, we describe the MRI findings of 26 patients with Parsonage-Turner syndrome. This study and the recently published study by Gaskin and Helms [5] are the largest reported series describing the imaging features of Parsonage-Turner syndrome to date.

Materials and Methods

We retrospectively reviewed the records of all the patients from our institution who were examined between 1997 and 2005 and had clinical or electrophysiologic evidence (or both) of acute brachial plexopathy without a definable cause (e.g., appropriate history of trauma or anatomic abnormality). This study was approved by the Mayo Clinic Institutional Review Board and met the criteria established for waiver of informed consent. Patients were identified by searching our database for those who underwent MRI of the brachial plexus or shoulder. The keywords that were used in the search included Parsonage-Turner, acute brachial neuritis, and denervation. Because we searched through an MRI database, patients with the clinical diagnosis of Parsonage-Turner syndrome who did not undergo MRI evaluation were excluded from the study. In addition, after the medical records and MRI examinations were reviewed, patients were excluded if they had a history of trauma, radiation, or neoplasm or if there was MRI evidence of rotator cuff tear, bone or soft-tissue mass, labral tear associated with paralabral cyst, or other conditions that could contribute to muscle weakness. Patients with a clinical diagnosis of Parsonage-Turner syndrome and normal findings after the MRI examination were also excluded because the purpose of our study was to evaluate and report only positive MRI findings of patients with this diagnosis. Twenty-six patients met the inclusion criteria for the study.

Keywords: denervation, idiopathic brachial neuritis, MRI, musculoskeletal imaging, Parsonage-Turner syndrome, shoulder DOI:10.2214/AJR.06.1136 Received August 25, 2006; accepted after revision January 15, 2007.
1Department

of Radiology, Mayo Clinic, 200 First St., SW, Rochester, MN 55905. Address correspondence to D. E. Wenger. of Biostatistics, Mayo Clinic, Rochester, MN.

2Division

WEB This is a Web exclusive article. AJR 2007; 189:W39W44 0361803X/07/1891W39 American Roentgen Ray Society

AJR:189, July 2007

W39

Scalf et al.
Fig. 1Bar graph illustrates spectrum of muscles of shoulder girdle and chest wall that were affected in patients with Parsonage-Turner syndrome and shows percentage of patients with involvement of each muscle. Supraspinatus and infraspinatus muscles were most commonly affected muscle groups, which explains why presenting symptoms mimicked rotator cuff tendon tears or mass in spinoglenoid notch that impinged on suprascapular nerve.

90 80 70

60 50 40 30 20 10 0
in or el to id us us is lis ap ul ar sp in at pi na t D to ra M im us ho m bo id s D or si

Affected Muscles

All MRI examinations were performed using a 1.5-T scanner (Signa, GE Healthcare). Each patient underwent MRI of the brachial plexus or shoulder, depending on the clinical suspicion of the cause of the pathologic process. Brachial plexus examinations included axial, oblique sagittal, and oblique coronal T1-weighted fast spin-echo images (TR/TE minimum, 400/600; field of view [FOV], 34 cm; slice thickness, 6 mm; intersection gap, 1 mm; matrix, 256 256) and oblique sagittal and oblique coronal T2-weighted fast spin-echo inversion recovery images (TR/TE, 3,200/68; inversion recovery time, 130 milliseconds; FOV, 2630 cm; slice thickness, 6 mm; intersection gap, 2 mm; matrix, 256 192). All brachial plexus images were obtained using a body or torso coil. Shoulder examinations were performed using a shoulder coil and included axial, oblique coronal, and oblique sagittal proton density images (3,200/32; FOV, 14 cm; slice thickness, 4 mm; intersection gap, 0.5 mm; matrix, 256 256). Axial, oblique coronal, and oblique sagittal fat-suppressed T2-weighted images (3,550/50; FOV, 14 cm; slice thickness, 4 mm; intersection gap, 0.5 mm; matrix, 256 256) were also obtained as a part of the shoulder imaging protocol. Neither intraarticular nor IV gadolinium was used for any of the examinations. All imaging studies were reviewed by consensus of three musculoskeletal radiologists. Each reviewer had completed a 1-year musculoskeletal fellowship, and two reviewers had 14 and 10 years of additional experience as musculoskeletal radiologists. Positive MRI findings were defined as features of denervation change that involved muscles of the rotator cuff or shoulder girdle, including intramuscular edema, muscular atrophy with or without fatty infiltration, or a combination of these findings. Intramuscular edema was defined as any area

of increased intramuscular signal on T2-weighted images. Muscular atrophy with fatty infiltration was defined as decreased muscle mass and linear foci of increased T1-weighted or proton density intramuscular signal when compared with the relative volume and signal intensity of adjacent normal muscle groups. Fatty infiltration was identified on proton density sequences as linear foci of increased intramuscular signal that also showed decreased signal on corresponding fat-saturated T2-weighted or inversion recovery sequences. The latter two T2weighted imaging sequences in combination were interpreted as the equivalent of T1-weighted images for recognition of fatty infiltration. All MRI data were correlated with available initial and follow-up clinical information. Statistical analyses were primarily descriptive. Categoric variables were summarized using frequencies and percentages, and continuous variables were summarized using medians and ranges because of skewed distribution. Associations between categoric variables (e.g., the affected side and the dominant handedness of the patient) were assessed using the chi-square test or the Fishers exact test as appropriate. Statistical tests were two sided, and p values less than 0.05 were considered statistically significant. All statistical analyses were performed using a commercial software package (version 8.0 SAS, SAS Institute).

Results We retrospectively identified 26 patients with clinical or electrophysiologic evidence (or both) of Parsonage-Turner syndrome who had positive MRI findings. All patients had clinical findings that were compatible with Parsonage-Turner syndrome. Nineteen of the 26 patients were evaluated using electromyo-

graphy (EMG) and had positive findings for Parsonage-Turner syndrome. Seven patients did not undergo EMG, but each had strong clinical evidence that supported the diagnosis of Parsonage-Turner syndrome and excluded other causes. The study patients included 20 males (77%) and six females (23%). The median age at diagnosis was 45 years (range, 877 years). No statistically significant correlation was identified between the affected side and the dominant handedness of the patient. Hand dominance was recorded for 20 patients: 17 (85%) were right-handed and three (15%), left-handed. Clinical symptoms were on the right side for 12 patients (46%), on the left side for 12 patients (46%), and bilateral for two patients (8%). Clinical symptoms were variable: Most patients presented with isolated pain, weakness, or numbness (or a combination of the three). The most common clinical presentation was pain and weakness (11 patients [42%]), and the second most common was pain alone (six patients [23%]). A small number of patients presented with pain, weakness, and numbness (four patients [15%]) or weakness alone (three patients [12%]). One patient presented with pain and numbness and another with weakness and numbness. None of the patients was asymptomatic. Patients presented with symptoms in various muscles of the shoulder girdle, including the supraspinatus, infraspinatus, deltoid, teres minor, subscapularis, latissimus dorsi, pectoralis, and rhomboids. Review of the MRI examinations showed that the supraspinatus and infraspinatus muscles were affected most frequently (Fig. 1). Most patients (88%) presented with more than one affected muscle group. Three or more muscles were affected in 17 patients (65%), two muscles in six patients (23%), and only one muscle in three patients (12%). Muscles were evaluated for evidence of denervation change in the shoulder girdle and chest wall. Each muscle was evaluated for the presence of edema and atrophy (Table 1). The supraspinatus and infraspinatus muscles were the most commonly affected muscles in our series. Of the patients with a detectable signal abnormality in the supraspinatus and infraspinatus muscles, one third showed MR evidence of only intramuscular edema and no evidence of atrophy or fatty infiltration on T1-weighted or nonfat-saturated proton density images (Fig. 2). Just over one half showed marked diffuse increased T2-weighted signal with atrophy and fatty infiltration on T1-weighted or nonfatsaturated proton density images (Fig. 3).

Patients (%)

es

Su bs c

Su pr a

In fr

Te r

Pe c

as

W40

La t

is s

AJR:189, July 2007

MRI of Patients with Parsonage-Turner Syndrome TABLE 1: MRI Findings of 26 Patients with Parsonage-Turner Syndrome
Affected Muscles Supraspinatus Infraspinatus Deltoid Teres minor Subscapularis Latissimus dorsi Pectoralis Rhomboids No. (%) of Patients Edema Only 9 (35) 8 (31) 5 (19) 4 (15) 3 (12) 1 (4) 0 (0) 1 (4) Atrophy Only 0 (0) 0 (0) 1 (4) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) Edema and Atrophy 14 (54) 15 (58) 5 (19) 4 (15) 2 (8) 2 (8) 1 (4) 0 (0) Unaffected 3 (12) 3 (12) 15 (58) 18 (69) 21 (81) 23 (88) 25 (96) 25 (96)

The MRI findings of two patients with typical imaging features of Parsonage-Turner syndrome are illustrated in Figures 4 and 5. The images in Figure 4 were obtained of an 18year-old man who presented with shoulder pain and progressive left upper extremity weakness. The overall EMG findings were those of a severe but incomplete left suprascapular neuropathy with evidence of early reinnervation. MRI of the brachial plexus showed diffuse increased T2-weighted intramuscular signal (Figs. 4A and 4B) and atrophy of the supraspinatus and infraspinatus muscles (Fig. 4C). Examination results were otherwise unremarkable; specifically, none of the images showed evidence of a labral tear, paralabral cyst, or rotator cuff tear. The images in Figure 5 were obtained of a 52-year-old man who presented with acute onset of severe left shoulder pain. His medical history included a presumed viral infection shortly before the onset of shoulder symptoms. EMG findings were most consistent with severe axillary and suprascapular neuropathies, but other elements of the upper trunk of the brachial plexus were spared. MRI of the left shoulder showed diffuse increased T2 signal in the supraspinatus, infraspinatus, and deltoid muscles (Figs. 5A and 5B); all were associated with mild fatty atrophy. No structural abnormalities were identified that could explain the denervation change. Discussion Acute brachial neuritis was first reported by Spillane in 1943 [2], but the disease name was not established until Parsonage and Turner [1] described a series of 136 clinical cases in 1948. Other terms used to describe this disease entity include brachial plexus neuropathy, paralytic brachial neuritis, and acute brachial radiculitis [3, 4, 6, 7]. However, the general term Parsonage-Turner syndrome is most commonly used.

Although the precise cause of ParsonageTurner syndrome has not been clearly established, viral and autoimmune processes have been proposed [8, 9]. Most patients present during the third to seventh decade, but reported ages range from 3 months to 82 years [2, 3, 6, 10]. In a clinical study, Beghi et al. [10] examined a series of 99 patients and reported an incidence rate of 1.64 cases per 100,000 people. Males are predominantly affected, with male-to-female ratios ranging from 2:1 to 11.5:1 [3, 4, 6]. Our findings correlate with those of other studies that show no evidence of a tendency for the right- or leftside upper extremity or correlation with hand dominance. Bilateral involvement has been reported for as many as one third of patients [3]. Phrenic nerve involvement with diaphragmatic dysfunction has been described [3]. Treatment is palliative and includes analgesics for pain and physical therapy for weakness. Parsonage-Turner syndrome typically has a self-limited course, although low recurrence rates have been reported [3]. Most patients with Parsonage-Turner syndrome present with severe onset of nontraumatic shoulder pain with or without associated weakness, paralysis, and paresthesias [1, 2]. Medical history and physical examination findings, EMG tests, and imaging studies are used to make an accurate diagnosis. EMG and nerve conduction velocities may show changes of acute denervation in the brachial plexus distribution [10]. Previous reports have shown that the most common EMG abnormalities of Parsonage-Turner syndrome occur in the distribution of the suprascapular nerve, a branch of the superior trunk of the brachial plexus [10]. This correlates with our results, which showed that the supraspinatus and infraspinatus muscles, both of which are innervated by the suprascapular nerve, were most commonly involved.

MRI is an important imaging tool for establishing the diagnosis of Parsonage-Turner syndrome. In addition to excluding more common causes of shoulder pain such as rotator cuff tear, impingement syndrome, or labral tear, MRI is sensitive for the detection of signal abnormalities in the shoulder girdle musculature that are related to denervation. MRI also provides an advantage for excluding structural abnormalities that may cause similar denervation changes in rotator cuff musculature such as a rotator cuff tear or spinoglenoid and suprascapular notch masses. In our series, MRI findings included a broad range of T1- and T2weighted signal abnormalities. Intramuscular changes observed in patients with ParsonageTurner syndrome reflect denervation changes and vary with the stage of disease [11, 12]. We used two imaging protocols in our series because the indications for imaging the brachial plexus or shoulder varied. We therefore had several limitations when comparing patients imaged with the brachial plexus protocol versus those imaged with the shoulder protocol. The FOV was larger for the brachial plexus protocol and includes more muscle groups than the smaller FOV used for the shoulder protocol. In addition, muscular edema was evaluated using inversion recovery sequences in the brachial plexus protocol and using a T2-weighted fast spin-echo sequence with fat saturation in the shoulder protocol. Although this difference in protocols resulted in minor differences in sensitivity for the detection of altered intramuscular signal intensity, we do not think that it is a significant limitation of the study. Both sequences sensitively detected intramuscular edema, and images were evaluated only for the presence or absence of edema. The assessment of fatty infiltration of the muscles also varied between the two protocols because T1-weighted images were used in the brachial plexus examination and T2weighted proton density images without fat saturation were used in the shoulder protocol. However, this did not limit image analysis because fatty infiltration could be assessed accurately by comparing the T2-weighted proton density images without fat saturation with the corresponding T2-weighted fast spinecho images with fat saturation. The mechanism and time course of MRI signal intensity changes in denervated skeletal muscle are not understood completely [11, 1315]. In the acute phase of denervation, the signal intensity of the muscles may be normal with MRI [11]. The earliest detectable change

AJR:189, July 2007

W41

Scalf et al.

Fig. 238-year-old man with mild pain and paresthesia along lateral aspect of arm. A, Axial T2-weighted fast spin-echo image with fat saturation of right shoulder shows increased T2 signal (arrowheads) throughout infraspinatus muscle that is compatible with edema. B, Axial T1-weighted fast spin-echo image shows no detectable atrophy or fatty infiltration of infraspinatus muscle (arrowheads).

Fig. 339-year-old man with pain and paresthesias in right shoulder and upper extremity. A, Axial T2-weighted fast spin-echo image of right shoulder with fat saturation shows increased T2 signal throughout infraspinatus (I), subscapularis (S), and portion of deltoid (D) muscles that is compatible with edema. B, Axial T1-weighted fast spin-echo image shows mild atrophy and fatty infiltration of affected muscles: infraspinatus (I), subscapularis (S), and deltoid (D) muscles.

in denervated muscles is a diffuse increased T2-weighted signal (due to edema), which may occur without a T1-weighted signal change [11, 15, 16]. In the subacute and chronic phases of denervation, T2-weighted signal abnormalities persist and muscular atrophy may develop [11, 12, 15]. Atrophic changes in muscle are reflected by decreased muscular mass and increased intramuscular, linear, T1-weighted signal due to fatty infil-

tration. MRI intramuscular signal change and clinical symptoms may revert to normal in several months after the chronic phase [11]. MRI signal change for patients with Parsonage-Turner syndrome has been described in several studies using small patient series. Helms et al. [7] reported increased T2weighted signal in the supraspinatus, infraspinatus, and deltoid muscles at initial presentation of three patients with Parsonage-

Turner syndrome. T1-weighted changes of atrophy without fatty infiltration during followup also were described. Until recently, no large studies that reported MRI features of patients with Parsonage-Turner syndrome had been published, to our knowledge. Coincident with our study, Gaskin and Helms [5] reported the MRI findings of 27 patients with ParsonageTurner syndrome; to our knowledge, their study is the largest published series to date.

W42

AJR:189, July 2007

MRI of Patients with Parsonage-Turner Syndrome

A
Fig. 418-year-old man with shoulder pain and progressive left upper extremity weakness. A and B, Oblique sagittal (A) and axial (B) T2-weighted images with fat saturation show increased T2 signal in supraspinatus (S) and infraspinatus (I) muscles. C, Axial T1-weighted image shows atrophy of infraspinatus muscle (arrowhead).

C The MRI findings and anatomic distribution of muscular involvement of patients in our study were similar to those of Gaskin and Helms [5]. Our study of 26 patients showed that the supraspinatus and infraspinatus muscles were most commonly involved. However, because our study also included patients who were imaged with the large FOV of the brachial plexus protocol, we noted that other shoulder muscle groups may also have abnormal MRI signal changes, including the latissimus dorsi and rhomboid muscles. In addition, MRI findings indicative of muscle atrophy, including evidence of fatty infiltration and decrease in muscle bulk, were identified more frequently when compared with previous studies. The latter findings are likely related to the relative delay between the onset of symptoms and the MRI examination. Further studies will need to evaluate MRI features relative to the time course of the disease in patients with Parsonage-Turner syndrome. The limitations of the study include the lack of control subjects, lack of information about the timing of MRI features relative to the disease course, and the low number of follow-up imaging studies available for review. The study was also limited by variations in the anatomic extent of the shoulder girdle musculature included on the examination because the varied symptoms meant that patients were referred for shoulder or brachial plexus examination. Follow-up MRI examinations were performed infrequently because most patients showed clinical improvement with resolution of symptoms and did not require further imaging. One patient in our study who initially had edema in the supraspinatus and infraspinatus muscles showed complete resolution of the denervation changes at an 8-month follow-up MRI examination. Others did not return for follow-up evaluation. An additional limitation was the absence of a definite reference standard for diagnosing ParsonageTurner syndrome that includes evaluation of clinical history, findings at presentation, and physical examination findings. For patients with shoulder pain of an unknown cause, signal intensity changes observed during MRI can be a valuable adjunct to clinical data and electrophysiologic testing for the diagnosis of Parsonage-Turner syndrome. This patient series presents the MRI features of patients with Parsonage-Turner syndrome, which include a spectrum of signal abnormalities in various muscle groups of the rotator cuff and shoulder girdle. Familiarity with these MRI features is important for radiologists because they may be the first to suggest the diagnosis.

AJR:189, July 2007

W43

Scalf et al.

Fig. 552-year-old man with acute onset of severe left shoulder pain. A and B, Oblique coronal (A) and oblique sagittal (B) T2-weighted images with fat saturation show increased T2 signal in supraspinatus (S), infraspinatus (I in B), and deltoid (D) muscles.

References
1. Parsonage MJA, Turner JW. Neuralgic amyotrophy: the shoulder-girdle syndrome. Lancet 1948; 251:973978 2. Spillane JD. Localised neuritis of the shoulder girdle: a report of 46 cases in the MEF. Lancet 1943; 242:532535 3. Tsairis P, Dyck PJ, Mulder DW. Natural history of brachial plexus neuropathy: report on 99 patients. Arch Neurol 1972; 27:109117 4. Magee KR, Dejong RN. Paralytic brachial neuritis: discussion of clinical features with review of 23 cases. JAMA 1960; 174:12581262 5. Gaskin CM, Helms CA. Parsonage-Turner syndrome: MR imaging findings and clinical information of 27 patients. Radiology 2006; 240:501507 6. Turner JW, Parsonage MJ. Neuralgic amyotrophy (paralytic brachial neuritis): with special reference

to prognosis. Lancet 1957; 273:209212 7. Helms CA, Martinez S, Speer KP. Acute brachial neuritis (Parsonage-Turner syndrome): MR imaging appearancereport of three cases. Radiology 1998; 207:255259 8. Pellas F, Olivares JP, Zandotti C, Delarque A. Neuralgic amyotrophy after parvovirus B19 infection. Lancet 1993; 342:503504 9. Suarez GA, Giannini C, Bosch EP, et al. Immune brachial plexus neuropathy: suggestive evidence for an inflammatory-immune pathogenesis. Neurology 1996; 46:559561 10. Beghi E, Kurland LT, Mulder DW, Nicolosi A. Brachial plexus neuropathy in the population of Rochester, Minnesota, 19701981. Ann Neurol 1985; 18:320323 11. Uetani M, Hayashi K, Matsunaga N, Imamura K, Ito N. Denervated skeletal muscle: MR imagingwork in progress. Radiology 1993;

189:511515 12. Sallomi D, Janzen DL, Munk PL, Connell DG, Tirman PF. Muscle denervation patterns in upper limb nerve injuries: MR imaging findings and anatomic basis. AJR 1998; 171:779784 13. Shabas D, Gerard G, Rossi D. Magnetic resonance imaging examination of denervated muscle. Comput Radiol 1987; 11:913 14. Polak JF, Jolesz FA, Adams DF. Magnetic resonance imaging of skeletal muscle: prolongation of T1 and T2 subsequent to denervation. Invest Radiol 1988; 23:365369 15. Fleckenstein JL, Watumul D, Conner KE, et al. Denervated human skeletal muscle: MR imaging evaluation. Radiology 1993; 187:213218 16. West GA, Haynor DR, Goodkin R, et al. Magnetic resonance imaging signal changes in denervated muscles after peripheral nerve injury. Neurosurgery 1994; 35:10771085

W44

AJR:189, July 2007