Atlas of peripheral nerve ultrasound

408 views

Published on

0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
408
On SlideShare
0
From Embeds
0
Number of Embeds
2
Actions
Shares
0
Downloads
0
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Atlas of peripheral nerve ultrasound

  1. 1. 17S. Peer, H. Gruber (eds.), Atlas of Peripheral Nerve Ultrasound,DOI 10.1007/978-3-642-25594-6_2, © Springer-Verlag Berlin Heidelberg 20132.1 General Considerationson Peripheral Nerve MRIEver since magnetic resonance imaging (MRI)got introduced in the armamentarium of medicalimaging methods, it received highest interestfrom neuroscience. This method’s superb soft tis-sue contrast allowed for unprecedented imagingof nervous tissue and developed very quickly intoa standard method of brain and spine imaging.Over the past few decades, MR scanners becamea routine tool and a mass product within afford-able cost and thus widely spread. At the sametime, rapid technical advances of the methodopened the door for even more detailed imagingand complete new applications: to date, scannerswith a field strength of 1.5–3T can be consideredstandard and allow for even more detailed imag-ing within acceptable imaging time. Cross-sectional imaging of nervous tissue can today becombined with numerous complementary scan-ning schemes, such as depiction of the vascula-ture with various MR angiography methods,insight into the biochemistry of nervous tissueand its pathologies with spectroscopy, and obser-vation of nerve tissue function by exploiting tis-sue perfusion and blood deoxygenation effectswith functional MRI. A more recent developmentin MRI is the ability to assess the Brownianmolecular motion of water molecules termeddiffusion-weighted imaging (DWI). This methodis routinely used for the early depiction of isch-emic stroke where hypoxic injury of the neuraltissue leads to a shift of free interstitial water intoW. JudmaierDepartment of Radiology, Innsbruck Medical University,Anichstrasse 35, 6020 Innsbruck, Tyrol, Austriae-mail: werner.judmaier@i-med.ac.at2Introduction to Magnetic ResonanceImaging of the Peripheral NervousSystem: General Considerationsand Examination TechniqueWerner JudmaierContents2.1 General Considerationson Peripheral Nerve MRI.......................... 172.2 Measurement Sequencesand Protocol Design................................... 24References................................................................. 27
  2. 2. 18 W. Judmaierswollen neurons (cytotoxic edema), thus reduc-ing the free diffusibility of water molecules whichresults in a drop of the calculated diffusioncoefficient. This imaging method is also used tofacilitate the differentiation of dense cellulartumor tissue from inflammatory or infectiouschanges as well as scar tissue after tumor resec-tion. Even more interestingly, DWI is also capa-ble of assessing the foremost direction of freeBrownian molecular motion, thus permitting thevisualization of the nerve bundles’ course withinthe spinal cord and cerebral tracts or, in case ofinjury, their anatomical or functional disruption.Despite these fascinating capabilities of MRI,its use has so far mainly been focused on centralnervous system diseases. The reasons for this areparticularly numerous and diverse:Microsurgical procedures were not developed•to the point to warrant a precise anatomicaldepiction of nerve injury.Electrophysiological studies were considered•adequate to ascertain the clinical diagnosis.The distribution of mainly low- to midfield•scanners with the at that time existing coiltechnology and scanning hardware and soft-ware did not allow for a reliable depiction ofperipheral nerve structures.MRI of the peripheral nervous system wasmainly performed to rule out compression ofneural bundles from surrounding structures, e.g.,depiction of soft tissue tumors, hematomas, orcysts, as well as anatomical variants such ashypertrophic muscles and ligamentous restrictionor impingement due to scarring upon a peripheralnerve: in most of these cases, it was sufficient toassess the course of the nerve in question withoutthe need to precisely depict the nerve itself. Onlyin few instances as in the carpal tunnel syndrome,an effort was made to assess the diameter and thedegree of flattening of the median nerve underthe retinaculum. Also, the degree of edema withswelling and hyperintensity in T2-weightedimages of the nerve proximal to its entry into thecarpal canal was evaluated in these cases(Mesgarzadeh et al. 1989; Campagna et al. 2009)(Fig. 2.1). More often, the CTS syndrome beingclinically clear and proven by electrophysiologicexaminations, MRI was merely ordered to ruleout tumorous or inflammatory reasons of nerveimpingement.Another indication for performing an MRIexamination is the depiction, classification, andfollow-up of primary neural tumors (Fig. 2.2).Especiallyincasesofhereditaryneurofibromatosiswhere multiple peripheral nerve tumors need tobe observed, MRI offers a variety of advantages:modern scanners allow for multiple scan regionswithin one examination so that eventual centralnervous tumors as well as peripheral schwanno-mas and plexiform neurofibromas in various partsof the body can be evaluated and reliably moni-tored using a repetitive standard imaging tech-nique and documentation. Comparison of theactual images with technically identical imagesof prior exams allows early to pinpoint tumorswith tendency of proliferation and rule out even-tual malignant transformation (Wasa et al. 2010).These repetitive and sometimes “whole-bodyexams” with MRI do not endanger the patientwith lifelong accumulating ionizing radiation.In cases of suspected solitary neural tumors,MRI might help in ascertaining its presence orabsence as in patients with neuroma formationafter limb amputation or in patients with Morton’sneuroma. Although ultrasound plays a predomi-nant role in demonstrating the neural origin ofsuch a tumor, MRI is of great value for the sur-geon providing a clear anatomical picture of thetumor and its surrounding structures in multipleplanes. Thus, MRI greatly facilitates resectionplanning (Figs. 2.3, 2.4).In posttraumatic nerve damage, MR is capableto reveal or rule out nerve root avulsion from thespinal cord. In more distally located nerve dam-age as in peripheral nerve traction injury, thereare at times only subtle signal changes to bedetected in the compromised section along thecourse of the nerve: overstretching of nerve fibersleads to edematous changes with subsequentswelling of the nerve. These changes can beobserved by MRI by comparing the diameter ofthe injured nerve with the contralateral nerve orthe portion of the same nerve proximal and distalto the site of injury and visualizing its water con-tent by using heavily T2-weighted sequenceswith spectral fat suppression. These changes in
  3. 3. 192 Introduction to Magnetic Resonance Imaging of the Peripheral Nervous SystemT2 values of the injured nerve and their timecourse have been shown to correlate with theobserved functional deficits in an animal modelof peripheral nerve traction injury (Shen et al.2010). Clinically, this method is mainly used toassess injuries to the brachial or lumbosacralplexus: the individual nerve roots subsequentlyconjoin to form distinct nerve trunks. All of thesepossess a sufficiently large diameter to makethem readily visible on MR images. These scansshould always include coronal T2-weightedimages with depiction of both sides of the plexus,so that the signal behavior of the affected areacan be compared to the healthy contralateral sideof a patient. Furthermore, MR is very sensitiveto demonstrate collateral damages, like muscularedema or tears, bone fractures, hematomas, ordiffuse edematous changes to the perineural fatindicating the site and severity of a sheertrauma.a bc dFig. 2.1 Recurrent carpal tunnel syndrome (CTS) afterinsufficient dissection of the transverse carpal ligament.Axial T2-weighted images: (a) median nerve (arrow) atthe level of the wrist joint before entering the carpal tun-nel; the nerve is massively enlarged with increased signalintensity on T2 (higher than the signal of the adjacentmusculature) indicating edematous changes of the nerve.(b) Marked flattening of the median nerve (arrow) at theentrance into the carpal tunnel due to the narrowed spacebetween the flexor tendons and the portion of the trans-verse ligament left intact (arrowhead). (c) Dissectedportion of the transverse ligament at the level of the carpaltunnel (arrows). Note the bowing of the ligament nearits insertion at the hamate bone indicating its laxity(arrowhead). The median nerve is decompressed, albeitshowing edematous changes with fluid collections in theepineural spaces surrounding the individual nerve fibers.(d) Typical and normal fibrotic changes after surgery inthe subcutaneous fatty tissue of the palm of the hand(arrows). The edematous reaction of the median nerve dueto the entrapment more proximally extends as far as itsbifurcation
  4. 4. 20 W. JudmaierPeripheral nerves can reliably be imaged inMRI down to a diameter of about 2 mm. InT1-weighted images, used for a detailed anatomicsurvey, nerve structures have intermediate signalintensity, comparable to peripheral musculature.In order to be able to separate nervous from mus-cle fibers, we therefore have to rely on tissue ofdifferent signal behavior that separates nervousstructures from musculature or skeletal bones.This is well known from sagittal images of thespine where the nerve roots and ganglia can beeasily identified within the neuroforamen at thelevel of the thoracic and lumbar spine due to thesurrounding perineural fat. In the cervical spine,however, there is substantially less fatty tissuewithin the neuroforamen, so that the nerve root isless easily identifiable, especially in patients withlow body mass index. In large nerves, there isalways a sufficient amount of surrounding fatcausing high signal in T1-weighted images toabFig. 2.2 (a) Axial T1-weighted image at the level of theright shoulder joint. A slightly oval mass of indeterminateorigin (lymph node, primary soft tissue tumor) is noted inthe axillary space (arrowhead). Note the progressive, conelike enlargement of the nerve at the proximal and distalmargin of the tumor. (b) Coronal T2-weighted inversionrecovery images (TIRM) show the mass (arrowhead) aris-ing from the enlarged and swollen trunk of the brachialplexus (arrows) indicating a primary peripheral nerve tumor.The lesion was histologically classified as a schwannoma
  5. 5. 212 Introduction to Magnetic Resonance Imaging of the Peripheral Nervous Systemmake them easily identifiable. In even largernerves, like the sciatic nerve, there is also adiposetissue interspersed between the perineural sheetsof the individual fascicles which renders themvisible. In smaller nerve fibers that abut muscula-ture or have an intramuscular course the lack ofsurrounding fatty tissue impedes their detection,even when their diameter is substantially largerthan the minimal resolution provided by theimaging sequence. Therefore, routine use of MRIin the peripheral nervous system is somewhatrestricted: apart from the cervical-brachial andlumbosacral plexus, larger nerves of the upperextremity like the radial, ulnar, and median nervedown to the forearm and the median and ulnarnerve at the level of the wrist are routinely exam-ined by MRI. In the lower extremity, the sciaticnerve as well as the tibial and peroneal nerves canbe imaged down to the level of the lower leg(Maravilla and Bowe 1998).Fig. 2.3 (a) SagittalT1-weighted (left) andT2-weighted inversionrecovery image (TIRM –right) of the left upper armshowing a “tumor on a string”(arrow): the enlarged mediannerve proximal and distal ofthe tumor can be clearlyidentified due to thesurrounding fat in theT1-weighted image (left) andthe bright edema on theT2-weighted image (right).Note the progressive,cone-like enlargement of thenerve at the proximal anddistal to the tumor. (b) AxialT2-weighted image showingthe extent of the tumor withinthe median bicipital grooveand its internal inhomogene-ity (arrows). (c) Series ofaxial T2-weighted imageswith spectral fat saturation:enlarged and abnormallybright median nerve(arrowhead) giving riseto the mass (arrows)ab
  6. 6. 22 W. JudmaierThere are limitations in the assessment ofpathologies of smaller nerves with MRI, espe-cially in patients with degenerative nervous dis-eases. In these cases, there is only a very limitededematous reaction of the nerve secondary tothe disease so that T2-weighted images fail todepict an enlarged and unusually hyperintensenerve. Contrast administration shows a mark-edly increased uptake in cases of acuteinflammation; however, in chronic changes theinflammatory changes are much too subtle toreveal pathologic hyperperfusion. Nevertheless,MRI can be of great diagnostic value also inthese cases: denervation of musculature causessignal changes in the muscle groups affected,which can readily be depicted by MR imaging.Muscle edema in early denervation and fattydegeneration and atrophy of the musculature ata later stage of a disease pinpoint the site of ner-vous damage, although the structural damage ofthe nerve itself cannot be visualized (Chhabrandand Andreisek 2012).cFig. 2.3 (continued)
  7. 7. 232 Introduction to Magnetic Resonance Imaging of the Peripheral Nervous SystemabFig. 2.4 Schwannoma (asterisk) of the ulnar nerve: (a)sagittal TSE T1 (left) and T2 TIRM (right) images of thehand. At the proximal end of the tumor, a portion of theulnar nerve can be identified with a cone-shaped enlarge-ment connected to the tumor: mass with “nervous tail.”(b) Axial TSE T1- (top) and T2-weighted (bottom) imagesshowing the extent and the inhomogeneous signal charac-teristics of the schwannoma (asterisk)
  8. 8. 24 W. Judmaier2.2 Measurement Sequencesand Protocol DesignImaging of peripheral nerves requires first of all avery detailed anatomical survey of the region ofinterest. This is best achieved by spin echo orturbo spin echo T1-weighted sequences. Theseimages should be angulated along the directionof a nerve under investigation and, in a secondplane, perpendicular to a nerves’ course produc-ing axial cross-sectional images. These sequencesyield a high tissue signal thus allowing forvery small imaging voxels with little backgroundnoise that would otherwise obscure image con-trast. These sequences can be measured withalmost all commercially available scanners; how-ever, the stronger the static magnetic field, thehigher the resonance signal gets. In lower-fieldscanners the same signal increase, and thus theability to measure equally fine detailed imagescan be achieved by increasing the number ofexcitations: two excitations means that the sameimage plane is measured twice and the signal isinternally averaged prior to image reconstruction.The gain in signal-to-noise ratio is unfortunatelyonly by a factor of √⎯2. However, an increase ofthe base magnetic field from 1.5T to 3T offers anapproximately twofold signal gain. In order toachieve the same signal-to-noise ratio with the1.5T scanner, four excitations would therefore beneeded thus quadrupling imaging time: althoughlower-field scanners allow for adequately detailedanatomical imaging to assess the peripheral ner-vous system, time limitations could make suchdiagnostic procedures problematic in daily clini-cal practice.Apart from the static magnetic field strength,coil technology is crucial for an optimal exploitof a resonance signal. Coils are antennas, pickingup the resonance signal from within the patient’sbody. The smaller the coil, the better its signalgain; however, the overall scan area and the depthof penetration are at the same time reduced.Sophisticated multichannel array surface coilsgreatly facilitate high-resolution imaging bycombining a set of smaller coils into one coilcompound. This technology is quite demandingon scanner hardware and software and thereforecost intensive but allows for detailed imaging ofmore extensive areas of the patient’s body whilemaintaining reasonable scan times.The second set of sequences routinely used inimaging of peripheral nerves are turbo spin echoT2-weighted images with frequency-selectivefat saturation. Alternatively, T2-weighted shortTI inversion recovery images (STIR) can bemeasured – a comparable sequence where a dif-ferent method to suppress the fat signal is used.These sequences are even more demanding forthe scanner as they yield substantially less signalfrom the area under examination and the imagescan easily be obscured by background noiseespecially in low-field scanners. These imageshowever are of high diagnostic interest: theymainly show the water content within anatomicstructures and thus are very sensitive to edema-tous changes in any tissue. Whereas a normal,small peripheral nerve can be overlooked onthese images, it is clearly shown as soon as it isenlarged due to an edema and its accordinglyelevated water content (Fig. 2.5). This effect canequally be detected in cases of nerve entrapmentor in posttraumatic lesions as in traction injury(mechanical edema), in infection (inflammatoryedema), or in tumors where a combination ofmechanical edema and inflammatory edemacaused by an immunological reaction to thetumor growth occurs.The use of contrast agents is generally notwarranted in peripheral nervous imaging.Intravenous gadolinium should however beadministered in patients with suspected infec-tious disease: inflammatory changes lead to ahyperperfusion of nerve bundles that can be visu-alized by “dying” the blood with a contrast agent.Inflammatory hyperpermeability of the endothe-lium also causes increased extravasation and thusaccumulation of contrast agent in the extracellu-lar space. To increase conspicuity, a fat-saturatedT1 sequence should be used to cancel the brightsignal from normal adipose tissue so that only thehyperintense signal from (hyper)perfused tissueremains. To distinguish contrast agent effectsfrom normal bright structures in an image, it ismandatory to measure the exact same sequencebefore and after contrast administration.
  9. 9. 252 Introduction to Magnetic Resonance Imaging of the Peripheral Nervous SystemIntravenous contrast agents should also beused in all cases where tumor formation is sus-pected. Gadolinium gives important informationabout the degree of vascularization of the tumor,helps to identify necrotic changes, and can attimes aid in classifying the entity of the tumor.Alternatively to fat-saturated spin echo orturbo spin echo T1 sequences, sometimes shortTE volumetric 3D gradient echo sequences witheither fat saturation or selective water excitationare used for pre- and postcontrast imaging.Although these sequences generally have a veryhigh T1 contrast showing contrast agent accumu-lation in tissue with great conspicuity, their ana-tomical clarity is inferior and they are prone toproduce various imaging artifacts.Similarly, T2*-weighted 3D sequences canalso be used instead or in adjunction to T2 orSTIR sequences for the visualization of edema-tous changes of peripheral nerves. The use ofsuch volumetric data acquisition methods isadvantageous when secondary reconstructions ofthe acquired 3D data sets are needed. So, even atortuous course of a nerve can be visualized in asingle plane by calculating secondary curvedreconstructions.Apart from conventional imaging with T1-and T2-weighted spin echo or turbo spin echosequences and 3D imaging, various modernand more sophisticated imaging techniquesare presently under investigation for assessingperipheral nerves anatomically and possibly alsofunctionally.One of these novel approaches is called mag-netization transfer imaging and emphasizes onthe ratio and exchange rate of protein-bound andfree water protons present in nerve tissue(Gambarota et al. 2012). Although the measuredmagnetization transfer ratio shows a physiologicvariation between differently sized nerves (pos-sibly due to differences in fascicle content), itseems to be sensitive for nerve damage withinequally sized nerves or when compared to ahealthy contralateral side. However, the value ofthis method has yet to be proven in daily clinicalpractice.Diffusion-weighted imaging (DWI) on theother hand is a standard imaging scheme andwidely used in central nervous system imaging. Ithas also been used to facilitate the detection ofperipheral nerves in a whole-body MR examina-tion (Yamashita et al. 2009). Apart from theintensity of the random microscopic motion(Brownian motion) of water molecules, their pre-vailing direction can also be assessed in diffu-sion-weighted sequences. Since the motion ofFig. 2.5 (a) Axial T1-weighted image of the elbow at thelevel of the humeral condyle. The arrow indicates anenlarged ulnar nerve surrounded by perineural fat andabutting the bone. (b) Axial T2-weighted imagedemonstrating edematous changes within the swollennerve (arrow). (c) Axial reconstructions of a 3D T2*-weighted sequence (DESS 3D) with selective waterexcitation: the ulnar nerve (arrows) can be easily followedin its course proximal, within, and distal to the ulnar nervegroove due to its bright signal (neural edema). (d) Singlesource image of the original 3D DESS data set in coronalorientation showing the portion of the ulnar nerve (arrow)running posterior to the medial humeral epicondylea b
  10. 10. 26 W. Judmaierwater molecules is impeded by membranes, itsforemost direction follows the course of nervebundles (Fig. 2.6). In the central nervous system,these anisotropic diffusion properties of nervoustracks are used to identify the course of nervefibers within the medulla and the brain. One of itsclinically important uses is to preoperatively helpidentify displaced tracts and thus permit to savetheir integrity when a space-occupying lesionneeds to be resected. This variant of DWI is calleddiffusion tensor imaging (DTI) and shows prom-ising results also for the peripheral nerve system,not only showing the nerve bundles themselvesbut also permitting to assess their integritycdFig. 2.5 (continued)
  11. 11. 272 Introduction to Magnetic Resonance Imaging of the Peripheral Nervous System(Skorpil et al. 2007; Kakuda et al. 2011). Thismeasurement scheme, however, suffers from aninherent low signal-to-noise level. To obtain sat-isfactory images in small peripheral nerves, it ismandatory to use a high-performance gradientsystem and, preferably, a high-field MR scanner.ReferencesCampagna R et al (2009) MRI assessment of recurrent car-pal tunnel syndrome after open surgical release of themedian nerve. AJR Am J Roentgenol 193(3):644–650Chhabrand A, Andreisek G (2012) Magnetic resonanceneurography. Jaypee Brothers Medical Publishers,New DelhiGambarota G et al (2012) Magnetic resonance imaging ofperipheral nerves: differences in magnetization trans-fer. Muscle Nerve 45:13–17Kakuda T et al (2011) Diffusion tensor imaging of periph-eral nerve in patients with chronic inflammatorydemyelinating polyradiculoneuropathy: a feasibilitystudy. Neuroradiology 53(12):955–960Maravilla KR, Bowe BC (1998) Imaging of the peripheralnervous system: evaluation of peripheral neuropathy andplexopathy. AJNR Am J Neuroradiol 19:1011–1023Mesgarzadeh M et al (1989) Carpal tunnel: MR imaging. PartII. Carpal tunnel syndrome. Radiology 171:749–754Shen J et al (2010) MR neurography: T1 and T2 measure-ments in acute peripheral nerve traction injury in rab-bits. Radiology 254(3):729–738Skorpil M et al (2007) Diffusion-direction-dependentimaging: a novel MRI approach for peripheral nerveimaging. Magn Reson Imaging 25(3):406–411Wasa J et al (2010) MRI features in the differentiation ofmalignant peripheral nerve sheath tumors andneurofibromas. AJR Am J Roentgenol 194(6):1568–1574Yamashita T et al (2009) Whole-body magnetic resonanceneurography. N Engl J Med 361:538–539Fig. 2.6 Recurrent CTS(same patient as in Fig. 2.1).Direction-sensitive diffusion-weighted imaging called“diffusion tensor imaging”(DTI) allows for a 3Dreconstruction of themolecular water motionwithin the nerve fascicles.Intact nerves can be followedby using seed points fromwhere the computer generatesimages along a nerve’s courseand displays color coded itsdirection. Note the compres-sion of the median nerve at itsentrance into the carpal tunneldepicted on one of the sagittalsource images (arrow)
  12. 12. http://www.springer.com/978-3-642-25593-9

×