| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Thoracic Imaging |
1 From the Department of Diagnostic Radiology, Royal Marsden National Health Service Trust, Downs Rd, Sutton, Surrey SM2 5PT, England. From the 1997 RSNA scientific assembly. Received September 17, 1998; revision requested November 23; final revision received June 28, 1999; accepted August 2. Address reprint requests to A.D.M.
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: Fifty patients with symptoms of brachial plexopathy (principally pain, weakness, and paresthesia) who had received treatment for breast cancer, which included surgery, radiation therapy, and cytotoxic chemotherapy, underwent MR imaging at 1.5 T. MR imaging was performed by using a body coil, which was supplemented with surface-coil imaging of the cervical spine and shoulder-coil imaging of the brachial plexus. At review, two observers attempted to discriminate between tumor recurrence and nonmalignant causes of symptoms. The diagnosis was verified with histologic analysis or a follow-up of at least 12 months.
RESULTS: Of 27 patients demonstrated to have tumor recurrence, 26 were correctly identified by using MR imaging; the recurrence was directly related to the brachial plexus in 17. During the follow-up, 21 patients remained free of recurrence, 20 of whom were determined to have a nonmalignant cause of symptoms. Two of the 50 patients were excluded from the analysis. The MR criteria used for detection of tumor yielded a sensitivity of 96%, specificity of 95%, positive predictive value of 96%, and negative predictive value of 95%.
CONCLUSION: MR imaging is reliable and accurate in the diagnosis of symptomatic brachial plexopathy following breast cancer therapy.
Index terms: Brachial plexus, 31.339 • Brachial plexus, MR, 31.121411, 31.121413, 31.30 • Breast neoplasms, therapy, 00.30, 07.30, 09.30 • Magnetic resonance (MR), coil arrays, 00.121411
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Recent literature confirms that the anatomy of the brachial plexus is well demonstrated with magnetic resonance (MR) imaging (8–11). Although various techniques have been employed, to our knowledge, all studies have been undertaken by using the body coil, and most reports are descriptive (8–12). However, Moore et al (6) demonstrated that MR imaging was able to facilitate a working diagnosis of fibrosis or recurrent tumor in patients with breast cancer. Our study was performed to investigate the accuracy of MR imaging in the detection of tumor recurrence by using surface-coil technology in a group of patients who developed brachial plexopathy following treatment for breast cancer.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Imaging Technique
All patients underwent MR imaging of the brachial plexus at 1.5 T (Vision; Siemens, Erlangen, Germany). To visualize the brachial plexus in its entirety, images of the cervical spine; neck; thoracic inlet; and supraclavicular, infraclavicular, and axillary regions were obtained. The cervical spine was imaged by using a phased-array surface coil (receive-only, radio-frequency pulses transmitted by a body coil). T1-weighted, sagittal, spin-echo images (320/14 [repetition time msec/echo time msec], two signals acquired, 250-mm field of view, 512 x 221 matrix) were obtained. T1-weighted, transverse, turbo spin-echo images (800/12 [effective], echo train length of three, two signals acquired, 175-mm field of view, 256 x 168 matrix) were obtained from segments C4 to T1. A section thickness of 5 mm, with no intersection gap, was used for both sequences.
The lower part of the neck, upper part of the chest, and axilla were examined by using body-coil imaging. T2-weighted, transverse, turbo spin-echo images (6,000 [nominal]/138 [effective], echo train length of 29, eight signals acquired) and T1-weighted, coronal, spin-echo images (340–450 [nominal]/14, two signals acquired) were obtained. The shortest possible repetition time within the range was used, which varied between patients and depended on the specific absorption rate. Electrocardiographic triggering was used for both sequences. For the transverse T2-weighted images, 15 sections were obtained, with a section thickness of 10 mm and an intersection gap of 2.5 mm (256 x 116 matrix). For the T1-weighted coronal images, 11 sections were obtained, with a section thickness of 6 mm and no intersection gap (256 x 192 matrix). In both cases, a field of view of 350 mm was used.
Local images of the distal brachial plexus were obtained by using a flexible surface coil designed for shoulder imaging (receive-only, radio-frequency pulses transmitted by a body coil). With the patient in the supine position, the flexible coil was positioned around the supraclavicular and infraclavicular regions on the symptomatic side. The coil was held in place by using a positioning device, a hinged clamp incorporating a ball joint. Positioning was improvised according to the build of the patient, as the coil was not being used for the exact purpose for which it was designed. The angle made by the flexible coil with the long axis of the magnet bore was kept below 45° to prevent coupling with the body coil.
T1-weighted, turbo spin-echo images (850/12 [effective], echo train length of three, two signals acquired) were obtained in the transverse and sagittal planes. Turbo short inversion time inversion-recovery (STIR) images (4,800/60 [effective]/150 [repetition time msec/echo time msec/inversion time msec], echo train length of 11, one signal acquired) were obtained in the transverse plane in 10 patients. A matrix of 256 x 192 and a field of view of 200 mm were used for all sequences. Fifteen 6-mm-thick sections, with no intersection gap, were obtained with each sequence.
The approximate overall imaging time was 60 minutes.
Image Interpretation
The images were retrospectively reviewed by two radiologists experienced in MR imaging (J.E.S.H., A.D.M.). The only clinical information provided was that the patient had symptoms suggestive of brachial plexopathy. The age of the patient, the time from diagnosis, the tumor size at presentation, the nodal status at presentation, and the radiation field were all concealed.
Interpretation was based on parameters of morphology and the signal intensity. Tumor recurrence was diagnosed only if a mass lesion was seen at any site along the course of the brachial plexus or if there was evidence of metastatic disease to the spine. High signal intensity on T2-weighted or STIR images was considered abnormal but, in the absence of a mass, not diagnostic of recurrent tumor. Thickening of the components of the brachial plexus, with or without signal intensity change, was the diagnostic criterion used for diagnosis of fibrosis.
Images were reviewed independently, and data were collected by a single investigator (A.Q.). Two instances of discrepant reading were resolved by consensus. Where consensus was not reached, imaging results were excluded from the analysis (two of the 50 patients). Images from each examination were reviewed for the presence of metastases in sites distant from the brachial plexus, for example, the mediastinum and the cervical spine.
Follow-up and Verification
All patients were followed up for a minimum of 12 months or until death; three patients died of disseminated carcinoma within 12 months. The absence of progression of symptoms or signs or spontaneous resolution of symptoms was interpreted as demonstration of nonmalignant disease such as fibrosis. The diagnosis of tumor recurrence was inferred from demonstration of a progressively enlarging mass either clinically or at subsequent imaging or from demonstration of a reduction of tumor bulk following treatment that was confirmed with both clinical methods and imaging studies. Histologic confirmation of tumor recurrence was obtained in five patients—with percutaneous biopsy in four patients and with open surgical biopsy in one patient.
Statistics
Statistical analysis was performed by using descriptive statistics and Bayesian analysis.
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
|
|
|
|
|
|
|
One false-positive diagnosis of tumor recurrence was made. In the patient with this false-positive diagnosis, there was sufficient thickening of the soft tissues around the cords of the brachial plexus to be interpreted as a mass, and this was classified as a tumor recurrence. However, follow-up MR imaging over 18 months, during which no specific treatment was undertaken, demonstrated no change. It therefore appears that the correct diagnosis was radiation fibrosis.
In two patients, the MR imaging findings were equivocal, such that a brachial plexus tumor or fibrotic change could not be diagnosed with confidence, and consensus could not be reached. These patients were excluded from the analysis. One of these patients was known to have pulmonary metastatic disease and died 2 months after imaging, without a firm cause of her symptoms established. In the other patient, a local recurrence was detected in the breast at MR imaging and was subsequently excised; local chemotherapy and radiation therapy followed. The brachial plexus in this patient was thickened, with an increased amount of axillary soft tissue that returned low signal intensity at T2-weighted imaging. There has been no clinical evidence of disease recurrence related to the brachial plexus or in the breast after 2 years, and the symptoms of brachial plexopathy have remained stable.
MR imaging accurately depicted 26 of 27 cases of malignant disease and helped to accurately exclude malignant disease in 20 of 21 cases. This resulted in a sensitivity of 96% and a specificity of 95%, with a positive predictive value of 96% and a negative predictive value of 95%.
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Radiation therapy reduces the risk of relapse, but at the cost of possible iatrogenic damage to the brachial plexus. There is no doubt that adjuvant radiation therapy improves local control and has a survival advantage (13,14). However, the side-effect profile of the treatment is such that physicians continue to adjust and refine fractionation regimens, with the intention of reducing the incidence of radiation-induced brachial plexopathy.
The pathogenesis of radiation-induced brachial plexopathy is poorly understood. The degree of tissue atrophy and fibrosis is variable and is thought to result from a combination of direct cell damage from ionizing radiation and progressive ischemic changes that continue to evolve after the initial injury and are responsible for the late complications of radiation therapy (15). Findings of post mortem examinations in patients with radiation-induced brachial plexopathy have shown areas of fibrosis surrounding the nerves of the brachial plexus, fibrous thickening of the neurilemmal sheath, demyelination, and fibrous replacement of nerve fibrils. Proximal to the irradiated area, the nerve structure is normal, but extensive myelin loss and atrophy are seen distal to the radiation field (16).
The brachial plexus is composed of mixed nerves, and the clinical manifestation of brachial plexopathy is with paresthesia, pain, and weakness. Sensory symptoms predominate initially. The frequency of radiation damage is low (less than 1%) but is dose dependent, and prior cytotoxic chemotherapy appears to increase the likelihood of early symptoms. Neurologic damage may become symptomatic at any time but generally does so 5–30 months after therapy, with a peak at 10–20 months (6).
Our results confirm that good-quality images and a simple diagnostic approach yield good results in the detection of tumor recurrence and in the discrimination of tumor recurrence from fibrosis. The results are encouraging, particularly in view of the sometimes extreme difficulty experienced in assessing patient conditions clinically. Some remarkably large masses were concealed clinically by induration and fibrosis of superficial tissues (Fig 1). In other instances, convincing mass lesions could be demonstrated in areas inaccessible to clinical examination, such as the area behind the clavicle (Fig 2).
MR imaging has traditionally been seen as competitive with CT. CT is certainly capable of demonstrating masses (17,18), but it requires intravenous administration of contrast material, involves ionizing radiation, and provides only transverse images. However, Moskovic et al (19) found that, in the absence of a palpable axillary mass, the prevalence of CT depiction of breast cancer recurrence is extremely low; they considered the technique unjustified for screening for clinically occult axillary disease in patients with arm symptoms following axillary surgery or radiation therapy for breast cancer.
The role of US in the investigation of brachial plexopathy has not been extensively investigated. The anatomy of the brachial plexus is technically difficult to depict by using US but in our experience has been useful in guiding cytologic aspiration of clinically suspected mass lesions in the upper axilla and root of the neck.
MR imaging is emerging as the optimal imaging modality for the brachial plexus. In the patient with breast cancer, it should be stressed that metastases may arise anywhere along the course of the brachial plexus, although the critical area is in the retroclavicular region around the cords of the brachial plexus. We recommend that the cervical spine always be included, as symptoms attributed to a peripheral brachial plexus lesion may be more proximal in origin. Body-coil images, particularly in the coronal plane, proved extremely useful in demonstrating asymmetry and occasionally demonstrated large clinically occult masses. Subsequent interrogation by using the local flexible coil enabled concentration on limited or subtle morphologic abnormalities.
The diagnostic parameters that were used were simple and were inferred from previous articles (6,8–12) on the subject. To our knowledge, tumor recurrence is usually reported as returning low signal intensity on T1-weighted images and high signal intensity on T2-weighted or STIR images (6,8,11). We found that the signal intensity characteristics were variable, and detection of a mass lesion proved to be a reliable indicator of tumor recurrence. The MR imaging literature draws attention to the low signal intensity returned by lesions with a high fibrotic content (20). Moore et al (6) found a reduction in the signal intensity of fat in the axillary and supraclavicular regions. They attributed these areas to the presence of fibrosis within radiation fields and found that these processes often result in a loss of clarity and distortion of the neurovascular bundles.
Our experience was that radiation-induced fibrosis followed this pattern on spin-echo images, which resulted in some thickening of nerve roots (Fig 3). Low signal intensity is certainly seen in established fibrosis; however, it should be remembered that in early fibrotic change, there is a substantial inflammatory component that in some cases will persist for years and that results in gadolinium-based contrast material enhancement on T1-weighted images and high signal intensity on T2-weighted and STIR images.
Images obtained with the STIR sequence demonstrated linear areas of high signal intensity along the brachial plexus in cases of radiation fibrosis. The STIR sequence was useful in highlighting abnormalities in proximity to the brachial plexus (Fig 4), but the anatomic detail was less clear than it was on spin-echo images. A combination of the sequences is therefore recommended. In this study, we did not use intravenous gadolinium-based contrast material; however, tumor vascularity may be helpful in characterizing possible recurrences, and dynamic enhancement studies should yield interesting information in future studies.
|
|
Interpretation of brachial plexus images depends fundamentally on adequate demonstration of anatomic detail. Even when this is achieved, difficulties may be encountered. In our study, analysis was retrospective and was undertaken by observers with previous experience. Consensus reading was used, because this fairly reflects our clinical practice, in which the majority of cases, including all those considered to have equivocal findings, are reviewed at clinicoradiologic meetings.
Agreement was not reached in two patients, who were excluded from the analysis. This skews the results to slightly better accuracy but underlines the fact that interpretation is occasionally controversial. A further limitation is the difficulty in obtaining detailed histologic correlation, as surgical resection is very rarely attempted, and radiologic interpretation must be verified by follow-up. Owing to constraints of time and finance, we did not examine an asymptomatic control group, which potentially introduces bias to the results. However, the findings confirm a valid role for MR imaging in the investigation of symptomatic brachial plexopathy.
Our study findings demonstrate that, by using a comprehensive but nevertheless straightforward imaging technique and simple diagnostic parameters, reliable diagnostic information can be obtained in a common and challenging clinical setting.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Author contributions: Guarantors of integrity of entire study, A.D.M., J.E.S.H.; study concepts and design, A.D.M., J.E.S.H.; definition of intellectual content, A.D.M., J.E.S.H.; literature research, A.Q.; clinical studies, A.D.M., A.R.P., P.R., J.E.S.H., A.Q.; data acquisition, A.Q., A.D.M., J.E.S.H.; data analysis, A.Q., A.D.M., A.R.P.; manuscript preparation, A.Q., A.D.M.; manuscript editing, A.D.M.; manuscript review, A.D.M., P.R., J.E.S.H.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
