(4.1.3)--5、Preoperativevisualizationofneu脑科学与影像新技术.pdf

《(4.1.3)--5、Preoperativevisualizationofneu脑科学与影像新技术.pdf》由会员分享，可在线阅读，更多相关《(4.1.3)--5、Preoperativevisualizationofneu脑科学与影像新技术.pdf（6页珍藏版）》请在淘文阁 - 分享文档赚钱的网站上搜索。

1、RIGEMINALneuralgia is a unique form of neuropathicpain most commonly caused by vascular compres-sion at the trigeminal nerve root entry zone in theCPA.Microvascular decompression of the trigeminal nerveproduces long-term improvement of symptoms in most pa-tients(70%are pain free at 10 years3),and th

2、e presence andseverity of vascular compression is highly correlated withoutcome after MVD.3,5Accurate preoperative character-ization of pathological neurovascular compression couldtherefore have an impact on the determination of appropri-ate treatment for TN.However,because the resolution istypicall

3、y not sufficient to allow meaningful interpretationof the complex anatomy of the CPA,conventional imagingtechniques have been shown to be inadequate.10Recently,2 strategies have been used in an attempt to im-prove the preoperative detection of vascular compression ofthe trigeminal nerve.First,high-r

4、esolution 3D steady-stateprecession MR imaging sequences,such as constructive in-terference in the steady state or BFFE,have been used toproduce highly T2-weighted images with an intense con-trast border between the CSF and solid structures.Theseimages are acquired using a volume acquisition at near

5、 iso-voxel resolution,are easily reconstructed in any plane withexcellent quality,and can be used to produce a 3D imageof the CPA viewed from any angle,effectively emulat-ing cisternoscopy.1,68,11Second,3D TOF MR angiographyand Gd-enhanced T1-weighted or spoiled gradient recalledsequences have been

6、used to objectively distinguish thenerves,arteries,and veins of the posterior fossa.2Balanced fast-field echo,MR angiography,and 3D Gd-enhancement each have advantages and significant limita-tions in their ability to detect neurovascular compression.Balanced fast-field echo scanning produces a very

7、high-resolution T2-weighted image with stark contrast betweenhigh-intensity CSF and all other structures,such as cra-nial nerves,blood vessels,brainstem,cerebellum,and sur-rounding bones outlined as low-intensity areas.Very fineanatomic detail is visible,but it is difficult to differentiateJ.Neurosu

8、rg./Volume 108/March 2008J Neurosurg 108:477482,2008Preoperative visualization of neurovascular anatomy intrigeminal neuralgiaJONATHANMILLER,M.D.,1FERIDUNACAR,M.D.,1BRONWYNHAMILTON,M.D.,2ANDKIMBURCHIEL,M.D.1Departments of 1Neurological Surgery and 2Diagnostic Radiology,Oregon Health&ScienceUniversit

9、y,Portland,OregonObject.The authors report on a novel technique to identify neurovascular compression in trigeminal neuralgia(TN).Using 3D reconstructed high-resolution balanced fast-field echo(BFFE)images fused with 3D time-of-flight(TOF)magnetic resonance(MR)angiography and Gd-enhanced 3D spoiled

10、gradient recalled sequence,it is possible to objec-tively visualize the trigeminal nerve and nearby arteries and veins.Methods.Magnetic resonance imaging was performed in 18 patients with unilateral TN using 3 sequences:BFFE,3D TOF angiography,and 3D Gd-enhanced imaging.The images were imported into

11、 OsiriX imaging software;aftertheir fusion,a 3D false-color reconstruction was produced using surface rendering.The reconstructed images objective-ly differentiate nerves and vessels and can be viewed from any angle,including the anticipated surgical approach.Results.Fifteen patients were predicted

12、to have neurovascular compression on the symptomatic side(9 arterial and6 venous compressions).All patients had a vascular structure that was identical in location and configuration to thatpredicted on preoperative analysis.The 3 patients without predicted compression underwent surgical exploration

13、be-cause they manifested the classic symptoms.As expected,exploration in 2 of these patients revealed no offending ves-sel.The third patient had a small vein embedded in the trigeminal nerve that was beyond the resolution of the 3D Gd-enhanced study.Conclusions.Combining BFFE with MR angiography and

14、 Gd-enhanced MR images capitalizes on the advantagesof both techniques,enabling MR angiography and contrast-enhanced MR imaging discrimination of vascular structuresat BFFE resolution.This results in an unambiguous 3D image that can be used to identify the neurovascular compres-sion and plan the sur

15、gical approach.(DOI:10.3171/JNS/2008/108/3/0477)KEYWORDSbalanced fast-field echo neurovascular compression OsiriX imagingtrigeminal neuralgia 3D magnetic resonance imaging T477Abbreviations used in this paper:BFFE=balanced fast-fieldecho;CPA=cerebellopontine angle;CSF=cerebrospinal fluid;MR=magnetic

16、 resonance;MVD=microvascular decompression;TN=trigeminal neuralgia;TOF=time-of-flight.artery and vein from nerve.Although it is sometimes pos-sible to deduce the identity of a structure by following it toits origin,this can be misleading and requires subtle judg-ment calls and a detailed anatomical

17、knowledge;vascularstructures embedded in neural structures are invisible.Onthe other hand,MR angiography and 3D Gd-enhanced im-ages differentiate nerves from blood vessels,but at presentdo not provide sufficient resolution with which to detectvery small structures reliably,leading to an unacceptably

18、low negative predictive value.2In the present study,we demonstrate that these 2 ap-proaches can be coupled into a strategy that combines thestrengths of both techniques,allowing for MR angiographyand 3D Gd-enhanced discrimination of nerves and vesselsat BFFE resolutions.This strategy provides a clea

19、r and ob-jective preoperative diagnostic tool for patients with TN.Using 3D-reconstructed BFFE images fused with 3D TOFMR angiography and 3D Gd-enhanced images,it is possi-ble to objectively visualize the nerves,arteries,and veins ofthe CPA with an unprecedented level of anatomic detail.Clinical Mat

20、erial and MethodsImaging ProtocolEighteen patients with symptoms characteristic of uni-lateral TN underwent prospective MR imaging examina-tions performed using a 3-T MR scanner(Philips Achieva),and images were obtained using a dedicated commerciallyavailable 6-channel head coil with sensitivity enc

21、oding(SENSE,Philips)parallel processing capability.Each pa-tient underwent imaging using a BFFE technique centeredon the pons in the region of the trigeminal nerve.Sequenceparameters are outlined in Table 1.Sequence parametersfor BFFE scans included a repetition time of 5.8 msec,anecho time of 2.4 m

22、sec,number of acquisitions 3,and a flipangle of 458.The scan time was between 3 minutes 11 sec-onds and 5 minutes 29 seconds.The field of view was 15cm with a 512 3 512 matrix and 0.6-mm slice thickness,producing isotropic voxels of 0.29 mm 3 0.29 mm 3 0.29mm.High-resolution 3D TOF angiography and 3

23、D spoiledgradient recalled sequences after injection of 0.1 mmol/kgintravenous Gd-diethylenetriaminepentaacetate were per-formed as previously described.2Postprocessing ProtocolPostprocessing was performed on an Apple MacintoshPower PC laptop(Apple,Inc.)running OS X and open-source OsiriX imaging so

24、ftware(Version 2.5.1;free down-load from http:/www.osirix- BFFE im-ages were inverted using the“Invert Data”plug-in,andthen fused to the 3D Gd-enhanced and MR angiographyimages.Regions of interest were identified that includedthe length of the trigeminal nerve and any structures in itsvicinity on bo

25、th the symptomatic and asymptomatic sides.The OsiriX surface-rendering algorithm was used to pro-duce 3D reconstructions of the BFFE images by setting thelower threshold to just above the signal intensity of CSF.Asecond surface-rendered image from the 3D Gd-enhancedor MR angiography data was superim

26、posed on this imageto produce a single false-color reconstruction of the CPA.Vessels visualized on angiography were defined as arteries,and those appearing on 3D Gd-enhanced images but not onangiography were identified as veins.This technique pro-duced 2 fused images:1(BFFE MR angiography)objec-tive

27、ly demonstrates the trigeminal nerve and nearby arter-ies,and the other(BFFE 3D Gd-enhanced)can be used todetermine the venous anatomy.The region of interest wasthen expanded to include the basilar artery and identify theoffending vessel.The images were rotated to be viewedfrom various angles,includ

28、ing the expected surgical ap-proach,producing a cisternogram that demonstrated neuralanatomy at very high resolution,with vascular anatomy vi-sible in a different color.Surgical TreatmentBased on symptoms and neuroimaging results,18 pa-tients underwent surgery.In each case,MVD was per-formed via a r

29、etromastoid craniotomy and exploration ofthe root entry zone of the affected nerve.The intraoperativemicroscopic view was compared with the preoperative re-constructions with special attention to the identity and na-ture of the offending vessel.ResultsDemographic DataEighteen patients were included

30、in this study,11 werewomen(61%),and the mean patient age was 52.9 years(range 2680 years).The imaging studies in all 18 patientswere successfully reconstructed using the OsiriX software.Five patients(28%)reported symptoms on the left side,and13(72%)reported symptoms on the right.Thirteen patients(72

31、%)described TN Type 1(intermittent and lancinating)pain,and 5(28%)described TN Type 2(constant)pain.4Patient characteristics are detailed in Table 2.Magnetic Resonance Imaging DataAfter image processing,the trigeminal nerve was clear-ly visible in each case,as were vascular structures in thevicinity

32、 of the nerve.Of the 18 patients in the study,9J.Miller et al.478J.Neurosurg./Volume 108/March 2008TABLE 1Sequence parameters for MR imaging studies*ModalityTR(ms)TE(ms)Flip Angle()FOV(cm)MatrixSlice Thickness(mm)VoxelBFFE5.8 2.4 4515 512 3 5120.6 0.29 mm3MRA33 2.4 2525512 3 5121 0.5 mm23 1.0 mmGd-D

33、TPA33 2.4 2525256 3 2561 1.0 mm3*DTPA=diethylenetriaminepentaacetate;FOV=field of view;MRA=magnetic resonance angiography;TE=echo time;TR=repetition time.(50%)were predicted to have arterial compression,6(33%)were predicted to have venous compression,and 3patients(16%)were predicted to have no vascu

34、lar com-pression despite manifesting the classic symptoms of TN.Six patients had some degree of neurovascular compres-sion predicted on the asymptomatic side(4 arterial and 2venous),although compression was more pronounced con-tralateral to symptoms in only 1 patient.The presence ofTN Type 1 symptom

35、s trended toward a greater likelihoodof arterial compression(78%for TN Type 1 compared with40%for TN Type 2 symptoms),but because of the smallsample size,the difference was not statistically significant(p=0.18,Fisher exact test).Surgical DataOf the 15 patients predicted to have neurovascular com-pre

36、ssion,100%were found at surgery to have compressionby an arterial or venous structure identical in location andconfiguration to that predicted on preoperative analysis.Representative images and intraoperative views for arterial(Fig.1)and venous(Fig.2)compression are shown.Of the3 patients in whom no

37、 compression was predicted,surgicalexploration in 1 patient revealed a very small vein embed-ded in the nerve(Fig.2,bottom row),and the other 2 pa-tients had no identifiable lesions at surgery.DiscussionIn recent years,there has been rapid development of di-agnostic imaging modalities for many neuro

38、surgical condi-tions.The most successful strategies applied to the eval-uation of TN have been 3D visualization of solid structuresby BFFE and demonstration of vascular structures by MRangiography and enhanced MR imaging.By combiningthese techniques,each complements the limitations of theother,and i

39、n many ways,capitalizes on the advantages ofeach.For instance,vessels can be traced forward on BFFEeven after they have become too small to be observed usingMR angiography and enhanced MR images,thus improv-ing sensitivity.The addition of objective vascular data toBFFE makes the images clear and una

40、mbiguous so that thecomplex orientation of structures around the brainstem isobvious even to the layperson.Most importantly,the useof this technology makes it possible to reliably grade thenature,degree,and location of the vascular compressionin patients with different symptoms and prognoses,whichsh

41、ould be helpful in studying the most effective treatmentfor TN.We report on a novel technique for identifying neurovas-cular compression in cases of TN using 3D reconstructedhigh-resolution BFFE images fused with 3D TOF MR an-giography and 3D Gd-enhanced MR images to objectivelyvisualize the trigemi

42、nal nerve and nearby arterial and ve-nous structures.To achieve this we used OsiriX imagingsoftware.OsiriX,a free,open-source Digital Imaging andCommunications in Medicine medical imaging softwaresystem,can be easily installed on any Macintosh computerthat meets the minimum system requirements(OS X

43、with2 gigabytes random access memory).The interface is intu-itive and interactive,and the reconstruction and fusion toolsare fast and accurate.This system has proven very usefulfor image manipulation because it does not require a dedi-cated workstation,and images can be processed by the sur-geon in

44、a clinical environment.Magnetic resonance imagesthat are immediately available to the consulting physiciantake on average 10 minutes to fuse and reconstruct,and cansubsequently be viewed and shared with the patient prior tocompletion of the consultation.In our clinic,the patientswere shown their rec

45、onstruction images,and in our expe-rience,this has proved helpful in discussing the causes oftheir symptoms and the proposed treatment plan.Recently,Satoh et al.9reported that volume-rendered fu-sion images using volumetric T2-weighted fast spin-echoand TOF MR angiography sequences can be used to de

46、lin-eate neurovascular anatomy.The method we describe dif-fers from the method of Satoh and colleagues in 3 respects.First,we used 3D Gd-enhanced imaging as well as MR an-J.Neurosurg./Volume 108/March 2008Preoperative visualization of neurovascular anatomy479TABLE 2Characteristics and findings in 18

47、 patients who underwent MVD for TN*Case No.Age(yrs),SexSideTN TypeSymptom Duration(mos)Predicted Offending VesselActual Offending VesselOutcome180,Mrt115SCASCApain free226,Flt23veinveinimproved335,Flt130nonenonepain free453,Frt130veinveinimproved533,Frt172SCASCApain free653,Frt17noneveinpain free771

48、,Mrt1108SCASCApain free862,Frt218SCASCAimproved962,Mrt1384SCASCApain free1065,Frt1120SCASCApain free1154,Flt136nonenonepain free1256,Mrt19SCASCApain free1337,Mrt26veinveinpain free1455,Frt11.5veinveinpain free1546,Frt172veinveinpain free1659,Mrt284SCASCApain free1751,Flt2120veinveinimproved1854,Mlt1

49、36SCASCApain free*SCA=superior cerebellar artery.giography so that venous and arterial anatomy was visible.Second,with the use of a 3-T field strength,it was possibleto achieve a level of resolution for all 3 sequences signifi-cantly higher than that practically available using 1.0-T or1.5-T techniq

50、ues,allowing for visualization of very smallarteries and veins that would not otherwise have been visi-ble.Finally,we used surface rendering techniques with astandard isosurface threshold rather than volume rendering,which ensures images from different patients are objective-ly comparable.One signif