author's personal copy - duke universitytransducers.bme.duke.edu/pubs/umb_2009_329-338.pdfple...

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

doi:10.1016/j.ultrasmedbio.2008.08.016

● Original Contribution

THE ULTRASOUND BRAIN HELMET: FEASIBILITY STUDY OFMULTIPLE SIMULTANEOUS 3D SCANS OF CEREBRAL VASCULATURE

STEPHEN W. SMITH,* NIKOLAS M. IVANCEVICH,* BROOKS D. LINDSEY,* JOHN WHITMAN,*EDWARD LIGHT,* MATTHEW FRONHEISER,* HEATHER A. NICOLETTO,† and

DANIEL T. LASKOWITZ†

*Department of Biomedical Engineering, Duke University, Durham, NC, USA; and †Division of Neurology, DukeUniversity Medical Center, Durham, NC, USA

(Received 26 May 2008; revised 25 June 2008; in final form 21 August 2008)

Abtract—We describe early stage experiments to test the feasibility of an ultrasound brain helmet to producemultiple simultaneous real-time three-dimensional (3D) scans of the cerebral vasculature from temporal andsuboccipital acoustic windows of the skull. The transducer hardware and software of the Volumetrics MedicalImaging (Durham, NC, USA) real-time 3D scanner were modified to support dual 2.5 MHz matrix arrays of 256transmit elements and 128 receive elements which produce two simultaneous 64° pyramidal scans. The real-timedisplay format consists of two coronal B-mode images merged into a 128° sector, two simultaneous parasagittalimages merged into a 128° x 64° C-mode plane and a simultaneous 64° axial image. Real-time 3D color Dopplerscans from a skull phantom with latex blood vessel were obtained after contrast agent injection as a proof ofconcept. The long-term goal is to produce real-time 3D ultrasound images of the cerebral vasculature from aportable unit capable of internet transmission thus enabling interactive 3D imaging, remote diagnosis and earliertherapeutic intervention. We are motivated by the urgency for rapid diagnosis of stroke due to the short timewindow of effective therapeutic intervention. (E-mail: [email protected]) © 2009 World Federation forUltrasound in Medicine & Biology.

Key Words: Transcranial, Ultrasound contrast, Intracranial arteries, 3D imaging, Phase aberration.

INTRODUCTION

Stroke encompasses a wide variety of cerebrovasculardiseases including ischemic stroke and hemorrhagicstroke. Stroke is the third leading cause of death in theUnited States, responsible for 160,000 deaths in the year2000 with a prevalence of 11.3 per 1000 or 4,000,000individuals. It is the leading cause of disability amongadults in the United States, costing some $56 billion pery (Gorelick and Alter 2002). As our population ages, thepublic health burden and number of deaths associatedwith stroke are expected to triple to approximately500,000 deaths by the year 2050.

At present, i.v. treatment with a thrombolytic drug,tissue plasminogen activator (tPA), is the only USFDA-approved pharmacologic intervention demonstrated toimprove mortality and functional outcome in ischemicstroke; however, i.v. thrombolysis is only effective whenadministered within three hours of the onset of symp-

toms (Marler et al. 1995). Unfortunately, the symptomsof stroke may vary widely depending on the location ofcerebral ischemia, and physician reluctance to make aclinical diagnosis contributes to the low utilization oftPA therapy for stroke patients. Thus, speed and diag-nostic accuracy are essential to tailor the treatment of theindividual patient during stroke onset and progression.There have been prolific advances in technologies toimage brain and cerebral vasculature that enable a morecomplete understanding of the pathophysiology of strokeand improve the safe management of stroke patients. Forexample, progress in neuroimaging includes visualiza-tion of cerebrovascular changes via computed tomogra-phy angiography (CTA), and magnetic resonance an-giography (MRA) in addition to traditional digital sub-traction angiography (Hennerici 2003).

However, these modalities suffer limitations of cost,availability, complexity and ease of use in the uncoop-erative patient. For example, Hanley and Hacke (2005)cite the Genentech Stroke survey, which demonstratedthat the mean time from stroke symptom onset to a

Address correspondence to: Stephen Smith, Box 90281, DukeUniversity, Durham, NC, USA 27705. E-mail: [email protected]

Ultrasound in Med. & Biol., Vol. 35, No. 2, pp. 329–338, 2009Copyright © 2009 World Federation for Ultrasound in Medicine & Biology

Printed in the USA. All rights reserved0301-5629/09/$–see front matter

329


computed tomography (CT) examination was 4 h, i.e.,already outside the time window set by the NationalInstitute of Neurologic Disorders and Stroke for use ofthrombolytic tPA (Marler et al. 1995). The paper alsonotes a population study demonstrating that only 11% ofpatients are eligible for tPA. Thus, what is needed is areal-time, portable, noninvasive, low cost bedside tech-nology for the rapid evaluation of the stroke patient thatwould be readily available in the prehospital and emer-gency room setting. It is our contention that real-timethree-dimensional (3D) transcranial ultrasound technol-ogy with improved image quality via correction of thephase aberration due to the skull bone is the answer tothese requirements.

Since the 1980s, transcranial ultrasound has been avaluable tool in neurology for the diagnosis and evalua-tion of stroke. In the transcranial ultrasound examination,a sonographer examines the major cerebral arteriesthrough the temporal and suboccipital acoustic windowsof the skull (Babikian and Wechsler 1999). The state ofthe art in transcranial ultrasound consists of a phased-array sector scan, operating at approximately 2 MHz,applied to the temporal and suboccipital acoustic win-dows, combined with the administration of microbubblecontrast agent and examination of blood vessels withcolor and spectral Doppler. Contrast-enhanced examina-tions have shown advantages over traditional ultrasoundin assessing ischemic cerebrovascular disease by visual-izing intracranial arteries. The extensive literature oftranscranial ultrasound imaging has been well reviewedin texts by Tegeler et al. (1996) and by Bogdahn et al.(1998). The continued success of this modality, in spiteof its limitations in image quality, points to its inherentstrong advantages over CT, MR and angiography in aclinical setting. All transcranial ultrasound is real-time,portable, noninvasive and low cost. It is a bedside tech-nology ideal for the restless or uncooperative patient andretains widespread use for acute and intensive care ap-plications as well as for therapy monitoring and periop-erative management.

Recently, a few laboratories have described off-linereconstructed 3D transcranial ultrasound (US) scanning,achieved by rotation of the transducer or electromagneticpositioning of the transducer. Advantages were demon-strated in the assessment of collateral circulation (Wes-sels et al. 2004), the diagnosis of vascular anatomy andlesion vascularity for tumors (Bauer et al. 1998) and theanalysis of intracranial aneurysms (Klötzsch et al. 1999).

In our own laboratories, we have concentrated onreal-time 3D ultrasound scanning originally developed atDuke University (Smith et al. 1991; von Ramm et al.1991) and commercialized by Volumetrics Medical Im-aging, Inc. (VMI, Durham, NC, USA). In 2004, wedescribed real-time 3D scans of the brain (Smith et al.

2004). In 2008, in the first human trial of real-time 3Dtranscranial ultrasound, we used our scanner from thetemporal and suboccipital acoustic windows with Defin-ity contrast agent enhancement (Lantheus Medical Im-aging, Billerica, MA, USA), to produce simultaneoustranscranial axial, coronal and parasagittal image planesof the human brain as well as steerable 3D spectralDoppler traces and 3D color flow images of cerebralvessels in seventeen normal subjects (Ivancevich 2008).In a single subject, we showed human in vivo correctionof the phase aberration of the skull. We used the echosignals from the brain tissue combined with the multi-lagleast-means-squares cross-correlation algorithm (Flaxand O’Donnell 1988; Liu and Waag 1994; Gauss 2001)previously adapted to 3D ultrasound (Ivancevich et al.2006). We compared the control and postskull correction3D ultrasound angiograms. The resulting corrected im-ages yielded an increase in detected cerebral vessels asdetermined by blinded observations of two neurosonolo-gists.

The success of that human trial led us to the subjectof this article, a feasibility study of an ultrasound brainhelmet that would include multiple simultaneous real-time 3D scans of the cerebral vasculature from the tem-poral and suboccipital acoustic windows of the skull.This article describes the hardware and software modi-fications as well as feasibility testing of the Volumetrics3D scanner to achieve dual simultaneous 3D scans with3D color Doppler from matrix arrays in a human skullphantom. Our long-term goal for the brain helmet is toproduce real-time 3D ultrasound images of the completecerebral vasculature with internet transmission of imagesto a stroke center, enabling remote diagnosis and earliertherapeutic intervention.

METHODS

As shown in Fig. 1A, our 3D ultrasound systemscans a full 64° pyramid using a matrix (checkerboard)array transducer at up to 30 volumes per s. Figure 1Ashows a schematic of the matrix phased array transducerproducing a pyramidal scan and displaying two simulta-neous orthogonal B-mode images, corresponding to axialand coronal image planes, as well as two C-mode planes,corresponding to parasagittal image planes. Real-timedisplay options in our 3D scanner also include 3D vol-ume rendering, 3D color flow imaging and a steerable 3Dultrasound beam (red) to be used for spectral Dopplermeasurements of cerebral blood flow or therapeutic ap-plications such as ultrasound enhanced thrombolysis (Al-exandrov et al. 2004). Figure 1B is a photograph of theVMI matrix array probe which is used for real-time 3Dscanning in both cardiac as well as transcranial applica-tions.

330 Ultrasound in Medicine and Biology Volume 35, Number 2, 2009


In Fig. 2A, we show an illustrative axial image ofthe Circle of Willis (CW), the ipsilateral middle cerebralartery (MCA) and contralateral skull from a normalsubject. By manipulating the thickness and orientation ofthe simultaneous coronal slice from the real-time 3Dscan, we show, in Fig. 2B, a coronal view of that sameMCA demonstrating patency as the vessel makes itstortuous path toward the outer surface of the brain.Figure 2C shows the off-line 3D-rendered view of thecontralateral skull and cerebral vasculature from thesame subject, a 3D ultrasound angiogram. This view canbe tilted and rotated to examine the vasculature from anyperspective. In like manner, for another subject scannedfrom the suboccipital window, Fig. 2D shows the colorDoppler 3D rendering of the vertebral arteries (VA)joining to form the basilar artery (BA).

Figure 3 illustrates the concept of the brain helmet:three matrix array transducers (T1-3) mounted in a sim-ple helmet to produce three simultaneous 3D ultrasoundscans of the brain through the temporal and suboccipitalwindows of the skull. In the first experiment of ourfeasibility study, we simply scanned the brain of a human

subject, who had given informed consent per the IRB-approved protocol (Ivancevich 2008), using two simul-taneous Volumetrics 3D scanners running asynchro-nously. The video output of the two displays was fed intoa video screen splitter (MicroImage, Boyertown, PA,USA) so that the image planes from two simultaneous3D scans could be viewed on a single television monitor.As shown in Fig. 4A through D, the combination of atemporal scan and a suboccipital scan yielded useful 3Dimages. After a Definity contrast injection, a matrixtransducer probe on the temporal acoustic window pro-duced color Doppler data from the cerebral vessels of theCircle of Willis in axial (Fig. 4A) and coronal (Fig. 4B)image planes. Simultaneously, a second matrix array onthe suboccipital window produced the coronal and para-sagittal image planes of Fig. 4C and Fig. 4D, yieldingviews of a vertebral artery (VA) and the atlas loop (AtL)in Fig. 4C and of the foramen magnum (FM) in Fig. 4D.In Fig. 4A through D, the green horizontal lines indicatethe positions of the simultaneous C-mode planes whichwere not shown on the split screen video display.

We were concerned that the two transducers wouldacoustically interfere with each other. This proved to bean insuperable problem when two transducer probeswere positioned across from each other over the twotemporal windows and caused significant interferenceeven for the temporal and suboccipital combination.

Thus, we were convinced of the necessity to modifyour 3D scanner for synchronous operation of multiplematrix arrays. In this project, we relied on our previousexperience developing multiple 3D intra-cardiac echo(3D-ICE) catheters wherein we had modified the 3Dscanner to switch between two 3D-ICE catheters in onesecond at the push of a button (Fronheiser et al. 2006). Innormal operation, the scanner includes 512 transmittersand 256 receive channels with 16:1 receive-mode paral-lel processing to generate 4096 B-mode image lines inthe pyramidal scan. We simulated matrix array designsfor multiple simultaneous transducers for a steering an-gle of (0°, 0°) using the Field II ultrasound simulationsoftware (Jensen and Svendsen 1992) yielding the axi-symmetric beam plots shown in Fig. 5. For each design,as a measure of image quality, we calculated the relativepeak pressure (a surrogate of image sensitivity), the �6dB beam width (equivalent to transducer lateral resolu-tion) and the grating lobe amplitude (a surrogate ofimage clutter). As a demonstration of image quality, foreach design, we also show the experimental C-scan im-ages of a 12 mm cyst phantom (contrast � �40 dB).

In its original configuration, as described by Light etal. (1998), the 2.5 MHz Volumetrics matrix array (� H2O

� 0.6 mm) was configured in a sparse periodic vernierpattern shown in Fig. 5A with 256 transmit (Tx) ele-ments (spacing � 0.35 mm) and 256 receive (Rx) ele-

Fig. 1. (A) Schematic of 2D matrix array scanning a pyramidand displaying simultaneous axial, coronal and parasagittalscans, as well as steerable spectral Doppler (red). (B) Photo-

graph of 2.5 MHz matrix array.

Ultrasound brain helmet ● S. W. SMITH et al. 331


ments (spacing � 0.7 mm). The simulations yielded apulse-echo sensitivity, which we assign to 0 dB, �6 dBbeamwidth � 3.9 mm at a depth of 70 mm, and a gratinglobe amplitude � �45 dB resulting in the associatedC-scan image.

In a later design, shown in Fig. 5B, every availableelement in the array was used in transmit mode resultingin 440 transmit elements and 256 receive elements yield-ing a sensitivity improvement of �4.7 dB relative to thatof Fig. 5A, �6 dB beamwidth � 3.4 mm at a depth of 70mm, but an increased grating lobe amplitude of �24 dB,resulting in the associated C-scan image of the cyst. This

design was used in the transcranial human study ofIvancevich et al. (2008).

For a brain helmet configuration of two simulta-neous 3D transducers, we analyzed the image qualitytrade-offs of a number of matrix array designs includingthe use of multiplexer circuits to switch between trans-ducers. In the end, based on criteria of cost, simplicityand sensitivity, we allocated 256 transmitters and 128receivers to each matrix array in the pattern shown inFig. 5C yielding a relative sensitivity of �6 dB, �6 dBbeamwidth � 5.5 mm at a depth of 70 mm, a grating lobeamplitude of �60 dB and associated C-scan image of the

Fig. 2. Images from real-time 3D transcranial scan including: (A) Axial plane showing the circle of Willis (CW) andipsilateral middle cerebral artery (MCA); (B) simultaneous coronal plane showing the same ipsilateral MCA; (C)off-line 3D rendering of contralateral skull and color Doppler of the cerebrovascular tree including MCA from temporalwindow; and (D) 3D color Doppler rendering of the cerebrovascular tree from suboccipital window including vertebral

arteries (VA) joining to form basilar artery (BA).



cyst. A comparison of the C-scans shows the enlargedspeckle size associated with increased beam width butgood cyst contrast for the brain helmet design.

Having chosen the transducer design, we fabricateda new transducer coupling system wherein the circuitryof 512 transmitters and 256 receivers, which normallyare connected to two ITT Cannon connectors (ModelDLM6–360), were rewired into four of these connectorsfor the two matrix arrays. To complete our hardwaredevelopments, we adapted a Transcranial Doppler Fixa-tion System (Spencer Technologies, Seattle, WA, USA),which is conventionally used for one-dimensional (1D)spectral Doppler, to mount our two simultaneous matrixarray probes. We then constructed a skull phantom byusing a polymer casting of a human skull (3B Scientific,Hamburg, Germany) filled with degassed water. Theskull adequately mimics the compact bone of the tem-poral window. We fixed a plastic mesh in the sagittalplane to mimic the midline falx structure and we sus-pended a latex tube (3 mm inner diameter) into a U-shaped loop lying in the coronal plane within the skull.The completed skull phantom with the transcranial trans-ducer fixation system, latex vessel and the two matrixarrays are shown in the photograph of Fig. 6.

We next modified the scanner software to producedual pyramidal scans. In normal operation, a 64° pyra-midal scan includes 64 � 64 � 4096 image lines. Pre-viously, we have developed 90° and 120° pyramidalscans for our 3D intra-cardiac catheters (Lee et al. 2004)and 3D endoscopes (Pua et al. 2007). In this case, wedeveloped a 64 � 64 � 4096 line pyramidal scan of 64°

x 128°. Under software control, we can enable or disableindividual array elements on each image line, so we splitthis pyramidal scan between the two matrix arrays byenabling only the elements of array no.1 during the first2048 image lines and enabling only the elements of arrayno. 2 during the second 2048 image lines. Thus wecreated two independent pyramidal scans each ofroughly 64°. Note that the scan line spacing was doubledin the elevation direction. All other features of the 3Dscanner, including Doppler capabilities, were unchanged.

As of yet, we have not customized the displayfeatures of the 3D scanner for the brain helmet so that thesimultaneous display planes now included: (1) a 128°elevation (coronal plane) sector scan which consisted ofadjacent 64° elevation planes from the dual transducers;(2) C-scan planes of adjustable orientation and depth,which consisted of adjacent parasagittal planes from thedual transducers; and (3) a 64° azimuth sector scanselected from one of the dual transducers.

RESULTS

Figure 7 shows the results of a real-time 3D scanwith color Doppler of the skull phantom with dual si-multaneous matrix array transducers positioned on thetemporal acoustic windows after a contrast agent injec-tion of agitated tap water into the loop of the latex tube.Figure 7A shows the blood vessel loop simultaneouslyimaged in the elevation (coronal) plane from both sidesof the skull so that the flow, which is away from the lefttransducer, is shown in blue and the flow, which istoward the right transducer, is shown in red. Due to thewidening U-shape of the loop, flow does not reversedirections within the field of view. Each image includesthe contralateral skull. Note also that the left transducershows reduced signal-to-noise ratio compared with theright transducer due to RF interference in the coupler forthe left transducer. Figure 7B shows the simultaneoustilted C-scan (parasagittal) plane cutting across the vesselloop in both scans to produce short axis images of flowin the vessel from the left transducer (blue) and the righttransducer (red).

The dotted yellow line in Fig. 7A shows the plane ofthat tilted C-scan. Note that the tilted C-scan cuts thevessel loop twice for the right transducer (red). Figure 7Cshows the simultaneous axial slice through the vesselloop for the left transducer. This slice orientation isdetermined by the white cursor arrow in Fig. 7A. Finally,after a quick track ball adjustment of that cursor, thedisplay of Fig. 7D appears, showing the axial slicethrough the vessel loop from the right transducer (red).This slice orientation is determined by the yellow arrowin Fig. 7A.

Fig. 3. Schematic of ultrasound brain helmet including simul-taneous temporal matrix arrays (T1, T2) and suboccipital ma-

trix array (T3).



DISCUSSION

We believe that dual simultaneous in vivo 3D scansfrom both temporal windows should provide additionaldiagnostic information compared with the single tempo-ral 3D scan of Fig. 2. In particular, note that the con-tralateral MCA is not visible from the temporal scan ofFig 2A through C. This problem would be eliminated bythe dual simultaneous temporal 3D scans, since each

MCA would be scanned from the ipsilateral side in oneof the two scans.

There are, of course, a number of hurdles which wemust address and overcome before the ultrasound brainhelmet can become a useful clinical device.

(1) We must test whether the multiple matrix arrayprobes within a helmet can be positioned over theacoustic windows of the skull by medical personnel

Fig. 4. Split screen video display of image planes from simultaneous temporal and suboccipital 3D scans using twoasynchronous scanners including: (A) Axial scan from temporal window showing ipsilateral (MCA), anterior cerebralartery (ACA), posterior cerebral artery (PCA) and its P2 segment; (B) simultaneous coronal scan from temporalwindow; (C) simultaneous coronal scan from suboccipital window including vertebral artery (VA) and Atlas loop (AtL);

and (D) simultaneous parasagittal plane from suboccipital window with foramen magnum (FM) visible.



of limited training such as in an ambulance environ-ment. However, we note that in a real-time 3D scan,precise positioning and aiming of the matrix arraymay not be as important as in conventional 1D and2D transcranial Doppler, since an entire volume ofthe brain will be scanned and stored. Thus, the dis-play planes may be selected interactively at any

desired orientation either in real time or in playbackmode by the members of a remote stroke team.

(2) The image quality of the 3D ultrasound brain scansmust be significantly improved. The ultimate goal isimage quality sufficient to reliably differentiate normalcerebral vasculature from that of an ischemic or hem-orrhagic stroke. At the current time, it is unknown ifthis ultrasound image quality is attainable. However,we are encouraged by the recent progress in phaseaberration correction of the skull bone both from ourown group (Ivancevich 2008) and other laboratories(Fink 1992; Hynynen and Sun 1999). In particular, wenote that for the case of the transducers positioned overthe temporal windows, each transducer can act as aphase correction beacon for the opposing transducer.The use of an external transducer as a phase correctionbeacon was pursued in our laboratories by Miller-Jones(1980) and more recently by the team of Vignon et al.(2006). Also encouraging is the recent development ofimproved transcranial phase inversion harmonic imag-ing to reduce the blooming artifact of contrast agents(Hölscher et al. 2005).

(3) The real-time 3D scanner must be substantially minia-turized for use in a helmet apparatus. Here we are

Fig. 5. Schematics of matrix arrays including transmit elements (Tx), receive elements (Rx), associated beam plot, andreal-time C-scan for: (A) original VMI design, (B) increased SNR and (C) matrix array from prototype brain helmet.

Fig. 6. Prototype brain helmet with dual matrix arrays, modifiedtranscranial fixation system and skull phantom.



encouraged by the ongoing drive to miniaturize allultrasound systems, which has recently resulted in theSiemens P-10, a 2D phased array scanner the size of apersonal digital assistant (PDA), and in the GE Volu-son i, a real-time mechanical 3D fetal scanner the sizeof a laptop computer. These trends should continue andwill likely yield the necessary miniaturization for thebrain helmet.

Another challenge is to miniaturize the thicknessof the matrix arrays with a low profile to fit within a

helmet configuration. Our past experience with arraysfor 3D catheter and endoscope applications gives usconfidence in meeting this challenge, as we have fab-ricated such arrays on multilayer flexible circuits ofpolyimide measuring as small as two millimeters intotal thickness (Lee et al. 2004).

(4) The remote transmission of interactively controlled 3Dultrasound scans from a battery powered portable unitmust be perfected. This would allow transmission ofimages via cell phone or internet using webcam tech-

Fig. 7. Image planes from brain helmet 3D scan of skull phantom with blood vessel including: (A) composite coronalplane of left and right matrix arrays, (B) composite parasagittal plane of left and right matrix arrays, (C) axial plane from

left array and (D) axial plane from right array.



nology from a remote hospital or from an ambulance toa neurologic team at a stroke center. The stroke teamwould select the imaging planes within a volume ofdata, thus enabling interactive 3D imaging, remote di-agnosis and earlier therapeutic intervention. We notethat with our own 3D scanner we easily transmitted thestreaming video output of the display over the internetusing a video capture card (ViewCast Osprey 100,Plano, TX, USA) with no evident loss of image quality.Yet to be completed is the remote control selection ofthe image display planes from the real-time 3D scanvolume analogous to the panning and tilting features ofwebcam technology. Finally, we note the current com-mercial availability of ambulance-based ECG tele-medicine systems giving us additional confidence inthe feasibility of this approach.

(5) The display features of the scanner need to be mod-ified for convenient analysis by the stroke team. Forexample, the simultaneous axial and coronal slicesfrom both transducers are probably the most usefuland they should be displayed in real time in standardorientations rather than the current display of Fig. 6Athrough C. Additionally, the 3D rendering colorDoppler scan needs to be displayed in real time.Finally, the 3D scans from multiple transducersshould be fused into a single 3D image for easyexamination by the neurosonologist. This is an areaof active research in the image processing commu-nity undergoing rapid advances (Stathaki 2008). Asshown by the double arrow in Fig. 2C and D, anexample of such 3D ultrasound image fusion can beenvisioned by mentally connecting the posterior ce-rebral artery of Fig. 2C obtained from the temporalskull window with the basilar artery of Fig. 2Dobtained from the suboccipital window.

SUMMARY

We have described early stage in vivo and in vitroexperiments to test the feasibility of an ultrasound brainhelmet to produce multiple simultaneous real-time 3Dscans of the cerebral vasculature from the temporal andsuboccipital acoustic windows of the skull. Our long-term goal is to produce real-time 3D ultrasound imagesof virtually the entire cerebral vasculature. This scannerwould transmit its images from a remote hospital or froman ambulance via cellular networks or internet to a neu-rologic team at a stroke center. The team would becapable of remotely selecting the imaging plane to dis-play, thus enabling interactive 3D imaging.

Acknowledgments—This study was supported by grants RR024128 andHL072840 from the National Institutes of Health.

REFERENCES

Alexandrov AV, Molina CA, Grotta JC, Garami Z, Ford SR, Alvarez-Sabin J, Montaner J, Saqqur M, Demchuk AM, Moye LA, Hill MD,Wojner AW, Al-Senani F, Burgin S, Calleja S, Campbell M, ChenCI, Chernyshev O, Choi J, El-Mitwalli A, Felberg R, Ford S,Garami Z, Irr W, Grotta J, Hall C, Iguchi Y, Ireland J, Labiche L,Malkoff M, Morgenstern L, Noser E, Okon N, Piriyawat P, Rob-inson D, Shaltoni H, Shaw S, Uchino K, Yatsu F, Alvarez-Sabin J,Arenillas JF, Huertas R, Molina C, Montaner J, Ribo M, RubieraM, Santamarina E, Saqqur M, Alchtar N, O[0019]Rourke F, Hus-sain S, Shuaib A, Abdalla E, Demchuk A, Fischer K, Hill MD,Kennedy J, Roy J, Ryckborst KJ, Schebel M, Investigators C.Ultrasound-enhanced systemic thrombolysis for acute ischemicstroke. New Engl J Med 2004;351:2170–2178.

Babikian VL, Wechsler LR, eds. Transcranial Doppler ultrasonogra-phy. Boston: Butterworth-Heinemann, 1999.

Bauer A, Bogdahn U, Haase A, Schlief R. [3-dimensional echo-en-hanced transcranial Doppler ultrasound diagnosis]. Radiologe1998;38:394–398.

Bogdahn U, Becker G, Schlachetzki F, eds. Echoenhancers and trans-cranial color duplex sonography. Berlin; Boston: Blackwell Sci-ence, 1998.

Fink M. Time-reversal of ultrasonic fields. 1. Basic principles. IEEETrans Ultrason Ferroelectr Freq Control 1992;39:555–566.

Flax SW, O’Donnell M. Phase-aberration correction using signals frompoint reflectors and diffuse scatterers - Basic principles. IEEE TransUltrason Ferroelectr Freq Control 1988;35:758–767.

Fronheiser MP, Light ED, Idriss SF, Wolf PD, Smith SW. Real-time,3-D ultrasound with multiple transducer arrays. IEEE Trans Ultra-son Ferroelectr Freq Control 2006;53:100–105.

Gauss R, Trahey GE, Soo MS. Wavefront estimation in the humanbreast. Proc SPIE 2001;172-181.

Gorelick PB, Alter M. The prevention of stroke. Boca Raton: ParthenonPublishing Group, 2002.

Hanley D, Hacke W. Critical care and emergency medicine neurologyin stroke. Stroke 2005;36:205–207.

Hennerici MG, ed. Imaging in stroke. London: Remedica, 2003.Hölscher T, Wilkening WG, Lyden PD, Mattrey RF. Transcranial

ultrasound angiography (tUSA): A new approach for contrast spe-cific imaging of intracranial arteries. Ultrasound Med Biol 2005;31:1001–1006.

Hynynen K, Sun J. Trans-skull ultrasound therapy: The feasibility ofusing image-derived skull thickness information to correct thephase distortion. IEEE Trans Ultrason Ferroelectr Freq Control1999;46:752–755.

Ivancevich NM, Dahl JJ, Trahey GE, Smith SW. Phase Aberrationcorrection with a real-time 3D ultrasound scanner: Feasibilitystudy. IEEE Trans Ultrason Ferroelectr Freq Contr 2006;53:1432–1439.

Ivancevich NM, Pinton GF, Nicoletto HA, Bennett E, Laskowitz DT,Smith SW. Real-time 3D contrast-enhanced transcranial ultrasoundand aberration correction. Ultrasound Med Biol 2008 (in press).

Jensen JA, Svendsen NB. Calculation of pressure fields from arbitrarilyshaped, apodized, and excited ultrasound transducers. IEEE TransUltrason Ferroelectr Freq Control 1992;39:262–267.

Klötzsch C, Bozzato A, Lammers G, Mull M, Lennartz B, Noth J.Three-dimensional transcranial color-coded sonography of cerebralaneurysms. Stroke 1999;30:2285–2290.

Lee W, Idriss SF, Wolf PD, Smith SW. A miniaturized catheter 2-Darray for real-time, 3-D intracardiac echo cardiography. IEEE TransUltrason Ferroelectr Freq Control 2004;51:1334–1346.

Light ED, Davidsen RE, Fiering JO, Hruschka TA, Smith SW. Progressin two-dimensional arrays for real-time volumetric imaging. Ultra-sonic Imaging 1998;20:1–15.

Liu DL, Waag RC. Time-shift compensation of ultrasonic pulse focusdegradation using least-mean-square error-estimates of arrival time.J Acoust Soc Am 1994;95:542–555.

Marler JR, Brott T, Broderick J, Kothari R, Odonoghue M, Barsan W,Tomsick T, Spilker J, Miller R, Sauerbeck L, Jarrell J, Kelly J,Perkins T, Mcdonald T, Rorick M, Hickey C, Armitage J, Perry C,



Thalinger K, Rhude R, Schill J, Becker PS, Heath RS, Adams D,Reed R, Klei M, Hughes S, Anthony J, Baudendistel D, Zadicoff C,Rymer M, Bettinger I, Laubinger P, Schmerler M, Meirose G,Lyden P, Rapp K, Babcock T, Daum P, Persona D, Brody M,Jackson C, Lewis S, Liss J, Mahdavi Z, Rothrock J, Tom T,Zweifler R, Dunford J, Zivin J, Kobayashi R, Kunin J, Licht J,Rowen R, Stein D, Grisolia J, Martin F, Chaplin E, Kaplitz N,Nelson J, Neuren A, Silver D, Chippendale T, Diamond E, LobatzM, Murphy D, Rosenberg D, Ruel T, Sadoff M, Schim J, SchleimerJ, Atkinson R, Wentworth D, Cummings R, Frink R, Heublein P,Grotta JC, Degraba T, Fisher M, Ramirez A, Hanson S,Morgenstern L, Sills C, Pasteur W, Yatsu F, Andrews K,Villarcordova C, Pepe P, Bratina P, Greenberg L, Rozek S,Simmons K, Kwiatkowski TG, Horowitz SH, Libman R, Kanner R,Silverman R, Lamantia J, Mealie C, Duarte R, Donnarumma R,Okola M, Cullin V, Mitchell E, Levine SR, Lewandowski CA,Tokarski G, Ramadan NM, Mitsias P, Gorman M, Zarowitz B,Kokkinos J, Dayno J, Verro P, Gymnopoulos C, Dafer R,Dolhaberriague L, Sawaya K, Daley S, Mitchell M, Frankel M,Mackay B, Barch C, Braimah J, Faherty B, Macdonald J, Sailor S,Cook A, Karp H, Nguyen B, Washington J, Weissman J, WilliamsM, Williamson T, Kozinn M, Hellwick L, Haley EC, Bleck TP,Cail WS, Lindbeck GH, Granner MA, Wolf SS, Gwynn MW,Mettetal RW, Chang CWJ, Solenski NJ, Brock DG, Ford GF,Kongable GL, Parks KN, Wilkinson SS, Davis MK, Sheppard GL,Zontine DW, Gustin KH, Crowe NM, Massey SL, Meyer M,Gaines K, Payne A, Bales C, Malcolm J, Barlow R, Wilson M,Cape C, Bertorini T, Misulis K, Paulsen W, Shepard D, Tilley BC,Welch KMA, Fagan SC, Lu M, Patel S, Masha E, Verter J, BouraJ, Main J, Gordon L, Maddy N, Chociemski T, Windham J, Zadeh

HS, Alves W, Keller MF, Wenzel JR, Raman N, Cantwell L,Warren A, Smith K, Bailey E, Welch KMA, Tilley BC, FroehlichJ, Breed J, Easton JD, Hallenbeck JF, Lan G, Marsh JD, WalkerMD. Tissue-plasminogen activator for acute ischemic stroke. NewEngl J Med 1995;333:1581–1587.

Miller-Jones SM. Automated arrival time correction for ultrasoniccephalic imaging. Durham, NC: Duke University, 1980.

Pua EC, Idriss SF, Wolf PD, Smith SW. Real-time three-dimensionaltransesophageal echocardiography for guiding interventional electro-physiology: Feasibility study. Ultrasonic Imaging 2007;29:182–194.

Smith SW, Chu K, Idriss SF, Ivancevich NM, Light ED, Wolf PD.Feasibility study: Real-time 3D ultrasound imaging of the brain.Ultrasound Med Biol 2004;30:1365–1371.

Smith SW, Pavy HE, von Ramm OT. High speed ultrasound volumetricimaging system part I: Transducer design and beam steering. IEEETrans Ultrason Ferroelectr Freq Control 1991;38:100–108.

Stathaki T. Image fusion: Algorithms and applications. New York:Academic Press, 2008.

Tegeler CH, Babikian VL, Gomez CR. Neurosonology. St. Louis:Mosby, 1996.

Vignon F, Aubry JF, Tanter M, Margoum A, Fink M. Adaptivefocusing for transcranial ultrasound imaging using dual arrays. JAcoust Soc Am 2006;120:2737–2745.

von Ramm OT, Smith SW, Pavy HE. High speed ultrasound volumetricimaging system part II: Parallel processing and display. IEEE TransUltrason Ferroelectr Freq Control 1991;38:109–115.

Wessels T, Bozzato A, Mull M, Klötzsch C, Intracranial collateralpathways assessed by contrast-enhanced three-dimensional trans-cranial color-coded sonography. Ultrasound Med Biol 2004;30:1435–1440.


author's personal copy - duke universitytransducers.bme.duke.edu/pubs/umb_2009_329-338.pdfple...

Documents