Dual-Mode Intracranial Cath e ter In te grat ing 3D Ul tra sound Im ag ing and Hyperthermia for Neuro-on col ogy: Fea si bil ity Study

Transcription

1 ULTRASONIC IMAGING 31, (2009) Dual-Mode Intracranial Cath e ter In te grat ing 3D Ul tra sound Im ag ing and Hyperthermia for Neuro-on col ogy: Fea si bil ity Study CARL D. HERICKHOFF, 1 EDWARD D. LIGHT, 1 KRISTIN F. BING, 1 SRINIVASAN MUKUNDAN, 2 GERALD A. GRANT, 3 PATRICK D. WOLF 1 AND STEPHEN W. SMITH 1 1 Department of Biomedical Engineering Duke University Dur ham, NC carl.herickhoff@duke.edu 2 Department of Radiology Brigham and Women s Hos pi tal Boston, MA Division of Neurosurgery Duke University Medical Center Dur ham, NC In this study, we in ves ti gated the fea si bil ity of an intracranial cath e ter trans ducer with dual-mode ca - pa bil ity of real-time 3D (RT3D) im ag ing and ul tra sound hyperthermia, for ap pli ca tion in the vi su al iza - tion and treat ment of tu mors in the brain. Fea si bil ity is dem on strated in two ways: first by us ing a 50-el e ment lin ear ar ray trans ducer (17 mm 3.1 mm ap er ture) op er at ing at 4.4 MHz with our Volu met - rics diagnostic scanner and custom, electrical impedance-matching circuits to achieve a temperature rise over 4 C in ex cised pork mus cle, and sec ond, by de sign ing and con struct ing a 12 Fr, in te grated ma trix and linear-array catheter transducer prototype for combined RT3D imaging and heating capability. This dual-mode cath e ter in cor po rated 153 ma trix ar ray el e ments and 11 lin ear ar ray el e ments diced on a 0.2 mm pitch, with a to tal ap er ture size of 8.4 mm 2.3 mm. This 3.64 MHz ar ray achieved a 3.5 C in vi tro tem per a ture rise at a 2 cm fo cal dis tance in tis sue-mim ick ing ma te rial. The dual-mode cath e ter pro to - type was com pared with a Siemens 10 Fr AcuNav TM cath e ter as a gold stan dard in ex per i ments as sess ing im age qual ity and ther a peu tic po ten tial and both probes were used in an in vivo ca nine brain model to im - age an a tom i cal struc tures and color Dopp ler blood flow and to at tempt in vivo heating. Key words: Catheter transducer; dual-mode array; real-time 3D imaging; ultrasound hyperthermia. IN TRO DUC TION The Amer i can Can cer So ci ety es ti mates that in the United States, 21,810 new cases of pri - mary ma lig nant brain and cen tral ner vous sys tem (CNS) tu mors were di ag nosed in 2008 and 13,070 peo ple died from this con di tion, com pris ing just over 2% of all can cer deaths. 1 The Cen tral Brain Tu mor Reg is try of the United States es ti mates that in the year 2000, more than 81,000 per sons in the U.S. were liv ing with a pri mary ma lig nant brain/cns tu mor di ag no - sis. 2 Ma lig nant gliomas ac count for 81% of all ma lig nant pri mary brain/cns tu mors and 51% of gliomas are of the WHO grade IV sub type glioblastoma multiforme (GBM), the most com mon intracranial neo plasm. 2, /09 $18.00 Copy right 2009 by Dynamedia, Inc. All rights of re pro duc tion in any form re served.

2 82 HERICKHOFF ET AL Cur rently, pri mary brain tu mors are most com monly treated by sur gery, ra dio ther apy and che mo ther apy. 4 Che mo ther apy may be ad min is tered sys tem i cally, by mouth or in tra ve nous infusion, or locally, by direct intra-arterial or interstitial injection into a tumor target. 5 The ma jor chal lenge to ef fec tively de liv er ing drugs to CNS tu mors is to achieve a high drug con - cen tra tion within the tu mor bed by over com ing the phys i o log i cal blood-brain bar rier, the dy namic in ter face that pro tects the brain and also ac tively ex trudes drugs from the brain through ac tive drug efflux trans port ers. 6 Blood-brain bar rier (BBB) dis rup tion must be tran sient and re vers ible rapid clo sure of the bar rier af ter dis rup tion ef fec tively traps drug within the tis sue. 7, 8 High-in ten sity fo cused ul tra sound (HIFU) has been stud ied as a method for achiev ing this ef fect The use of low-pres sure ul tra sound (e.g., peak neg a tive pres sure am pli tudes of MPa) at low di - ag nos tic fre quen cies ( MHz), in con junc tion with microbubble con trast agent such as Optison or Definity, has also been shown to achieve lo cal ized BBB dis rup tion Mild hyperthermia (41 C) us ing ul tra sound (20 min, 0.4 W/cm 2 ) has been shown to re vers ibly en - hance pas sive dif fu sion of hy dro pho bic drugs through microvessel en do the lial cell mono - layers, al low ing the drugs to by pass efflux trans port ers and dem on strat ing the abil ity to permeate the BBB. 15 Ef forts to de velop transcranial, im age-guided fo cused ul tra sound sys tems for brain tu mor ab la tion have shown en cour ag ing re sults; how ever, these stud ies, as well as the afore men - tioned BBB dis rup tion ex per i ments, have con cerns as so ci ated with the skull ei ther a highly in va sive ap proach is taken, where the skull bone must be re moved prior to sonication, or a noninvasive ap proach is taken, where the high ul tra sound at ten u a tion and re flec tion from the bone is ac cepted and the phase ab er ra tion from the skull s vari able thick ness is ei - ther man aged or cor rected Min i mally-in va sive endovascular tech niques for treat ment of var i ous intracranial dis - eases, in clud ing arteriovenous mal for ma tion, intracranial an eu rysm and dural ve nous si nus throm bo sis, have been used in the field of interventional neuroradiology and these pro ce - 20, 21 dures can be ex tended to the treat ment of intracranial tu mors. For treat ment of ce re bral ve nous throm botic dis ease, a 7 Fr to 9 Fr in tro ducer sheath or guide cath e ter may be used for ac cess, usu ally via a transfemoral ap proach or a di rect ret ro grade in ter nal jug u lar stick, and a 5 Fr cath e ter may be in serted as far as the fron tal por tion of the su pe rior sagittal si nus Case re ports de tail ing the use of cath e ter de vices in the dural ve nous si nuses to treat throm bo sis (by urokinase in fu sion or rheolytic thrombectomy) dem on strate the fea si bil ity of this ap proach Dual-mode ul tra sound trans duc ers for im age-guided ther apy have been de signed and im - plemented for several thermal applications, including treatment of cardiac arrhythmias and tu mors of the liver, breast, biliary ducts and pros tate These have been de vel oped with a cath e ter, en do scope or extracorporeal ap proach, us ing ei ther a sin gle el e ment or phased ar - ray for ther apy (phased ar rays al low for the most con trol over the en ergy de po si tion pat tern of ul tra sound). 34 When com pared with mag netic res o nance (MR) im ag ing, ultrasound is an at trac tive mo dal ity for guid ance and treat ment mon i tor ing, be cause of its low cost, por ta bil - 32, 33 ity and high frame rate. Dual-mode im ag ing/ther apy ar rays are de sir able be cause of the in her ent spa tial reg is tra tion be tween im ag ing, treat ment and mon i tor ing. 35 For ex am ple, min ia tur ized lin ear ar rays, us ing a 2.3 mm 49 mm ap er ture at 3.1 MHz, have been de vel - oped for minimally-invasive interstitial thermal treatment and imaging of liver tissue. 28 In the 1990 s, real-time 3D (RT3D) ul tra sound im ag ing was de vel oped at Duke Uni ver - sity for transthoracic car diac ap pli ca tions and, more re cently, for intracardiac cath e ters, intravascular cath e ters, en do scopes, lap aro scopes, 3D ul tra sound guid ance of sur gi cal ro - bot ics, 3D transcranial ul tra sound and 3D intracranial ul tra sound en do scopes for neu ro sur - 30, gery and nee dle guid ance in the brain. Our Duke/Volu met rics Med i cal Im ag ing, Inc. (Dur ham, NC, USA) 3D ul tra sound sys tem scans a pyr a mid us ing a ma trix ar ray

3 DUAL-MODE INTRACRANIAL CATHETER FOR NEURO-ONCOLOGY 83 trans ducer to pro duce 3D scans at rates of 30 vol umes/s. Real-time dis play op tions in our 3-D scan ner in clude up to five im age planes ori ented at any de sired an gle and depth within the py ram i dal scan, as well as real-time 3D vol ume ren der ing, 3D steer able pulsed Dopp ler and 3D color Dopp ler flow im ag ing. The de vel op ment of microbubble ul tra sound con trast agents (UCAs) and tem per a - ture-sen si tive liposomes for tar geted im ag ing and tar geted drug de liv ery is also pro ceed ing rap idly. Microbubble UCAs are blood pool agents, mean size of m, administered in - tra ve nously such that a to tal in jec tion of only 0.1 ml in creases the sig nal of blood ves sels by over a fac tor of 10, en abling de tec tion of cap il lary-sized ves sels. The con trast microbubbles are used clin i cally in car diac ex ams, in ex per i men tal per fu sion stud ies of kid ney, trans cra - 46, 47 nial brain and liver and in blood flow mea sure ments of breast, pros tate and liver tu mors. UCAs of fer the ad van tage of por ta ble, in ex pen sive, real-time use with out ion iz ing ra di a tion and are thus ideal for se rial stud ies as in treat ment mon i tor ing. Lipo somes are mem - brane-en closed ves i cles com posed of a lipid bilayer shell (which can trap hy dro pho bic and amphipathic drugs) sur round ing an aque ous core (which can con tain hy dro philic drugs), with a di am e ter typ i cally near 100 nm. 48 Liposomes can be readily con ju gated to an ti bod ies or other ad he sion lig ands and UCAs can also be chem i cally mod i fied to en able ac tive tar get - ing of tu mor angiogenesis; for ex am ple, ul tra sound mo lec u lar im ag ing of tu mor angio - genesis in a rat af ter in tra ve nous in jec tion of microbubbles tar geted to v integrins has dem on strated en hance ment of a pri mary glioblastoma and a small me tas ta sis. 49 Hy brid drug de liv ery ve hi cles, con ju gat ing liposomes and microbubbles, are also under investigation. 48 Temperature-sensitive liposomes containing the cancer drug doxorubicin have been shown to quickly re lease their con tents when the sur round ing tem per a ture is raised to 41 C, or 4 C above normal body temperature. 50 Therefore, temperature-sensitive liposomes, in combination with lo cal hyperthermia (which may be cre ated and di rected by ul tra sound), pro vide tar geted 50, 51 con trol of drug re lease that may aug ment chemotherapeutic ef fi cacy in many clin i cal set tings. The goal of this pro ject is to ex tend real-time 3D ul tra sound to min i mally-in va sive cath e - ters for intracranial im ag ing of the brain and ul tra sound hyperthermia for neuro-on col ogy. This tech nol ogy could be used to vi su al ize a tu mor tar get in 3D and then trig ger the re lease of chemotherapeutic drugs con tained within microbubble or liposomal agents mo lec u larly tar - geted to re gions of tu mor angiogenesis. The cath e ter will be de signed to be both thin and flex i ble enough (ca pa ble of bend ing with a 1 cm ra dius of cur va ture) to be ma nip u lated through the in ter nal jug u lar vein into the dural ve nous si nuses (Fig. 1), which pro vide min i - mally-in va sive ac cess to vir tu ally the en tire brain vol ume. In this pa per, we de scribe a se ries of ex per i ments con ducted in an ef fort to dem on strate proof of con cept. An ini tial pro to type dual-mode (RT3D im ag ing & hyperthermia) cath e ter trans ducer was de signed and con structed to de ter mine the fea si bil ity of our long term goal. The im ag ing and ther mal per for mance of the dual-mode pro to type was com pared with a Siemens (Siemens Med i cal So lu tions, Ul tra sound Di vi sion, Issaquah, WA, USA) AcuNav TM catheter in in vi tro ex per i ments and in an in vivo ca nine brain model, with ac cess to the su pe - rior sagittal si nus af forded through a 1 cm burr hole in the skull. These ex per i ments dem on - strate that a com pletely in tra ve nous ap proach for such a treat ment may be suc cess ful, thus elim i nat ing the need for a burr hole. METH ODS The trans ducer ar ray we pro pose in te grates a 2D ma trix ar ray, which can ac quire 3D scans, with ad ja cent lin ear ar rays, which con trib ute power to a beam for in duc ing hyperthermia (Fig. 2).

4 84 HERICKHOFF ET AL FIG. 1 In tended cath e ter path way af ford ing min i mally-in va sive ac cess to the brain vol ume via dural ve nous si nuses. Preliminary experiment: hyperthermia with RT3D scan ner To es tab lish the fea si bil ity of a dual-mode im ag ing/ther apy cath e ter trans ducer, we first sought to dem on strate that our di ag nos tic RT3D Volu met rics scan ner could be used to achieve and main tain a suf fi cient tem per a ture rise ( = 4 C). In this pre lim i nary ex per i ment, a 50 chan nel, 4.4 MHz PZT-5H, 17 mm 3.1 mm lin ear ar ray trans ducer pre vi ously de scribed was used with elec tri cal im ped ance-match ing cir cuits cus tom ized for each chan nel in or der to max i mize power trans fer from the scan ner to the trans ducer el e ments. 52 The match ing cir - cuits were sim ple L-sec tion net works con sist ing of two re ac tive com po nents: an in duc tor in se ries with the load and a ca pac i tor in par al lel. To de sign these net works, the sys tem (source) im ped ance and the cor re spond ing wa ter-cou pled trans ducer el e ment (load) im ped - ance were mea sured at the cho sen op er at ing fre quency, for each chan nel. Then, be cause max i mum real power trans fer oc curs when the in put im ped ance (L-sec tion plus load) equals the com plex con ju gate of the source im ped ance, we found the op ti mal val ues of in duc tance and capacitance for the L-section components algebraically. The necessary series induc - tance ranged from 9 to 30 H and the par al lel ca pac i tance ranged from 47 to 268 pf. With the match ing cir cuits in use, the scan ner gen er ated a 12-cy cle pulse at 6.6 khz PRF set tings found to yield max i mum sta ble power out put from the scan ner and the spa - tial-peak, temporal-average intensity (I SPTA ) was mea sured at a 3 cm fo cal depth with a cal i - brated mem brane hydrophone (S4-251, Sonora Med i cal Sys tems, Longmont, CO, USA) in

5 DUAL-MODE INTRACRANIAL CATHETER FOR NEURO-ONCOLOGY 85 FIG. 2 In te grated 2D ma trix & lin ear ar ray de sign for RT3D im ag ing & hyperthermia. The cen tered 2D ar ray scans a py ram i dal vol ume in real time and the ad ja cent lin ear ar rays pro duce a hyperthermia beam (shown in red within the vol ume), steer able in az i muth (dif fer ent pos si ble foci in di cated by red x s). a wa ter tank, in ac cor dance with the pro ce dures out lined in the AIUM/NEMA UD stan dard. 53 A ther mo cou ple (5TC se ries, Omega En gi neer ing, Inc., Stam ford, CT, USA) was used to mon i tor the tem per a ture of the scan ner s in ter nal elec tron ics dur ing trans mit. With the same scan ner set tings, the hydrophone was re placed by a piece of de gassed, ex cised pork mus cle; a type T, 33 gauge, stain less steel hy po der mic nee dle ther mo cou ple (HYP-0, Omega En gi neer ing, Inc.) was in serted ap prox i mately 2 mm be neath the sur face of the tis - sue and placed at the fo cus to mea sure the achiev able tem per a ture rise. The data ac qui si tion sys tem (Integra model 2700, Keithley In stru ments, Inc., Cleve land, OH, USA), which con - tains built-in sig nal con di tion ing hard ware for fil ter ing and noise re duc tion, re corded tem - per a ture data at a sam pling fre quency of 3.8 Hz. The tis sue tem per a ture rise mea sure ment was rep li cated four times. The nonat ten u at ing wa ter path be tween the trans ducer and the tis - sue served to iso late the tis sue from con duc tive heat ing so as to better cor re late tem per a ture measurements with intensity measurements. Possible artifacts associated with this type of ther mal mea sure ment are vis cous heat ing, re flec tions and con duc tion along the wire; our small-di am e ter, fine-wire ther mo cou ple sheathed in stain less steel was cho sen to min i mize 54, 55 these ef fects and we dis re gard them for the proof-of-con cept ex per i ments de scribed here. Integrated catheter transducer design and fabrication The re sults from our pre lim i nary ex per i ment gave us con fi dence to de velop a new de sign for a pro to type cath e ter de vice in te grat ing a ma trix ar ray and a lin ear ar ray. We used a cus - tom 10 Fr multi-layer flex i ble cir cuit made by Microconnex (Snoqualmie, WA, USA) as a con struc tion tem plate. The flex cir cuit could ac com mo date an ar ray with 198 ma trix el e - ments and 18 lin ear el e ments, diced on a 0.2 mm pitch, for a to tal ap er ture size of 8.4 mm

6 86 HERICKHOFF ET AL FIG. 3 Field II sim u la tion ap er tures and beam plots. (a, b, c) Ac tive ap er tures for each case: the en tire in te grated ar ray with 2 cm fo cus, the in te grated cath e ter s lin ear ar rays with unfocused trans mit and the AcuNav with 2 cm fo - cus, re spec tively. (d, e, f) Rel a tive pres sure am pli tude in the zero-el e va tion plane for each case, re spec tively. (g, h, i) Rel a tive pres sure am pli tude at a depth of 2 cm for each case, re spec tively. 2.3 mm. Two dif fer ent schemes were con sid ered as meth ods to achieve op ti mal tis sue heat - ing: one with all the el e ments (both the ma trix and lin ear ar rays) fo cused at 2 cm driven by the Volu met rics scan ner and one unfocused scheme us ing only the lin ear ar ray el e ments driven by a sin gle-chan nel, 25-Watt rf power am pli fier (Model 525LA, ENI, Roch es ter, NY, USA) con nected to a wave form gen er a tor (Model 33250, Agilent, Santa Clara, CA, USA). The unfocused lin ear ar ray with sin gle-chan nel power am pli fier was con sid ered as an al ter - na tive to us ing the Volu met rics scan ner at high power set tings that could po ten tially harm its in ter nal elec tron ics. Both schemes used a 4.4 MHz op er at ing fre quency. To aid our de sign pro cess, the beam pat tern for each of these pro to type schemes was sim u - lated and com pared with that of a Siemens 10 Fr AcuNav cath e ter probe us ing Field II. 56 The AcuNav is a lin ear ar ray with 64 el e ments on a 0.11 mm pitch, for an ap er ture size of ap prox - i mately 7.1 mm 2.5 mm; this ar ray was fo cused at 2 cm, op er at ing at 5.33 MHz. Each com - putation included a 0.5 db/cm/mhz attenuation factor to simulate brain tissue. 57 The sim u la tions did not cal cu late the first 5 mm in front of the trans ducer be cause the model is less ac cu rate for very short dis tances due to use of a far-field ap prox i ma tion of the re sponse of each trans ducer el e ment. Fig ure 3 il lus trates this stage of the de sign pro cess. Fig ures 3a, 3b and 3c show the foot - print of the ac tive ap er ture in each case: the en tire in te grated ar ray (2 cm trans mit fo cus), the in te grated de sign s lin ear ar ray el e ments only (unfocused trans mit) and the AcuNav (2 cm trans mit fo cus), re spec tively. Fig ures 3d, 3e and 3f dis play the beam s rel a tive pres sure am - pli tude in the zero-el e va tion plane for each case, re spec tively, and fig ures 3g, 3h and 3i show each beam s rel a tive am pli tude at a depth of 2 cm. For the case of the en tire in te grated ar ray, the 6 db az i muthal beam width was 7.5 mm at 0.5 cm deep. For the case of the in te grated de sign s unfocused lin ear ar ray el e ments, the 6 db beam beam width was 7.8 mm at 0.5 cm deep. For the fo cused AcuNav, the 6 db beam width was 4.6 mm at 0.5 cm deep.

7 DUAL-MODE INTRACRANIAL CATHETER FOR NEURO-ONCOLOGY 87 FIG. 4 (a) Bare flex cir cuit above a ruler with mm mark ings. (b) Diced ma trix and lin ear ar rays with MicroFlat ca bles sol dered to lin ear ar ray con tacts. (c) Com pleted 12 Fr dual-mode catheter transducer. Fab ri cat ing the in te grated ar ray in volved adapt ing the flex cir cuit (Fig. 4a), which was ini - tially de signed for a ma trix ar ray alone. 42 A 0.35-mm-thick piece of PZT-5H was bonded to the flex cir cuit with sil ver ep oxy, slightly over lap ping the grid of el e ment con tacts onto the bor der ing ground con tact. This piece was diced on a 0.2 mm pitch both ver ti cally and hor i - zon tally to cre ate the ma trix ar ray on a el e ment grid. The lin ear ar ray was then built in the ad ja cent spaces: first, each space was filled with sil ver ep oxy and cured to cre ate a level back side con tact; then, a piece of PZT-5H was bonded to each side and diced (0.2 mm pitch), phys i cally sep a rat ing the nine el e ments on each end. MicroFlat ca bles from W.L. Gore & As so ci ates, Inc. (New ark, DE, USA) were sol dered with ap pro pri ate spac ing to the ex posed bor der con tact and these sol der joints were re in forced with UV-cur ing ep oxy be fore care - fully re-dic ing the lin ear ar ray to en sure elec tri cal sep a ra tion of the lin ear ar ray chan nels (fig ure 4b shows the com ple tion of these steps: the 8.4 mm 2.3 mm to tal ap er ture). A layer of dou ble-sided metallized liq uid crys tal poly mer (LCP) was at tached to the face of the trans ducer for a front ground ing con tact and over all elec tronic shield ing for the trans - ducer. There was no acous tic match ing layer for this pro to type. MicroFlat ca bles were sol - dered to the pads of the flex cir cuit to con nect the ma trix ar ray chan nels and an acous tic back ing was at tached be fore bend ing the flex cir cuit into a side-view ing con fig u ra tion and

8 88 HERICKHOFF ET AL pack ag ing the trans ducer into the 12 Fr cath e ter lu men. The prox i mal ends of the ca bles were sol dered to cus tom cir cuit boards fit ted with Samtek, Inc. (New Al bany, IN, USA) con - nec tors (model # BSH F-D-A) to link with the ul tra sound sys tem han dle. The com - plete, pack aged trans ducer is shown in fig ure 4c (on a larger scale than fig ures 4a,b). In vi tro temperature measurement For mea sure ments as sess ing the hyperthermia ca pa bil ity of the pro to type dual-mode cath - e ter trans ducer, tis sue-mim ick ing ma te rial from Na tional Phys i cal Lab o ra tory (Teddington, 58, 59 UK) was used as a model, hav ing ther mal prop er ties sim i lar to brain tis sue. Be cause the dual-mode pro to type had a smaller ap er ture than the 50-chan nel lin ear ar ray used in our pre - lim i nary ex per i ment, a 2 cm fo cus was cho sen in stead of 3 cm to in crease fo cal gain. A hy - po der mic nee dle ther mo cou ple was in serted ap prox i mately 2 mm be neath the sur face and placed at the fo cus to mea sure the achiev able tem per a ture rise. As be fore, the wa ter path be - tween the trans ducer and the tar get served to iso late the tis sue from con duc tive heat ing. In or der to com pare hyperthermia ex per i ments and as sess whether the heat ing was a re sult of con duc tion or ul tra sound ab sorp tion, the max i mum heat ing rate was cal cu lated by di vid ing by the amount of time taken to achieve the ini tial 2.0 C of the to tal tem per a ture rise. Large an i mal study pro to col The In sti tu tional An i mal Care and Use Com mit tee at Duke Uni ver sity ap proved the fol low ing pro ce dure used on two ca nine sub jects one for a RT3D im ag ing study us ing an earlier ma - trix ar ray cath e ter pro to type, and one for a com bined RT3D im ag ing and hyper thermia study using the integrated matrix and linear array prototype. 60 In each study, an in tra ve nous (IV) line was es tab lished in a pe riph eral vein and the dog was se dated with thiopental so dium, 20 mg/kg ad min is tered in tra ve nously. An es the sia was in duced with in ha la tion of isoflurane gas 1% to 5 % de liv ered through a nose cone. The an i mal was intubated, placed on a wa ter heated ther - mal pad and started on a ven ti la tor. A fem o ral ar te rial line was placed via a percutaneous punc ture or cutdown. Elec tro lyte and ven ti la tor ad just ments were made based on se rial elec - tro lyte and ar te rial blood gas mea sure ments. An IV main te nance drip of D-5 lac tated Ringer s so lu tion was started and main tained at 5 ml/kg/min. Blood pres sure, lead II elec - tro car dio gram and tem per a ture were con tin u ously mon i tored through out the pro ce dure. The an i mal was placed in a prone po si tion, the top of the head was shorn and a scalp in ci sion was made near the skull midline. The skin and mus cle were dis sected free and re tracted to ex pose the skull. A 1 cm burr hole was care fully drilled through the bone just above the ex - ter nal oc cip i tal pro tu ber ance to ex pose the dura ma ter en cas ing the su pe rior sagittal ve nous si nus. An in ci sion in the dura was made and a cath e ter de vice was in serted into the su pe rior sagittal si nus; the cath e ter was ad vanced and ro tated in the ves sel to ori ent and op ti mize the field of view and iden tify struc tures in the cra nial cav ity us ing both B-mode and Dopp ler im - ag ing. For the cath e ter hyperthermia por tion of the study, a sec ond 1 cm burr hole was made ap prox i mately 1.5 cm lat eral to the midline and the po si tion of the in serted trans ducer. A type T, 33 gauge, stain less steel hy po der mic nee dle ther mo cou ple was in serted through the ex posed dura into the ce re brum to mon i tor tem per a ture dur ing hyper thermia tri als. RESULTS Preliminary experiment: hyperthermia with RT3D scan ner With the 50-chan nel lin ear ar ray trans mit ting to a 3 cm fo cal depth in a wa ter tank, a spa - tial-peak, temporal-average intensity (I SPTA ) of 6.88 W/cm 2 was mea sured with our mem -

9 DUAL-MODE INTRACRANIAL CATHETER FOR NEURO-ONCOLOGY 89 FIG. 5 (a) Cen ter of lin ear ar ray trans ducer used in in vi tro hyperthermia experiment with Volumetrics RT3D scan ner and match ing cir cuits. (b) Ther mo cou ple data show ing a 5.0ºC tem per a ture rise at 3 cm fo cus in de gassed, ex cised pork mus cle. brane hydrophone. Us ing iden ti cal scan ner set tings, we mea sured a tem per a ture rise of 5.0 C in de gassed, ex cised pork mus cle. The in ter nal tem per a ture of the scan ner re mained at an ac cept able level as the power out put was main tained for sev eral min utes. Fig ure 5a is a photo of the cen ter of the lin ear ar ray used in this ex per i ment; note that a gap was left where the ma trix ar ray would be ac cord ing to the in te grated de sign given in fig ure 2. The data in fig ure 5b show the tem per a ture rise over 4 C within one min ute af ter turn-on, lev el ing off at about 5 C above the am bi ent tem per a ture. In vi tro imaging and thermal performance tests The dual-mode RT3D/hyperthermia pro to type cath e ter con struc tion yielded 153 ma trix ar ray el e ments (77% el e ment yield) and 11 lin ear ar ray el e ments (61% yield). El e ment yield

10 90 HERICKHOFF ET AL FIG. 6 (a) AcuNav im age of 4-cm deep, 1.5-cm di am e ter cyst phan tom. (b) Dual-mode cath e ter sin gle-b scan of cyst phan tom. (c, d, e) Dual-mode cath e ter 3D scan of cyst phan tom: co ro nal, ax ial and sagittal im age planes. will steadily im prove with prac tice and re fine ment of fab ri ca tion tech niques. The dualmode cath e ter and Volu met rics scan ner were com pared with a Siemens 10 Fr AcuNav and SONOLINE Anteres TM sys tem as a gold stan dard to as sess im ag ing ca pa bil ity and ther a peu - tic po ten tial. Var i ous wire, tu mor, and cyst phan toms were im aged with each probe. Fig ure 6 shows im ages of a test phan tom con tain ing a 1.5 cm di am e ter cyst, 4 cm deep: fig ure 6a is an im age taken with the AcuNav at 5.33 MHz, fig ure 6b is sin gle-b scan us ing both the ma - trix and lin ear ar ray el e ments of the in te grated cath e ter ar ray at 3.64 MHz and fig ures 6c, 6d and 6e are si mul ta neously-ac quired co ro nal, ax ial and sagittal planes, re spec tively, of a 3D scan us ing only the ma trix ar ray el e ments. It should be noted that the con cen tric ring pat tern of the ax ial im age (Fig. 6d) is an ar ti fact of the sys tem s scan con verter. The speckle pat terns in these im ages show that the pro to type dual-mode cath e ter yielded poorer spa tial res o lu tion than the AcuNav due to the re duced ap - er ture and band width of the ma trix ar ray for 3D im ag ing. In ad di tion, the AcuNav showed su pe rior color Dopp ler sen si tiv ity (Fig. 9b). To as sess the po ten tial of the cath e ters to cre ate hyperthermia, the max i mum spa tial-peak, tem po ral-av er age in ten sity was mea sured in a wa ter tank with our mem brane hydrophone. At a trans mit fo cus of 2 cm, the Volu met rics scan ner de liv ered an I SPTA = 2.43 W/cm 2 us ing a 12-cy cle, 3.64 MHz pulse at a 6.6 khz PRF with the dual-mode cath e ter. An I SPTA = 4.54 W/cm 2 was also mea sured at 1 cm for the case of the unfocused lin ear ar ray el e ments driven at 4.6 MHz by a 25-Watt power am pli fier. Fi nally, the AcuNav could de liver a 10-cy cle, 5.33 MHz pulse at 6.6 khz PRF and 55% sys tem volt age as pro jected max i mum set tings safe for the sys tem. To avoid po ten tially-dam ag ing self-heat ing in the AcuNav, I SPTA mea sure - ments were taken for lower volt age set tings on the Siemens sys tem and we ex trap o lated to 55% yield ing an es ti mate of I SPTA = 4.10 W/cm 2. Us ing the same set tings as for the I SPTA = 2.43 W/cm 2 data, the dual-mode cath e ter pro to type cre ated a tem per a ture rise of 3.5 C ( Fig. 7) at a 2 cm trans mit fo cus in tis sue-mim ick ing ma te rial.

11 DUAL-MODE INTRACRANIAL CATHETER FOR NEURO-ONCOLOGY 91 FIG. 7 Thermocouple data showing 3.5ºC temperature rise in tissue-mimicking material using the dual-mode cath e ter fo cused at 2 cm. TA BLE 1 In ten sity and ther mal ex per i ment re sults for three ar rays and five con fig u ra tions. Frequency (MHz) Focus (cm) I SPTA (W/cm 2 ) Temperature Rise ( C) in vi tro in vivo Maximum heating rate ( C/s) Lin ear ar ray (w/match ing cir cuits n/a AcuNav * n/a n/a n/a AcuNav n/a Dual-mode catheter (all el e ments) Dual-mode catheter (linear arrays) n/a n/a unfocused 4.54 n/a * ex trap o lated es ti mate trans ducer self-heat ing, con duc tion weak ened el e ments This was achieved in less than 2.5 min utes, at which time trans mis sion was halted be cause some of the scan ner s in ter nal elec tronic com po nents were over heat ing (above 85 C). It took sev eral min utes for the tis sue-mim ick ing ma te rial to dis si pate the added heat and re turn to equi lib rium tem per a ture. To avoid risk ing dam age to the scan ner, fur ther ther apy ex per i - ments were car ried out with the dual-mode pro to type by driv ing only the lin ear ar ray el e - ments as a sin gle chan nel (unfocused) us ing an ex ter nal, 25-Watt rf power am pli fier con nected to a wave form gen er a tor. Ta ble 1 shows a com plete chart of ex per i men tal re sults.

12 92 HERICKHOFF ET AL FIG. 8 In vivo ca nine model: the dual-mode cath e ter (C) placed in the su pe rior sagittal si nus through a burr hole cre ated in the skull and a ther mo cou ple (T) in serted into the ce re brum through the dura ma ter ex posed by a sec ond burr hole, for hyperthermia tri als. In vivo ca nine model im ag ing Each ul tra sound cath e ter was placed in the su pe rior sagittal si nus ac cord ing to the large an i mal pro to col. Fig ure 8 is a photo from the sur gi cal pro ce dure show ing the ex posed skull, with the dual-mode cath e ter in serted an te ri orly into the su pe rior sagittal si nus through the burr hole cre ated on the midline and the ther mo cou ple in serted me di ally into the ce re brum through the dura ma ter ex posed by an ad ja cent burr hole. Both echo and Dopp ler im ages were ac quired with each probe. The AcuNav pro duced echo im ages with clear de lin ea tion of ce re bral gyri and sulci (Fig. 9a), as well as color Dopp ler im ages of the in ter nal ca rotid ar - tery (Fig. 9b) ac quired with out the use of con trast.

13 DUAL-MODE INTRACRANIAL CATHETER FOR NEURO-ONCOLOGY 93 FIG. 9 Intracranial AcuNav im ages. (a) Echo im age show ing var i ous gyri and sulci of the ce re brum. (b) Color Doppler image showing the internal carotid artery. The dual-mode cath e ter pro to type ac quired RT3D im age data from a py ram i dal vol ume ex tend ing from the trans ducer face to ward the base of the cra nial cav ity. The Volu met rics scan ner si mul ta neously dis played two per pen dic u lar B-mode sec tors (sagittal and co ro nal planes) and two C-scans (ax ial plane, par al lel to the trans ducer face), which can be in clined at any de sired an gle. The co ro nal plane im ages were found to be very use ful for ori ent ing the trans ducer s field of view, en abling the op er a tor to iden tify struc tures more quickly, with greater confidence and accuracy. The lateral ventricles were the most prominent anatomical land mark, and could be clearly seen in all three or thogo nal im age planes. Due to the lim ited spa tial res o lu tion of the dual-mode cath e ter, the nar row lu men of the lat eral ven tri cles and the highly-echogenic ven tric u lar walls in the ca nine, the ven tric u lar lu men does not ap pear

14 94 HERICKHOFF ET AL FIG. 10 (a, b, c) Intracranial RT3D echo im ages in co ro nal, ax ial and sagittal planes, re spec tively, com pared to corresponding anatomical images (reproduced with permission from Fletcher et al, University of Minnesota Col - lege of Vet er i nary Med i cine). The lat eral ven tri cles (LV) are clearly seen in (a) and (b). The tentorium (T) and cer e - bel lum (C), as well as a pos te rior horn of the lat eral ven tri cle, are vis i ble in (c). as an echo-free re gion. The tentorium cerebelli mem brane, which sep a rates the cer e bel lum from the ce re brum, and pul sa tions of ar ter ies at the base of the brain were also vis i ble. Fig - ures 10a, 10b, and 10c com pare si mul ta neous co ro nal, ax ial and sagittal planes, re spec tively, ver sus cor re spond ing an a tom i cal slices. 61 The lat eral ven tri cles can be clearly seen in the co - ro nal and ax ial planes and the sagittal plane just ad ja cent to the midline shows the cer e bel - lum, the tentorium and a pos te rior horn of the lat eral ven tri cles. For 3D color Dopp ler im ag ing, a bolus of ag i tated sa line was needed for con trast and was de liv ered into the left in ter nal ca rotid ar tery. 62 Several vessels became visible immediately af ter the bolus in jec tion, in clud ing those that com prise the Cir cle of Wil lis. The con trast-en - hanced 3D color Dopp ler scans are shown in fig ure 11. To elim i nate ef fects of aliasing due to undersampling, the color Dopp ler look-up ta ble was mod i fied to dis play only a sin gle hue, to better visualize the vessels (albeit eliminating directional flow information). Figure 11a is the co ro nal plane of the scan, clearly show ing the an te rior com mu ni cat ing ar tery and left mid dle ce re bral ar tery as well as the left and right in ter nal ca rotid ar ter ies. Fig ure 11b shows the Cir cle of Wil lis, in clud ing the an te rior, mid dle and pos te rior ce re bral, an te rior and pos te - rior com mu ni cat ing ar ter ies, in an ax ial im age plane. Fig ure 11c is a sagittal plane of the scan, show ing the left in ter nal ca rotid, mid dle and pos te rior ce re bral, and pos te rior com mu - ni cat ing ar ter ies. Fig ure 11d is an an a tom i cal di a gram show ing a la tex-in jected sheep brain of sim i lar scale for ref er ence. 63 In vivo ca nine model hyperthermia Af ter im ag ing, the sec ond burr hole was cre ated, and a hy po der mic nee dle ther mo cou ple was in serted. The ther mo cou ple was lo cated and aligned within 1 cm of the trans ducer by im ag ing and both were held in place. Fig ure 12a shows the ther mo cou ple aligned in an

15 DUAL-MODE INTRACRANIAL CATHETER FOR NEURO-ONCOLOGY 95 FIG. 11 (a, b, c) Intracranial RT3D color Dopp ler im ages in co ro nal, ax ial and sagittal planes, re spec tively (note: the color Doppler look-up table was modified, eliminating directional flow information). (d) Latex-injected sheep brain vasculature for an a tom i cal ref er ence (re pro duced with per mis sion from R.R. Miselis, Uni ver sity of Penn syl - vania School of Veterinary Medicine). The internal carotid arteries (ICA), left mid dle ce re bral ar tery (MCA) and an te rior com mu ni cat ing ar tery (ACoA) are vis i ble in (a). The Cir cle of Wil lis is shown in (b), with the an te rior ce re - bral arteries (ACeA), posterior cerebral arteries (PCeA) and posterior communicating arteries (PCoA) also indicated. AcuNav scan. Scan ning was halted and high-power pulse se quences (10-cy cles at 5.33 MHz, 6.6 khz PRF and 55% sys tem volt age) were ini ti ated. As seen in fig ure 12e, af ter about 20 sec onds, the tem per a ture be gan to rise slowly and lin early with time for the next 3.5 min utes up to 4.5 C above base line un til it was re ported that the AcuNav felt hot where it was be ing held six inches from the trans ducer. Af ter turn ing off power, the ther mo cou ple re corded a slow de cline in tem per a ture while the probe it self re mained warm. The cas ing around the trans ducer had melted and de formed and the probe could no lon ger ac quire im ages, in di cat ing that the mea sured heat ing was due to en ergy dis si pa tion within the trans ducer and con duc tion through the tis sue rather than ul - tra sound ab sorp tion at the fo cus. Com par i son of the slow heat ing rate (0.020 C/s) with those of our in vi tro ul tra sound hyperthermia ex per i ments ( C/s) re in forces this determination. For the dual-mode cath e ter, we had de cided that driv ing the lin ear ar ray el e ments as a sin - gle chan nel (unfocused) with an ex ter nal, 25-Watt power am pli fier was pref er a ble to us ing the Volu met rics scan ner at high power set tings that could po ten tially harm its in ter nal elec - tron ics. Here, the ther mo cou ple was aligned with one side of the lin ear ar ray at a depth of 0.5 cm, where we ex pected max i mal heat ing ac cord ing to our sim u la tions (Fig. 3e). Fig ures 12b, 12c and 12d show si mul ta neous co ro nal, ax ial and sagittal planes of the in serted ther - mo cou ple. 3D scan ning with the ma trix ar ray was paused as ther mo cou ple mon i tor ing was ac ti vated and the power am pli fier turned on to drive the lin ear ar ray at 4.6 MHz. The mea -

16 96 HERICKHOFF ET AL FIG. 12 (a) AcuNav im age of ther mo cou ple (ar row) placed at 1 cm depth in ce re brum. (b, c, d) RT3D co ro nal, ax ial and sagittal im ages, re spec tively, of in serted ther mo cou ple (ar rows), ac quired with dual-mode cath e ter. (e) Tem per a ture rise achieved us ing the AcuNav (trans ducer self-heat ing and con duc tion). (f) Tem per a ture rise achieved by driv ing the dual-mode cath e ter s lin ear ar ray el e ments as a sin gle chan nel (unfocused) with an rf power am pli fier. sured tem per a ture rise was just over 0.85 C (Fig. 12f); how ever, it was found later that the trans mit ef fi ciency of the lin ear ar ray el e ments had been con sid er ably weak ened dur ing the ex per i ment, in di cat ing that this mea sure ment did not give the max i mum pos si ble tem per a - ture rise achiev able by a dual-mode probe of this de sign. For this ex per i ment, be cause we did not ex ceed 2.0 C of heat ing, the max i mum heat ing rate (0.146 C/s) was cal cu lated by di - vid ing by the amount of time taken to achieve the ini tial C of the to tal tem per a ture rise. This da tum was in agree ment with our in vi tro heating-rate measurements. SUM MARY AND DISCUSSION We suc cess fully cre ated a tem per a ture rise of 5 C in ex cised pork mus cle us ing the Volu - metrics RT3D scanner and customized, channel-specific impedance matching circuits with a 50 chan nel lin ear ar ray, dem on strat ing that hyperthermia might be achiev able in brain tis - sue us ing cath e ter ar rays. An in te grated ma trix and lin ear ar ray cath e ter trans ducer pro to - type was de signed and fab ri cated. When com pared to a Siemens AcuNav probe, this dual-mode catheter device exhibited satisfactory RT3D image quality and created a 3.5 C tem per a ture rise in tis sue-mim ick ing ma te rial at a depth of 2 cm with out us ing elec tri cal im - pedance-match ing cir cuits. We be lieve that achiev ing a higher el e ment yield and add ing match - ing cir cuits at the prox i mal end of the cath e ter to max i mize power out put ef fi ciency would make a 4.0 C or greater tem per a ture rise pos si ble us ing the Volu met rics scan ner with this ar ray de sign. To our knowl edge, we have shown the first intracranial ul tra sound im ages in the brain us - ing cath e ter probes. In an in vivo ca nine model, the AcuNav and dual-mode pro to type each

17 DUAL-MODE INTRACRANIAL CATHETER FOR NEURO-ONCOLOGY 97 were placed in the su pe rior sagittal ve nous si nus to im age the brain vol ume and at tempt to cre ate hyperthermia. The higher sen si tiv ity and Dopp ler res o lu tion of the AcuNav was ap - par ent in 2D scans of ce re bral gyri and sulci and of an in ter nal ca rotid ar tery; these im ages demonstrate the immediate clinical imaging potential of intracranial catheters. The dualmode cath e ter pro to type in ter ro gated a 3D py ram i dal vol ume in real-time, which made it eas ier to rec og nize struc tures within the cra nial cav ity. The lat eral ven tri cles were eas ily dis - tin guish able in echo im ages and in jec tion of an ag i tated sa line bolus for ves sel con trast al - lowed visualization of the Circle of Willis and several connecting arteries. The AcuNav was able to cre ate a tem per a ture rise of 4.5 C in ca nine brain tis sue but it was ev i dent that this was pri mar ily, if not ex clu sively, due to trans ducer self-heat ing and the re - sult ing heat con duc tion through out the tis sue vol ume, not ab sorp tion of en ergy from ul tra - sound waves. This il lus trates the im por tance of con sid er ing the con duc tion of in ter nal heat away from the trans ducer el e ments when de sign ing a dual-mode de vice: we sus pect that the poly mer cas ing en cap su lat ing the AcuNav may have been ef fec tively trap ping ther mal en - ergy and ex ac er bat ing the self-heat ing of the probe. The dual-mode pro to type was able to cre ate an in vivo tem per a ture rise less than one de - gree Cel sius by driv ing the lin ear ar ray el e ments as a sin gle chan nel at 4.6 MHz us ing a 25-Watt power amplifier. Post-experiment assessment of the transducer s performance re - vealed that the acous tic power out put had been de graded by a fac tor of 30 dur ing use in the ca nine brain model; there fore, the in vivo re sult is not con clu sive. The max i mum heat ing rate for the dual-mode pro to type in vivo was com pa ra ble to the in vi tro measurement, which in di cates that the heat ing was due to ul tra sound ab sorp tion and sug gests that a 4 C in vivo tem per a ture rise may be achiev able with this probe de sign. Fur ther test ing is needed to de - ter mine the op ti mal tis sue heat ing method (fo cused or unfocused beam) and trans mit cir - cuitry set tings and the ad di tion of match ing cir cuits to max i mize power out put ef fi ciency should en hance the de po si tion of ther mal en ergy as well. As a whole, the re sults from these ex per i ments, in par tic u lar the in vivo RT3D im ages and measured in-vitro tem per a ture rises, lead us to be lieve that the de vel op ment of an intracranial ul tra sound cath e ter com bin ing RT3D im ag ing and hyperthermia ca pa bil ity is fea si ble. In ad di tion to at tain ing a higher el e ment yield, build ing elec tri cal im ped ance match ing cir cuits and op ti miz ing the driv ing cir cuitry, the cath e ter it self must be built to man age the curves it must make through the vasculature for the in tended, min i mally-in va sive pro ce dure (Fig. 1). This will re quire a smaller, more flex i ble cath e ter for a next-gen er a tion de vice, perhaps similar to the 4.8 Fr de sign pre vi ously de scribed by Light and Smith for 3D intra vascular imaging. 40 This 4.8 Fr de vice con tains con duc tors for 100 chan nels and may be ad e quate to re pro duce our ar ray de sign in a smaller size while achiev ing suf fi cient acous tic out put for heat ing. In our long term plan, a brain tu mor will be lo cal ized by CT or MR im ag ing. Fluoroscopic angiography/venography will be used to guide the 3D ul tra sound cath e ter ret ro grade from the jug u lar vein to the dural ve nous si nuses so that the cath e ter is ad ja cent to or in close prox - im ity to the tu mor. A microbubble or liposomal agent mo lec u larly tar geted to re gions of tu - mor angiogenesis will be ad min is tered via in tra ve nous in jec tion. The dual-mode cath e ter will pro duce RT3D ul tra sound im ages with 3D color Dopp ler of the le sion and an ul tra sound hyperthermia beam will suf fi ciently warm the given agent, trig ger ing the re lease of a chemotherapeutic drug at the tu mor site. AC KNOWL EDGE MENTS The au thors would like to thank El len Dixon-Tulloch and Ste phen J. Hsu for their as sis - tance. This re search was sup ported by a 2008 Chandran Fam ily Foun da tion Re search Award and NIH grant EB

18 98 HERICKHOFF ET AL REFERENCES 1. American Cancer Society, ACS Can cer Facts & Fig ures 2008 (2008). 2. Cen tral Brain Tu mor Reg is try of the United States, CBTRUS Sta tis ti cal Re port: Pri mary Brain Tu mors in the United States, (2008). 3. Kleihues P, Cavenee WK. Pa thol ogy and Ge net ics of Tu mours of the Ner vous Sys tem (IARC Press, Lyon, 2000). 4. Mareel M. Anti-in va sive brain tu mor ther apy: gen eral as pects and fu ture strat e gies, in Brain Tumor Invasion: Biological, Clinical, and Therapeutic Considerations, Mikkelsen T, ed., pp. 393 (Wiley-Liss, 1998). 5. Prados MD. Sys temic che mo ther apy, in Neuro-Oncology: The Essentials, Bernstein M, Berger MS, eds., pp (Thieme Medical Publishers, 2000). 6. Lesniak MS. Tar get ing drugs to tu mors of the cen tral ner vous sys tem, in Brain Tu mors, Ali-Osman F, ed., pp. 353 (Humana Press, 2005). 7. Kroll RA, Neuwelt EA. Out wit ting the blood-brain bar rier for ther a peu tic pur poses: os motic open ing and other means, Neu ro sur gery 42, (1998). 8. Zunkeler B, Car son RE, Olson J, et al. Quan ti fi ca tion and pharmacokinetics of blood-brain bar rier dis rup tion in hu mans, J Neurosurg 85, (1996). 9. Mesiwala AH, Farrell L, Wenzel HJ, et al. High-in ten sity fo cused ul tra sound se lec tively dis rupts the blood-brain bar rier in vivo, Ul tra sound Med Biol 28, (2002). 10. Choi JJ, Pernot M, Small SA, Konofagou EE. Noninvasive, transcranial and lo cal ized open ing of the blood-brain bar rier us ing fo cused ul tra sound in mice, Ul tra sound Med Biol 33, (2007). 11. Hynynen K, McDannold N, Vykhodtseva N, Jolesz FA. Noninvasive MR im ag ing-guided fo cal open ing of the blood-brain bar rier in rab bits, Radiology 220, (2001). 12. McDannold N, Vykhodtseva N, Hynynen K. Use of ul tra sound pulses com bined with definity for tar geted blood-brain barrier disruption: a feasibility study, Ul tra sound Med Biol 33, (2007). 13. McDannold N, Vykhodtseva N, Hynynen K. Blood-brain bar rier dis rup tion in duced by fo cused ul tra sound and circulating preformed microbubbles appears to be characterized by the mechanical index, Ultrasound Med Biol 34, (2008). 14. Vykhodtseva N, McDannold N, Hynynen K. Prog ress and prob lems in the ap pli ca tion of fo cused ul tra sound for blood-brain bar rier dis rup tion, Ultrasonics 48, (2008). 15. Cho CW, Liu Y, Cobb WN, et al. Ul tra sound-in duced mild hyperthermia as a novel ap proach to in crease drug up take in brain microvessel en do the lial cells, Pharm Res 19, (2002). 16. Aubry JF, Tan ter M, Pernot M, Thomas JL, Fink M. Ex per i men tal dem on stra tion of noninvasive transskull adaptive focusing based on prior computed tomography scans, J Acoust Soc Am 113, (2003). 17. Guthkelch AN, Carter LP, Cassady JR, et al. Treat ment of ma lig nant brain tu mors with fo cused ul tra sound hyperthermia and ra di a tion: re sults of a phase I trial, J Neuro-Oncol 10, (1991). 18. Hynynen K, Jolesz FA. Dem on stra tion of po ten tial noninvasive ul tra sound brain ther apy through an in tact skull, Ul tra sound Med Biol 24, (1998). 19. Hynynen K, Clem ent GT, McDannold N, et al. 500-el e ment ul tra sound phased ar ray sys tem for noninvasive fo cal sur gery of the brain: a pre lim i nary rab bit study with ex vivo hu man skulls, Magn Reson Med 52, (2004). 20. Connors JJ, Wojak JC. Interventional Neuroradiology: Strat e gies and Prac ti cal Tech niques (Saunders, Phil - adelphia, 1999). 21. Mor ris P. Interventional and Endovascular Ther apy of the Ner vous Sys tem: A Prac tice Guide (Springer, New York, 2002). 22. Novak Z, Coldwell DM, Brega KE. Se lec tive in fu sion of urokinase and thrombectomy in the treat ment of acute cerebral sinus thrombosis, Am J Neuroradiol 21, (2000). 23. Chow K, Gobin YP, Saver J, et al. Endovascular treat ment of dural si nus throm bo sis with rheolytic thrombectomy and intra-ar te rial thrombolysis, Stroke 31, (2000). 24. Dowd CF, Malek AM, Phatouros CC, Hemphill JC. Ap pli ca tion of a rheolytic thrombectomy de vice in the treat ment of dural si nus throm bo sis: a new tech nique, Am J Neuroradiol 20, (1999).

19 DUAL-MODE INTRACRANIAL CATHETER FOR NEURO-ONCOLOGY Horowitz M, Purdy P, Unwin H, et al. Treat ment of dural si nus throm bo sis us ing se lec tive catheterization and urokinase, Ann Neurol 38, (1995). 26. Kuether TA, O Neill O, Nesbit GM, Barnwell SL. Endovascular treat ment of trau matic dural si nus throm bo - sis: case re port, Neurosurg 42, (1998). 27. Opatowsky MJ, Mor ris PP, Regan JD, Mewborne JD, Wil son JA. Rapid thrombectomy of su pe rior sagittal si - nus and trans verse si nus throm bo sis with a rheolytic cath e ter de vice, Am J Neuroradiol 20, (1999). 28. Makin IRS, Mast TD, Faidi W, et al. Min ia tur ized ul tra sound ar rays for in ter sti tial ab la tion and im ag ing, Ul - tra sound Med Biol 31, (2005). 29. Pua EC, Qiu YP, Smith SW. In te grated en do scope for real-time 3D ul tra sound im ag ing and hyperthermia: fea si bil ity study, Ultrasonic Imaging 29, 1-14 (2007). 30. Gen try KL, Smith SW. In te grated cath e ter for 3-D intracardiac echo car di og ra phy and ul tra sound ab la tion, IEEE Trans Ultrason Ferroelec Freq Contr 51, (2004). 31. Wu F, Chen WZ, Bai J, et al. Patho log i cal changes in hu man ma lig nant car ci noma treated with high-in ten sity focused ultrasound, Ul tra sound Med Biol 27, (2001). 32. Ebbini ES, Yao H, Shrestha A. Dual-mode ul tra sound phased ar rays for im age-guided sur gery, Ultrasonic Imaging 28, (2006). 33. Bouchoux G, Lafon C, Berriet R, et al. Dual-mode ul tra sound trans ducer for im age-guided in ter sti tial ther mal therapy, Ul tra sound Med Biol 34, (2008). 34. Hynynen K. Re view of ul tra sound ther apy, in Proc IEEE Ultrasonics Symp, pp (IEEE cat. no. 97CH36118, 1997). 35. Barthe PG, Slayton MH, Jae ger PM, et al. Ul tra sound ther apy sys tem and ab la tion re sults uti liz ing min ia ture imaging/therapy arrays, in Proc IEEE Ultrasonics Symp, pp (IEEE cat. no. 04CH37553, 2004). 36. von Ramm OT, Smith SW, Pavy HG. High-speed ul tra sound vol u met ric im ag ing sys tem. 2. par al lel pro cess - ing and im age dis play, IEEE Trans Ultrason Ferroelec Freq Contr 38, (1991). 37. Light ED, Mukundan S, Wolf PD, Smith SW. Real-time 3-D intracranial ul tra sound with an en do scopic ma - trix ar ray trans ducer, Ul tra sound Med Biol 33, (2007). 38. Smith SW, Pavy HG, von Ramm OT. High- speed ul tra sound vol u met ric im ag ing sys tem.1. trans ducer de - sign and beam steer ing, IEEE Trans Ultrason Ferroelec Freq Contr 38, (1991). 39. Lee W, Idriss SF, Wolf PD, Smith SW. A min ia tur ized cath e ter 2-D ar ray for real-time, 3-D intracardiac echo cardiography, IEEE Trans Ultrason Ferroelec Freq Contr 51, (2004). 40. Light ED, Smith SW. Two di men sional ar rays for real time 3D intravascular ul tra sound, Ultrasonic Imaging 26, (2004). 41. Pua EC, Idriss SF, Wolf PD, Smith SW. Real-time 3d transesophageal echocardiography, Ultrasonic Imag - ing 26, (2004). 42. Light ED, Idriss SF, Sullivan KF, Wolf PD, Smith SW. Real-time 3D lap aro scopic ultrasonography, Ul tra - sonic Imaging 26, (2005). 43. Pua EC, Fronheiser MP, No ble JR, et al. 3-D ul tra sound guid ance of sur gi cal ro bot ics: a fea si bil ity study, IEEE Trans Ultrason Ferroelec Freq Contr 53, (2006). 44. Ivancevich NM, Pinton GF, Nicoletto HA, et al. Real-time 3-D con trast-en hanced transcranial ul tra sound and aberration correction, Ul tra sound Med Biol 34, (2008). 45. Smith SW, Chu K, Idriss SF, et al. Fea si bil ity study: real-time 3-D ul tra sound im ag ing of the brain, Ul tra - sound Med Biol 30, (2004). 46. Blomley MJK, Cooke JC, Unger EC, Monaghan MJ, Cosgrove DO. Sci ence, med i cine, and the fu ture - microbubble con trast agents: a new era in ul tra sound, Brit Med J 322, (2001). 47. Goldberg BB, Raichlen JS, Forsberg F. Ultrasound Contrast Agents : Basic Principles and Clinical Applications (Dunitz, Lon don, 2001). 48. Kheirolomoom A, Dayton PA, Lum AFH, et al. Acous ti cally-ac tive microbubbles con ju gated to liposomes: characterization of a proposed drug delivery vehicle, J Con trol Re lease 118, (2007). 49. Bloch SH, Dayton PA, Ferrara KW. Tar geted im ag ing us ing ul tra sound con trast agents, IEEE Eng Med Biol 23, (2004). 50. Ponce AM, Vujaskovic Z, Yuan F, Needham D, Dewhirst MW. Hyperthermia me di ated liposomal drug de - liv ery, Int J Hyperther 22, (2006).