Spatial mapping of the brachial plexus using three-dimensional ultrasound C J C Cash, MRCP, FRCR1, A M Sardesai, MD, DA, FRCA2, L H Berman, MRCP, FRCR1, M J Herrick, FRCA2, G M Treece, MA, PhD3, R W Prager, MA, PhD3 and A H Gee, MA, PhD3
[size=-1]Cambridge University Departments of 1 Radiology and 2 Anaesthesia, Addenbrookes Hospital NHS Trust, Hills Road, Cambridge CB2 2QQ and 3 Cambridge University Department of Engineering, Trumpington Street, Cambridge CB2 1PZ, UK
Imaging of the brachial plexus with MRI and standard two-dimensional (2D) ultrasound has been reported, and 2D ultrasound-guided regional anaesthetic block is an established technique. The aim of this study was to map the orientation of the brachial plexus in relation to the first rib, carotid and subclavian arteries, using three-dimensional (3D) ultrasound. A free-hand optically tracked 3D ultrasound system was used with a 12 MHz transducer. 10 healthy volunteers underwent 3D ultrasound of the neck. From the 3D ultrasound data sets, the outlines of the brachial plexus, subclavian artery and first rib were manually segmented. A surface was interpolated from the series of outlines to produce a spatially orientated 3D reconstruction of the brachial plexus. The brachial plexus could be mapped in all volunteers, although a variation in image resolution between individuals existed. Anatomical variations were demonstrated between the 10 volunteers; the most notable and clinically relevant was the alignment of the plexus divisions. 3D reconstructions illustrated the plexus, changing its orientation from a vertical alignment in the interscalene region to a more horizontal alignment in the supraclavicular fossa. Spatial mapping of the brachial plexus is possible with 3D ultrasound using the subclavian artery and first rib as landmarks. There is a deviation from the conventionally described anatomy and this may have implications for the administration of regional anaesthesia.
Modern ultrasound machines are capable of imaging individual nerves of the brachial plexus from their extraforaminal roots to their cords in the infraclavicular region [1–4]. Two-dimensional (2D) ultrasound is increasingly being used to guide regional anaesthesia [5–13]. A high-resolution freehand three-dimensional (3D) ultrasound system has been developed by Cambridge University Engineering Department. The system software, Stradx, preserves much of the original 2D data, ensuring a highly accurate and reproducible 3D data set (http://mi.eng.cam.ac.uk/
rwp/stradx). Rigorous laboratory calibrations have demonstrated a point location accuracy of within 0.5 mm [14]. The system has been used for clinical research in neonatal musculoskeletal imaging, organ volumes, subcutaneous injection assessment [15] and spatial localization of breast lesions. Conventional ultrasound of the brachial plexus requires careful positioning of the transducer, particularly within the supraclavicular region. The freehand Stradx 3D ultrasound system, however, allows almost limitless placement and orientation of the transducer in acquiring the data sets. The transducer can be moved with as much freedom as conventional 2D ultrasound. The aim of this study was to identify the normal brachial plexus in healthy individuals and to demonstrate the plexus in three dimensions with respect to key anatomical landmarks.
Following local ethical approval and written consent, 10 healthy asymptomatic volunteers (8 male and 2 female, average age 39 years, range 28–52 years) underwent bilateral 3D ultrasound neck examinations. An Aplio ultrasound machine (Toshiba Medical Systems, http://www.toshiba-europe.com) was used with a 12 MHz linear matrix transducer. Colour Doppler facility was available, but was only rarely required during the first few examinations. A position sensor (Polaris; Northern Digital Inc., http://www.ndigital.com/) continually tracks the orientation of 15 infrared emitting diodes attached to the transducer. Ultrasound images together with their spatial orientation data are simultaneously acquired on an 800 MHz PC running Stradx (http://mi.eng.cam.ac.uk/
rwp/stradx). This software is also used to analyse and display the 3D data. The volunteers are supine, reclined at an angle of approximately 45° with their head turned 30° to the contralateral side and their arms by their sides. Using standard acoustic gel, the transducer was placed in a transaxial position over the apex of the posterior triangle and swept smoothly down the lateral side of the neck and into the supraclavicular fossa. Using conventional definitions, the transducer is orientated in an axial plane at the start of the sweep and in the sagittal plane at the end of the sweep (Figure 1
). In order to follow the plexus as far laterally as possible, the position of the transducer at the end of the sweep was angled anteriorly to demonstrate anatomy lying deep to the lateral third of the clavicle.
Figure 1. The whole data set is illustrated. The position of the transducer at the beginning and end of the sweep is demonstrated on the left and right, respectively.
Data analysis
Following acquisition, an algorithm within the Stradx software was applied to the data set to correct for any misregistration caused by operator-induced variations in transducer pressure or physiological tremor [16]. From representative B-scans from the data set, manual segmentation of individual components of the brachial plexus was performed by a single observer (CJCC) from the first visible appearance as roots emerging from the cervical foramina to its disappearance under the clavicle (Figure 2
). A surface was interpolated from this series of outlines to produce a surface rendered reconstruction of the brachial plexus in relation to the key anatomical landmarks [17].
Figure 2. (a) The whole data set from the extraforaminal to the costo-clavicular section of a right brachial plexus. (b) An individual B-scan from the data set from the interscalene triangle and illustrates the vertical alignment of the plexus trunks at this stage. The individual components of the plexus on this single image have been manually segmented (dotted lines). (c) A series of representative outlines of the brachial plexus, from the whole data set.
Figure 3. The first rib, subclavian artery and components of the brachial plexus have been manually segmented. Arrowheads indicate a nerve containing fibrillar structures occasionally seen within the brachial plexus at this level.
The data sets were individually reformatted into an equivalent coronal plane for each case. Distances from the skin surface to the most superficial part of the brachial plexus and to the apical pleural surface were measured.
Data sets were obtained from both sides of the neck in 8 of the 10 volunteers. In two volunteers only one side was examined due to time constraints. Each data set took an average of 8 s to acquire. Typically two data sets were recorded per side, although only one data set was kept for further analysis. The data set with the least movement artefact showing the best anatomical definition was kept. 18 satisfactory data sets were analysed further. The average data set contained 586 images (standard deviation (SD)=144 images). The Stradx algorithm to correct for misregistration between images caused by sub-millimetre irregularities in probe pressure from the operator was applied to the whole data set. The effect of this correction on a reconstruction plane chosen through the data set to produce a multiplanar reformat (MPR) along the longitudinal axis of the supraclavicular brachial plexus is shown in Figure 4
Figure 4. (a) A multiplanar reconstruction (MPR), reconstructed from a data set from a left brachial plexus along the longitudinal plane of the plexus. The image quality improves by correcting for operator-induced misregistration (b). (c,d,e) The original B-scans from the data set of the trunk, divisions and cord level, respectively. A hyper-reflective sheath clearly outlines the brachial plexus trunks and divisions (c,d).
Individual B-scan images were similar to previously published reports on 2D ultrasound imaging of the brachial plexus [1, 4]. Within the interscalene triangle, the brachial plexus trunks appear as 3–5 well-defined circular hyporeflective structures vertically aligned between the muscles of scalenus anterior and scalenus medius. A hyper-reflective sheath surrounds the brachial plexus (Figure 4c,d
). The structure of the plexus within the supraclavicular fossa is more complex, but again supports previous published findings of hyporeflective nerves with occasional internal echo-reflective fibrillar structures [1, 4], (Figure 3
). Typically there were 6–8 hyporeflective structures defined within the supraclavicular fossa. It was assumed that this increase in number of identified nerves represented the divisions within the plexus. The plexus was clearly identified from all 18 data sets.
All data sets were segmented and a surface interpolated. This process took approximately 15 min per data set. Examples of the 3D reconstructions are shown in Figures 5 and 6
. Using the landmarks of the first rib and segments of the carotid and subclavian arteries, the spatial orientation of the plexus as it descends the neck can be defined. The mean distance from skin to the most superficial part of the brachial plexus within the supraclavicular fossa, measured from the coronal MPRs, was 0.53 cm (SD=0.15 cm; range 0.3–0.8 cm). From the same reconstructions, the mean distance from skin to the most superficial aspect of the apical pleural surface was 2.19 cm (SD=0.32 cm; range 1.57–2.53 cm).
Figure 5. Examples of surface rendered 3D reconstructions of the brachial plexus in relation to segments of the carotid and subclavian artery and the first rib. A right brachial plexus viewed from (a) a lateral and (b) an anteroposterior perspective, respectively.
Figure 6. Surface rendered reconstructions of a right and left brachial plexus from the same individual viewed laterally (a, b) and from above (c, d). This individual demonstrates clumped right sided divisions, white arrows (a, c), and flatter left sided divisions, arrowheads.
An advantage of recording a complete data set is the ability to repeatedly review, frame by frame, the entire examination. During the post-processing analysis, subtle variations in the anatomy of the brachial plexus between the individual volunteers were demonstrated. Alignment of the divisions within the supraclavicular fossa
The orientation of the plexus changes from a vertical alignment in the interscalene region to a more horizontal flat arrangement in the supraclavicular region as it traverses the first rib. This pattern was observed in 12 of 18 (67%) sides. In the remaining third of brachial plexuses, the vertical alignment converts to a clumped (tightly grouped) arrangement over the first rib. One individual demonstrated a clumped arrangement on the right (Figure 6a,c
). These two patterns can also be appreciated from the original B-scans taken from the right and left data set at approximately the same anatomical point over the first rib (Figure 7
). On both sides the plexus can be clearly identified within a hyper-reflective fascial sheath. The maximum anteroposterior width of this sheath is 10 mm on the right and 15 mm on the left. In this subject, the left plexus lies in a more posterior position relative to the subclavian artery than on the right. This difference in relationship with the subclavian artery is also appreciated by the 3D reconstructions (Figure 6
Figure 7. Original B-scans illustrating the different alignment of the divisions over the first rib within a single individual. The anteroposterior width of the plexus at this level measures 10 mm on the right and 15 mm on the left.
The course of C5 from its extraforaminal origins to the interscalene region
In the majority of volunteers, we observed the nerve roots emerging from the cervical foramina with varying clarity. They descend into a more superficial position, becoming vertically grouped together between the muscles of scalenus anterior and medius, as previously described (Figure 2b
). In two (11%) of our volunteers, however, C5 passed anterior, rather than posterior, to scalenus anterior then coursed posteriorly over the superficial surface of scalenus anterior to unite with the remainder of the plexus as these nerve trunks emerged from their interscalene position (Figure 8
Figure 8. The course of the C5 nerve root (dashed line) appears to run anterior to scalenus anterior (SA) (a, b) rather than between scalenus anterior and medius (SM). It runs over the lateral border of scalenus anterior before reuniting with the remaining nerve roots, C6 and C7 (c, d). The course of the C6 and C7 nerve roots (dotted line) runs between scalenus anterior and medius (c, d).
Position of the divisions of the brachial plexus in relation to the scalene muscles
Two patterns amongst the volunteers were defined. In 5 of 18 (28%) cases, the divisions formed whilst the plexus was still between the scalene muscles (Figure 9a
Figure 9. (a) The plexus has formed divisions whilst still apparently within the scalene muscles. (b) The divisions first appear once the plexus has emerged from between the scalene muscles.
Relationship to local vascular structures
There were wide variations in the number, position and size of vascular structures orientated in an anteroposterior direction. Doppler ultrasound demonstrated these to be arterial. In 16 of 18 (89%) cases, a vessel was seen crossing the plexus in the supraclavicular region (Figure 10
), and in four (22%) of these cases a second vessel was seen. The diameter of these vascular structures varied between 1 mm and 3 mm. In each of the 18 cases the vessel was within a sub-millimetre distance from the brachial plexus sheath.
Figure 10. In the majority of volunteers, an artery was seen crossing the brachial plexus in an anteroposterior direction within the supraclavicular fossa. This artery is most likely either the superficial cervical or suprascapular branch of the subclavian artery.
Proximity to the subclavian artery varied with a mean closest distance of 2.77 mm (SD=2.07 mm; range 0.4–6.7 mm).
The brachial plexus is formed from the union of the ventral rami of C5–C8 and T1. For descriptive purposes the plexus may be divided into roots, trunks, divisions and cords (Figure 11
) [1]. There have been several comprehensive reports of using 2D ultrasound to map the brachial plexus from its extraforaminal origins to the infraclavicular region [1–4]. With modern high-resolution ultrasound machines, the brachial plexus can be clearly imaged with a steep but short learning curve. Many authors refer to the use of colour Doppler to differentiate between vessel and nerve, but most workers have found, as with our experience, that this became unnecessary as familiarity of the anatomy developed. In this study the resolution of the images resulted in all brachial plexuses being definable from each data set. However, for the purposes of this small study, there was bias in the selection of volunteers towards leaner individuals. As with conventional 2D ultrasound, body habitus and length of neck will influence the quality of the original B-scans and therefore the quality of a 3D data set.
Figure 11. The brachial plexus may be divided into roots, trunks, divisions and cords. The upper trunk is formed from the union of the roots C5 and C6, the middle trunk is a continuation of C7 root and the lower trunk is formed by the union of C8 and T1 roots. Each trunk divides into anterior and posterior divisions. The lateral cord is formed by the union of the anterior divisions of the upper and middle trunks. The medial cord is formed by the anterior divisions of the lower trunk. The posterior cord is formed by the union of the posterior divisions of all the trunks.
Previous authors have described the appearance of three hyporeflective circular structures within the interscalene region representing the three trunks. We noted a variation of between three and five structures defined between the scalene muscles. Depending on the angle of the transducer, a combination of nerve roots and trunks may be imaged in the interscalene region, i.e. if the transducer is placed over the superficial upper trunk, but is medially angled, the middle trunk, C8 and T1 rami would be included in the image, producing four hyporeflective structures in a single 2D image.
MRI is the imaging investigation of choice in brachial plexopathies [18, 19]. However, artefacts from respiration and neighbouring vascular structures may be a problem [18]. Although MRI is multiplanar, the five ventral rami can usually not be seen in one single plane due to the normal cervical lordosis [19]. To our knowledge this is the first time 3D ultrasound has been used to reconstruct the brachial plexus. Although the reconstructions are relatively basic, the inclusion of certain key structures such as segments of the carotid artery, subclavian artery and the first rib allows spatial orientation of the brachial plexus to be determined. This has revealed an interesting finding in the majority of volunteers of a propeller-like rotation in the brachial plexus from a vertical to a horizontal plane. This flattening out occurred in most cases as the plexus crosses the first rib. A definite variation existed amongst our volunteers regarding the degree of flattening observed, with the plexus remaining tightly clumped as it crossed the first rib in a third of our volunteers. Although time-consuming, a 3D ultrasound data set does allow frame-by-frame interpretation of the anatomy, which in this study has shown further anatomical variations that were not appreciated during initial data acquisition. The variation in the course of C5 nerve root was an interesting finding. As far as we are aware, this has not been previously described. This anatomical variant could have implications for interscalene nerve blocks, although given the relatively large volumes of local anaesthetic delivered, typically up to 20 ml [8], it is likely that satisfactory local anaesthesia would still be achieved. When using anatomical landmarks and a nerve stimulator to guide blocks, anaesthetists use the interscalene groove, the position of the subclavian artery and the first rib to guide the needle tip. The results of this study show that the relationship of the brachial plexus to these structures is variable both between and within individuals. Anatomical variations in the branches of the subclavian artery are documented [20, 21] and not unexpected. The vessels coursing in an anteroposterior direction identified in this study are most likely to represent either the superficial cervical artery and/or the suprascapular artery. Complications of supraclavicular brachial plexus block performed without image guidance include pneumothorax and subclavian artery puncture with a reported incidence of 0.6–5% and 25.7%, respectively [4]. The vascular variations we describe and the proximity of arterial structures (0.4 mm in one of our cases) to the plexus explain this high incidence of "subclavian artery" traumatic puncture. The reported incidence of pneumothorax is relatively low considering the proximity of the pleural surface to the skin, just 1.57 cm in one of our volunteers. Ultrasound-guided regional anaesthesia of the brachial plexus has been claimed to be more successful, quicker and to result in a more effective block [5, 7] with fewer complications [5, 10] than performing regional blocks using anatomical landmarks and nerve-stimulators as guidance. Cost-savings have been demonstrated [22]. The current small study has demonstrated considerable variation in the anatomy of the brachial plexus both between and within individuals. The existence of such anatomical variations supports the use of image guidance for brachial plexus intervention. Whilst it would be unnecessary to use 3D ultrasound to guide regional anaesthesia in real time, the spatial orientation and the degree to which the plexus appears to rotate and flatten is appreciated by the 3D ultrasound reconstructions. Such reconstructions could potentially aid in the diagnosis of brachial plexus injuries and help guide surgery in the exploration of traumatic brachial plexopathy. Another potential clinical use of brachial plexus spatial mapping using 3D ultrasound is in radiotherapy planning. Radiation-induced brachial plexopathy is a significant cause of morbidity and litigation in patients who have received radiotherapy for supraclavicular nodal disease [23]. With the advent of 3D conformal radiotherapy and intensity-modulated radiotherapy (IMRT) [24], a spatial map of the brachial plexus could be imported into the planning system and the dose to the plexus adjusted accordingly. Received for publication January 6, 2005. Revision received April 20, 2005. Accepted for publication May 12, 2005.
Demondion X, Herbinet P, Boutry N, Fontaine C, Francke J-P, Cotton A. Sonographic mapping of the normal brachial plexus. AJNR Am J Neuroradiol 2003;24:1303–9.[Abstract/Free Full Text] Martinoli C, Bianchi S, Santacroce E, Pugliese F, Graif M, Derchi LE. Brachial plexus sonography: a technique for assessing the root level. AJR Am J Roentgenol 2002;179:699–702.[Abstract/Free Full Text] Sheppard DG, Iyer RB, Fenstermacher MJ. Brachial plexus: demonstration and US. Radiology 1998;208:402–6.[Abstract/Free Full Text] Yang WT, Chui PT, Metreweli C. Anatomy of the normal brachial plexus revealed by sonography and the role of sonographic guidance in anesthesia of the brachial plexus. AJR Am J Roentgenol 1998;171:1631–6.[Abstract/Free Full Text] Marhofer P, Schrogendorfer K, Koinig H, Kapral S, Weinstabl C, Mayer N. Ultrasonographic guidance improves sensory block and onset time of three-in-one blocks. Anesth Analg 1997;85:854–7.[Abstract] Marhofer P, Schrogendorfer K, Wallner T, Koinig H, Mayer N, Kapral S. Ultrasonographic guidance reduces the amount of local anesthetic for 3-in-1 blocks. Reg Anesth Pain Med 1998;23:584–8.[CrossRef][Medline] Williams SR, Chouinard P, Arcand G, Harris P, Ruel M, Boudreault D, et al. Ultrasound guidance speeds execution and improves the quality of supraclavicular block. Anesth Analg 2003;97:1518–23.[Abstract/Free Full Text] Chan VW, Perlas A, Rawson R, Odukoya O. Ultrasound-guided supraclavicular brachial plexus block. Anesth Analg 2003;97:1514–7.[Abstract/Free Full Text] Chan VW. Applying ultrasound imaging to interscalene brachial plexus block. Reg Anesth Pain Med 2003;28:340–3.[CrossRef][Medline] Ootaki C, Hayashi H, Amano M. Ultrasound-guided infraclavicular brachial plexus block: an alternative technique to anatomical landmark-guided approaches. Reg Anesth Pain Med 2000;25:600–4.[CrossRef][Medline] Peterson MK, Millar FA, Sheppard DG. Ultrasound-guided nerve blocks. Br J Anaesth 2002;88:621–4.[Free Full Text] Sandhu NS, Capan LM. Ultrasound-guided infraclavicular brachial plexus block. Br J Anaesth 2002;89:254–9.[Abstract/Free Full Text] De Andrés J, Sala-Blanch X. Ultrasound in the practice of brachial plexus anesthesia. Reg Anesth Pain Med 2002;27:77–89.[CrossRef][Medline] Treece GM, Gee AH, Prager RW, Cash CJ, Berman LH. High-definition freehand 3-D ultrasound. Ultrasound Med Biol 2003;29:529–46.[CrossRef][Medline] Cash CJC, Berman LH, Treece GM, Gee AH, Prager RW. Two- and three-dimensional ultrasound in the development of a needle-free injection system. Br J Radiol 2004;77:236–42.[Abstract/Free Full Text] Treece GM, Prager RW, Gee AH, Berman L. Correction of probe pressure artefacts in freehand 3D ultrasound. Med Image Anal 2002;6:199–214.[CrossRef][Medline] Treece GM, Prager RW, Gee AH, Berman L. Surface interpolation from sparse cross-sections using region correspondence. IEEE Trans Med Imaging 2000;19:1106–14.[CrossRef][Medline] Kichari JR, Hussain SM, Den Hollander JC, Krestin GP. MR imaging of the brachial plexus: current imaging sequences, normal findings, and findings in a spectrum of focal lesions with MR-pathologic correlation. Curr Probl Diagn Radiol 2003;32:88–101.[Medline] Hyodoh K, Hyodoh H, Akiba H, Tamakawa M, Nakamura N, Yama N, et al. Brachial plexus: normal anatomy and pathological conditions. Curr Probl Diagn Radiol 2002;31:179–88.[CrossRef][Medline] Snell RS. The head and neck. In: Clinical anatomy for medical students. Boston: Little, Brown and Company, 1986:717. Agur AMR. The neck. In: Grants atlas of anatomy. Baltimore: Williams and Wilkins, 1991:560. Sandhu NS, Sidhu DS, Capan LM. The cost comparison of infraclavicular brachial plexus block by nerve stimulator and ultrasound guidance. Anesth Analg 2004;98:267–8.[Free Full Text] Fathers E, Thrush D, Huson SM, Norman A. Radiation-induced brachial plexopathy in women treated for carcinoma of the breast. Clin Rehabil 2002;16:160–5.[Abstract/Free Full Text]
Webb S. Advances in three-dimensional conformal radiation physics with intensity modulation. Lancet Oncol 2000;1:30–6.[CrossRef][Medline]
[size=+2]This article has been cited by other articles:
[size=-1]P. M. Hopkins Ultrasound guidance as a gold standard in regional anaesthesia
Br. J. Anaesth., March 1, 2007; 98(3): 299 - 301. [Full Text][PDF]
|关于我们|业务合作|手机版|新青年麻醉论坛
( 浙ICP备19050841号-1 )
发表于 2009-5-31 23:31:30
QQ好友和群
收藏
淘帖
鲜花
鸡蛋
提升卡
置顶卡
变色卡
发表于 2009-6-27 20:06:23
