Cath e ter Ul tra sound Phased-Ar ray Trans duc ers for Ther mal Ab la tion: A Fea si bil ity Study

Transcription

1 ULTRASONIC IMAGING 27, (2005) Cath e ter Ul tra sound Phased-Ar ray Trans duc ers for Ther mal Ab la tion: A Fea si bil ity Study KENNETH L. GENTRY, 1 NASHEER SACHEDINA AND STEPHEN W. SMITH De part ment of Bio med i cal En gi neer ing Duke University Dur ham, NC kgentry@wisc.edu The feasibility of catheter single-element ultrasound transducers for cardiac ablation has been shown pre vi ously. We de scribe the de sign and test ing of cath e ter-sized lin ear phased ar rays trans duc ers for ul - tra sound ab la tion. One ar ray has 86 PZT-4 el e ments op er at ing at 8 MHz and 5 MHz. The over all ar ray size is 14.9 mm by 3.1 mm (10 Fr). The other ar ray has 50 PZT-5 el e ments op er at ing at 4 MHz and is 17 mm by 3.1 mm (10 Fr). In or der to pro duce the in ten sity needed to cre ate le sions in heart tis sue, we mod - i fied a real-time, 3D scan ner to pro duce 100 V pp 256-cy cle trans mit pulses at a pulse rep e ti tion fre quency of 14.1 khz. This made it pos si ble for the PZT-4 and PZT-5 trans duc ers to pro duce I SPTA of 3.26 W/cm 2 and 142 W/cm 2, re spec tively. When driv ing the trans duc ers at high duty fac tor, the trans mit cir cuitry in the scan ner was dam aged. A me chan i cally-fo cused trans ducer with the same di men sions as the PZT-4 trans ducer was built. When trans mit ting con tin u ously at 9 MHz, it pro duced an I SPTA of 29.3 W/cm 2. This cre ated a le sion 5 mm across and 5 mm deep in beef tis sue while rais ing the fo cal tem per a ture 23 C. Ab - la tion is within the ca pa bil i ties of a cath e ter phased ar ray trans ducer in te grated into a di ag nos tic ul tra - sound scan ner. Key words: Cath e ter trans ducer; ultrasound ablation. I. IN TRO DUC TION Sev eral re search ers have shown the fea si bil ity of us ing cath e ter-de liv ered ul tra sound to ab late car diac tis sue for treat ment of arrhythmias. 1, 2 This tech nol ogy is be ing ex plored as a re place ment for the cur rent method of us ing intracardiac radio fre quen cy (rf) cath e ters to ab - late the site of ar rhyth mia. For rf ab la tion, typ i cally 50 W of power is de liv ered for up to 60 s per ablation site. 3 Ini tial clin i cal stud ies us ing ul tra sound ab la tion cath e ters have shown that 40 W of acous tic power ap plied for s can cre ate cir cumfer ential le sions in the pul mo - nary vein ostia in the left atrium. 4 Ex per i men tal an i mal stud ies have also dem on strated the ef fi cacy of us ing ul tra sound ab la tion cath e ters transurethrally for treat ment of the pros tate. Here in di vid ual le sions are cre ated by ap ply ing 20 W of acous tic power for 90 s. 5 We have pre vi ously de scribed two gen er a tions of de vices in te grat ing real-time, three-di - men sional ul tra sound im ag ing and ul tra sound ab la tion: a 12 Fr (out side di am e ter = 4.0 mm) side-view ing cath e ter with a 5 MHz im ag ing ar ray ad ja cent to a 10 MHz ab la tion pis ton transducer and five integrated electrocardiogram electrodes 6 (Fig. 1a) and a for ward-view - ing 14 Fr (out side di am e ter = 4.5 mm) cath e ter with a 5 MHz im ag ing ar ray and a 10 MHz 1 Present address: 3115 Engineering Centers Building, 1550 Engineering Drive, Madison, WI /05 $18.00 Copy right 2005 by Dynamedia, Inc. All rights of re pro duc tion in any form re served.

2 90 GENTRY ET AL FIG. 1 (a) The first generation device was a side-looking device with a 64 element, two-dimensional imaging array (I), an ablation piston (A) and five electrocardiogram electrodes (E). (b) The second generation device was a forward-looking device with a 112 element, two-dimensional imaging array (A) and an ablation annulus (I). FIG. 2 In the proposed device, two linear phased arrays transmit an ablation beam. They surround a 112 element, two-dimensional imaging array to be used for real-time, three-dimensional ultrasound imaging. Here the device is shown imaging two pulmonary vein ostia (PV) from in the left atrium and ablating a linear lesion (xxx) to electrically isolate the PVs. PZT-4 ablation annulus (outside diameter of 4.5 mm and inside diameter of 3.1 mm) placed 7, 8 around the imaging array (Fig. 1b). These earlier devices produced spot lesions with diameters on the order of 4 mm after s ablation. The treatment of arrhythmias in the atria frequently requires the creation of linear lesions in the myocardium 3 mm wide and up to 2 cm long using rf techniques.9 Therefore in a recent work, using computer simulations we examined designs for a side-looking, linear phased array catheter transducer to produce linear lesions by scanning a distance of 2 cm from the target tissue (Fig. 2).10 The basic design includes a 112 element, 5 MHz imaging array11 for

3 CATH E TER UL TRA SOUND TRANS DUC ERS FOR AB LA TION 91 real-time, three-di men sional im ag ing. This ar ray is in the mid dle of two 43 el e ment, 8 MHz lin ear phased ar rays, which, in com bi na tion, trans mit an ab la tion beam that can be fo cused and steered in the az i muth di rec tion, i.e., along the length of the cath e ter, to cre ate lin ear le - sions. We es ti mated that such an ar ray could pro duce a spa tial-peak, tem po ral-av er age in - tensity (I SPTA ) of 25 W/cm 2 when fo cused through 2 cm of blood and 42 W/cm 2 when fo cused through wa ter. Fi nite el e ment anal y sis (FEA) fur ther showed that in the lat ter case, the tem - per a ture rise in tis sue at the fo cus (18.4 C af ter 120 s) is ad e quate for pro duc ing a le sion as 12, 13 de ter mined us ing a ther mal dose model. In this pa per, we de scribe our ef forts to ex per i men tally con firm our FEA find ings in wa ter by build ing a lin ear phased ar ray trans ducer and con nect ing it to our Model 1 real-time, 3D ul tra sound scan ner (Volu met rics Med i cal Im ag ing, Dur ham, NC) for steer ing and fo cus - ing. 14, 15 We de scribe build ing and test ing two ar ray de signs, one op er at ing at 8 MHz and one op er at ing at 4 MHz. Ad di tion ally, we ex plain mod i fy ing the scan ner to pro duce higher power and mea sur ing acous tic out put from the ar rays while driven by the scan ner. In Re - sults, we re port our find ings from the two ar rays and the mod i fied scan ner. In the Dis cus sion sec tion, we in ves ti gate some prob lems we ex pe ri enced us ing the scan ner, and re port on a me chan i cally fo cused ar ray. Fi nally, we con clude with some thoughts on pos si ble fu ture work. Scanner modification II. METH ODS The typ i cal pulses used in di ag nos tic ul tra sound im ag ing are 1-10 cy cles long with a pulse rep e ti tion fre quency (prf) around 1-15 khz. Though the in stan ta neous in ten sity from these pulses may be high, due to their low duty fac tor the I SPTA pro duced in wa ter is on the or der of 1 W/cm 2. For ex am ple, the Model 1 scan ner and a 5 MHz transthoracic 2-D phased ar ray probe 16 pro duce an I SPTA of 782 mw/cm 2 when the scan depth is 10 cm (prf = 6.67 khz) and the trans mit pulse is 2 cy cles long. This cor re sponds to a duty fac tor of less than 0.3%. As fac tory con fig ured, the scan ner s max i mum duty fac tor is only 6.4% at 5 MHz (32 cy cle pulse with 10 khz prf) but pre vi ous ex pe ri ence has shown that a duty fac tor of at least 50% is de sir able for ther apy. There fore it was nec es sary to re con fig ure the Model 1 scan ner to pro - duce pulses with a higher num ber of cy cles and at a higher prf. By chang ing the scan ner s soft ware to de crease the min i mum scan depth to 3 cm from 6 cm, we were able to in crease the prf to 14.1 khz. In creas ing the num ber of cy cles was more dif fi cult, how ever, due to nu mer ous safety checks built into both the scan ner hard ware and soft ware. The user can vary the num ber of cy cles per trans mit pulse from 1 to 32, though the scan ner lim its the max i mum to four when both the prf and trans mit am pli tude are high. This num ber is sent to a field pro gram ma ble gate ar ray (FPGA) con trol ling the trans mit cir cuitry as a five bit num ber where it drives a se ries of five coun ters. Three coun ters (also con trolled by the same five bit num ber) were added to the se ries, in creas ing the max i mum num ber of cy cles to 256. Then the safety checks were by passed by us ing only the first two bits (i.e., one to four cy cles per pulse) to con trol all eight coun ters and cover the new range of pulse lengths. The FPGA pro duces a pair of dig i tal pulse trains, one as in put for a pos i tive-go ing am pli fier and the other for a neg a tive-go ing am pli fier. Com bined, these am pli fi ers pro duce the si nu soi dal trans mit pulse. The Model 1 scan ner is ca pa ble of trans mit ting at fre quen cies from 1.25 to 10 MHz. Our lin ear phased ar ray de sign is in the up per end of this range, so we first mea sured the peak-to-peak trans mit volt age across a 12k re sis tor from a sin gle chan nel across the range

4 92 GENTRY ET AL of fre quency. The scan ner is in ca pa ble of trans mit ting at any se lected fre quency. In stead, be cause its mas ter clock runs at 40 MHz and the trans mit pulses must be an in te ger num ber of those cy cles, the avail able fre quen cies are, for ex am ple, 5 MHz (8 clock cy cles), 6.7 MHz (6 clock cy cles) and 8 MHz (5 clock cy cles). Ad di tion ally, the trans mit volt age peaks at 2.9 MHz, and is 2 db down at 4 MHz and nearly 10 db down at 8MHz. Phased ar rays The to tal size of the 8 MHz lin ear phased ar ray pro posed in the pre vi ous pa per is 3.1 mm in the el e va tion di men sion by 14.9 mm in az i muth. 10 To build an in ex pen sive pro to type ar ray, PZT-4 (840, APC In ter na tional, Mackeyville, PA) with a 10 MHz nom i nal par al lel res o nant fre quency (230 m thick) was bonded to cop per-plated polyimide with sil ver ep oxy. The ar - ray and cop per lay ers were then diced on a 150 m (0.8 at 8 MHz) pitch, re sult ing in cop per traces elec tri cally con nected to the un der side of each trans ducer el e ment, but iso lated from each other. Liq uid crys tal poly mer (LCP) sput tered with 0.25 m gold on one side was bonded to the top of all el e ments as a com mon ground layer. Rather than at tempt to build the ar ray into a cath e ter, this pro to type was in stead bonded to an alu mi num sub strate with a 5 mm deep groove milled into it. The ar ray was po si tioned so that the trans ducer el e ments were over this groove, pro vid ing air back ing to in crease trans mit ef fi ciency. The layer thick - nesses were used in KLM model sim u la tions (PiezoCAD, Sonic Con cepts, Woodinville, WA) of the trans mit im pulse re sponse spec trum. 17 To com plete the trans ducer, the cop per traces were wired to in terface with the Model 1 scan ner trans ducer han dle. Be cause of the low trans mit volt ages at high fre quen cies, a sec ond pro to type de signed to op er ate at 4 MHz was also built. This sec ond ar ray used PZT-5H (TRS 610HD, TRS Tech - nol o gies, State Col lege, PA) with a nom i nal par al lel res o nant fre quency of 5 MHz (360 m thick) and was diced on a 300 m (0.8 at 4.0 MHz) pitch. In stead of alu mi num, it was built on a poly mer sub strate with a 5 mm deep groove to pro vide air back ing. Transducer characterization The trans mit im pulse re sponse from sev eral el e ments on both trans duc ers was mea sured. A pulser/re ceiver (5073PR, Panametrics, Waltham, MA) ex cited in di vid ual el e ments. The sig nal was re ceived on a cal i brated PVDF mem brane hydrophone (Model 804, Sonic Tech - nol o gies, Hatboro, PA) po si tioned ap prox i mately 1 cm away and was saved with an os cil lo - scope (744A, Tektronix, Wilsonville, OR). A dis crete Fou rier trans form was performed on the trans mit im pulse re sponse to ob tain the trans mit spec trum for com par i son to KLM sim u - lation. The I SPTA from the two ar rays was mea sured us ing the pro ce dures out lined by the Cen ter for De vices and Ra dio log i cal Health of the Food and Drug Ad min is tra tion. 18 The trans duc ers were po si tioned in a wa ter tank with the hydrophone at the fo cus. The scan ner trans mit ted one line fo cused at 2 cm on-axis. With the prf set to the min i mum 2.5 khz, the trans mit fre - quency was swept to find the max i mum re ceived pulse am pli tude. At the max i mum fre - quency, the num ber of cy cles per pulse was in creased from one cy cle to higher val ues and the in ten sity was mea sured at each set ting. Fi nally, the trans mit beam pro files of the trans ducer ar rays at a 2 cm fo cus were mea sured in both az i muth and el e va tion by re cord ing the peak-to-peak volt age as the hydrophone was swept in front of the ar rays at steps of 0.1 mm. Ad di tion ally, the az i muth beam pro files were mea sured when the 4 MHz PZT-5H ar ray was steered off-axis to an gles of (The dis - tance be tween these two foci is 17.9 mm to sim u late the pro duc tion of a lin ear le sion by phased ar ray scan ning.) These re sults were com pared to beam pro files sim u lated us ing the Field II pro gram. 19

5 CATH E TER UL TRA SOUND TRANS DUC ERS FOR AB LA TION 93 FIG MHz PZT-4 bonded to cop per-plated polyimide and diced. FIG. 4 5 MHz PZT-5 bonded to cop per-plated polyimide and diced. III. RESULTS Af ter bond ing and dic ing, the 8 MHz PZT-4 ar ray had 86 work ing el e ments (Fig. 3) and the 4 MHz PZT-5H ar ray had 50 work ing el e ments (Fig. 4) with the first five on one end shorted to gether. These yields were sus tained through later fab ri ca tion steps. The trans mit im pulse re sponse spec trum from the 8 MHz PZT-4 ar ray showed two peaks (Fig. 5), one at 7.6 MHz with a 6 db band width of 6.5% and a smaller peak at 5.2 MHz with a 6 db band - width of 12.8%. The KLM sim u la tion shows rea son able agree ment with three peaks; one at 4.6 MHz with 7.6% 6 db band width, one at 7.1 MHz with 5.6% 6 db band width and one at 11.2 MHz with 4.4% 6 db band width. The high and low fre quency peaks orig i nate in the back ing lay ers of the trans ducer. A typ i cal trans mit im pulse re sponse from the 4 MHz PZT-5H ar ray trans ducer shows a peak at 4.4 MHz with 35.6% 6 db band width (Fig. 6). The KLM sim u la tion of the el e ment has two peaks, but also shows rea son able agree ment with one at 4.7 MHz with 16.0% 6dB band width and one at 8.3 MHz with 10.8% 6dB band width. By de creas ing the min i mum scan depth to 3 cm, the max i mum prf was in creased to 14.1 khz. With the new max i mum pulse length of 256 cy cles, the max i mum duty fac tor is 45.1% at 8.0 MHz and 90.2% at 4 MHz. This pulse has sig nif i cant droop, how ever, be cause of sag in the power sup ply high volt age pro vided to the trans mit am pli fi ers. Fur ther, this sag is due to in suf fi ciently large ca pac i tors in the scan ner s high-volt age AC-to-DC con vert ers. The last cy cle of a 5 MHz 256-cy cle pulse is ap prox i mately half of the amplitude of the first cycle. The Model 1 scan ner was used to drive the trans duc ers at the min i mum prf of 2.5 khz as the trans mit fre quency was changed. The 8 MHz PZT-4 trans ducer pro duced a max i mum peak-to-peak out put volt age at 5 MHz and a lo cal max i mum at 8 MHz. The 4 MHz PZT-5H trans ducer s max i mum was at 4 MHz. We next at tempted to mea sure the in ten sity with in - creas ingly long pulse lengths. Un for tu nately, even at the min i mum prf lon ger pulses dam - aged the trans mit cir cuitry in the scan ner mak ing some chan nels in op er a ble. There fore,

6 94 GENTRY ET AL FIG. 5 The impulse re sponse spec trum from a typ i cal el e ment on the PZT-4 trans ducer shows a peak at 7.6 MHz. The 6 db band width is 6.5%. There is also a smaller peak at 5.2 MHz with a 6 db band width of 12.8%. The KLM sim u la tion shows three peaks; one at 4.6 MHz with 7.6% -6 db band width, one at 7.1 MHz with 5.6% -6 db band - width, and one at 11.2 MHz with 4.4% 6 db band width. FIG. 6 The im pulse re sponse spec trum from a typ i cal el e ment on the PZT-5H trans ducer shows a peak at 4.4 MHz. The 6 db band width is 35.6%. The KLM sim u la tion of the el e ment shows a cen ter fre quency of 4.7 MHz and a 6dB band width of 16.0%.

7 CATH E TER UL TRA SOUND TRANS DUC ERS FOR AB LA TION 95 TA BLE 1 The spa tial-peak, tem po ral av er age in ten sity from the trans duc ers was mea sured at a prf of 2.5 khz with four-cy cle pulses. This value was ex trap o lated to the the o ret i cal max i mum value pos si ble from the Model 1 scan ner by scal ing the in ten sity to a prf of 14.1 khz with 256-cy cle pulses. Trans ducer Frequency MHz Measured I SPTA W/cm 2 Extrapolated ISPTA W/cm 2 PZT PZT accurate I SPTA data with all trans ducer chan nels trans mit ting could only be col lected at a max - i mum of four cy cles af ter re plac ing the dam aged trans mit ters. These val ues were ex trap o - lated to the max i mum set tings pos si ble on the scan ner first by scal ing I SPTA to 14.1 khz in stead of 2.5 khz, and then by mul ti ply ing by 64 (to scale to 256 cy cles) and di vid ing by 2 (be cause of the volt age droop). These re sults in di cate that if the trans mit cir cuitry had not failed, the scan ner could achieve 3.26 W/cm 2 at 5 MHz with the PZT-4 trans ducer and 142 W/cm 2 at 4 MHz with the PZT-5H transducer (Table 1). The fo cal on-axis beam plots from the 4 MHz PZT-5H trans ducer were mea sured in both el e va tion and az i muth and com pared to sim u lated re sults (Figs. 7a, b). Be cause of its low in - ten sity, no beam plots were mea sured from the 8 MHz PZT-4 trans ducer. The 6 db beam - width of the PZT-5H trans ducer is 0.74 mm in az i muth (com pared to 0.63 mm sim u lated) and 3.4 mm in el e va tion (com pared to 3.5 mm sim u lated). Az i muthal side lobes are 12 db down at 1.0 mm and 11 db down at 1.1 mm. This com pares fa vor ably to sim u lated side lobes 8.5 db down at 0.8 mm. The beam was then steered off-axis at an gles of 26.6 and the fo cal beam plot was mea sured and com pared to sim u la tion (Fig. 8). The 6dB beam - width when steered at a pos i tive an gle is 0.73 mm in az i muth (com pared to 0.57 mm sim u - lated). In the other di rec tion, the mea sured 6dB beam width is 0.76 mm, com pared to a sim u lated value of 0.56 mm. Ad di tion ally, when steered off-axis the am pli tude of the trans - mit pulses are 0.3 db less than when on-axis. Con tours of 6 db beam widths for all three foci are plot ted on the same axes in fig ure 9 (el e va tion beam width is as sumed the same for all three foci) to il lus trate a lin ear le sion pro duced by phased ar ray scan ning. The di men - sions are 3.4 mm wide by 19 mm long which is good agree ment with Olgin et al. 9 IV. DISCUSSION The abil ity of the trans duc ers to pro duce ad e quate in ten sity for ab la tion was se verely lim - ited by the Model 1 scan ner. The PZT-4 trans ducer was de signed to op er ate at 8 MHz, but be cause of the air and other back ing lay ers it also trans mit ted sound at 5 MHz. Due to the roll-off in trans mit volt age over 3.5 MHz, this trans ducer ac tu ally pro duced higher in ten sity at the lower fre quency. The 4 MHz PZT-5H trans ducer op er ated closer to the trans mit peak of the scan ner and its own res o nant fre quency and thus was able to pro duce much higher in - ten sity. The mea sured and ex trap o lated in ten sity re sults were en cour ag ing, and seemed to in di cate that this trans ducer could pro duce enough in ten sity to pro duce a le sion in heart tis - sue, given a more ro bust scan ner. In or der to test this hy poth e sis, a third trans ducer was built sim i lar to the 8 MHz PZT-4 trans ducer. In the new trans ducer, the cop per layer was only partly diced so that all of the el - e ments are elec tri cally con nected on both top and bot tom. Rather than bond ing the

8 96 GENTRY ET AL FIG. 7 Mea sured beamplots (dots) are com pared to beamplots sim u lated in Field II (solid) for the 4 MHz PZT-5H ar ray steered on-axis. Az i muth (a) and el e va tion (b) plots are shown. FIG. 8 The mea sured beamplots (dots) are com pared to beamplots sim u lated in Field II (solid) for the 4 MHz PZT-5H ar ray steered off-axis. The top plot (a) is of the beam steered to an an gle of The bot tom plot (b) is of the beam steered to an an gle of 26.6.

9 CATH E TER UL TRA SOUND TRANS DUC ERS FOR AB LA TION 97 FIG. 9 The 4MHz PZT-5H trans ducer was steered on-axis and at ±26.6 off-axis. The 6 db beam widths at each foci are shown in di cat ing that it is pos si ble to cre ate a lin ear le sion 3.4 mm wide and 19 mm long by scan ning the ab la tion beam within the dashed lines. Fig. 10 The ra dius of cur va ture of the me chan i cally fo cused trans ducer is 20 mm. The ap er ture is oth er wise as shown in fig ure 3, ex cept 5 mm thick alu mi num re places the air. polyimide to an alu mi num sub strate with a groove, the trans ducer was bonded to the in side of an alu mi num half-cyl in der with a ra dius of 20 mm (Figs. 10a, b). As a re sult, all of the 86 trans ducer el e ments are me chan i cally fo cused 20 mm from the trans ducer face. The me chan i cally-fo cused trans ducer was driven by a sig nal source (8165A, Hewlett- Packard, Palo Alto, CA) and am pli fier (525 LA, ENI, Roch es ter, NY) at 9.0 MHz and the I SPTA was mea sured at a prf of 3.12 khz and a range of cy cles per pulse. The in ten sity is 5.1 W/cm 2 at 500 cy cles (Fig. 11). No in ten sity mea sure ments were made dur ing con tin u ous wave op er a tion to safe guard the hydrophone. When ex trap o lated to con tin u ous wave op er a - tion, the in ten sity from the trans ducer is 29.3 W/cm 2. Next, a ther mo cou ple was em bed ded 2 mm deep in 8 mm thick beef tis sue. The trans ducer was po si tioned so that the me chan i cal fo - cus was just in the tis sue and was driven con tin u ously for 120 s. The ther mo cou ple mea - sured the tem per a ture in the tis sue ev ery 15 s and the re sults com pared fa vor ably with our ear lier FEA re sults (Fig. 12). 10 This ab la tion pro ce dure pro duced a le sion 3 mm across and 5 mm deep in 8 mm thick beef tis sue. Fig ure 13 shows a pho to graph of a cross-sec tion of the le sion with a dashed line de not ing its ex tent. Con vert ing a di ag nos tic scan ner to a com bined ab la tion/im ag ing sys tem places large de - mands on all com po nents of the sys tem. A suc cess ful phased-ar ray ab la tion sys tem must be highly ef fi cient and be able to han dle high av er age power trans mis sion, rather than just high peak power as in a di ag nos tic sys tem. There are nu mer ous ways that the trans ducer ef fi - ciency might be in creased. This would re duce the power drawn from the Model 1 scan ner, pos si bly al low ing it to be used with out dam age. For ex am ple, elec tri cal match ing could re -

10 98 GENTRY ET AL FIG. 11 The mechanically-focused array produced a maximum I SPTA of 5.11 W/cm 2 at a prf of 3.12 khz (solid mark ers). When ex trap o lated to 17.9 khz, i.e., con tin u ous wave op er a tion at 500 cy cles(dashed lines, empty mark - ers), the mechanically-focused array produced a maximum I SPTA of 29.3 W/cm 2. Fig. 12 The tem per a ture at the fo cus of the me chan i cally-focused trans ducer (empty cir cles) rose 23 C in two min utes as mea sured by a ther mo cou ple em bed ded in bo vine tis sue. These re sults com pare fa vor ably to those found us ing FEA in the pre vi ous pa per cited in the text (solid line).

11 CATH E TER UL TRA SOUND TRANS DUC ERS FOR AB LA TION 99 FIG. 13 The le sion (out lined re gion in cen ter of im age) was nearly sym met ric on the tis sue sur face with a di am e - ter of 3 mm. It was 5 mm deep into the tis sue duce the cur rent drawn by the trans ducer el e ments and low-loss match ing lay ers could better cou ple acous tic en ergy into the tis sue. While the scan ner was able to pro duce the needed trans mit volt age on in di vid ual trans ducer el e ments, when the en tire trans ducer was driven with long pulses at high prf the trans mit cir cuitry was dam aged. Ad di tion ally, the scan ner was in ca pa ble of pro duc ing high fre quency trans mit pulses with large peak-to-peak volt age. Even within these lim i ta tions, how ever, the 4 MHz PZT-5H trans ducer was able to pro duce an in ten sity of 786 mw/cm 2 with a duty fac tor of 0.25%. As shown by the me chan i cally-fo - cused trans ducer, an I SPTA of 29.3 W/cm 2 will pro duce a le sion. If the value mea sured from the 4 MHz PZT-5H trans ducer is ex trap o lated to a 256-cy cle pulse (i.e., mul ti ply by 64 and di vide by 2), a prf of 2.9 khz (duty fac tor of 19%) will trans mit at that in ten sity. We have also shown an abil ity to steer the 4 MHz PZT-5H trans ducer off-axis with out a sub stan tial de crease in in ten sity or in crease in beam width. Thus, the cre ation of a lin ear le sion 19 mm long by 3.4 mm wide is achiev able. AC KNOWL EDG MENTS The au thors would like to thank John Schachte for his help pro gram ming the dig i tal trans - mit cir cuitry, Ron Hileman for his help an a lyz ing the an a log trans mit cir cuitry, and Chris Pua for his help re-pro gram ming the trans mit con trol soft ware. This work was sup ported by NIH grants HL and HL and NSF grant DMR REF ER ENCES 1. Zimmer JE, Hynynen K, He DS, Marcus F. The fea si bil ity of us ing ul tra sound for car diac ab la tion, IEEE Trans Biomed Eng 42, (1995).

12 100 GENTRY ET AL 2. Lesh MD, Guerra PG, Roithinger FX, et al. Novel cath e ter tech nol ogy for ab la tive cure of atrial fi bril la tion, J Interv Card Electr 4, (2000). 3. McRury ID, Haines DE. Ab la tion for the treat ment of arrhythmias, Proc IEEE 84, (1996). 4. Meininger GR, Calkins H, Lickfett L, et al. Ini tial ex pe ri ence with a novel fo cused ul tra sound ab la tion sys tem for ring ab la tion out side the pul mo nary vein, J Interven Car diac Electrophys 8, (2003). 5. Seip R, Biswas M, Kuznetsov M, Sanghvi N. Transurethral high in ten sity fo cused ul tra sound: cath e ter based prototypes and experimental results, in Proc IEEE Ultrasonics Symp, pp ( IEEE cat. no. 00CH37121, 2000). 6. Gen try KL, Light ED, Smith SW. Com bined 3-D intracardiac echo and ul tra sound ab la tion, in Proc SPIE Medical Imaging, pp (SPIE cat. no. 5035, 2003). 7. Gen try KL, Smith SW. In te grated cath e ter for 3-D intracardiac echocardiography and ul tra sound ab la tion, in Proc IEEE Ul tra sonics Symp, pp (IEEE cat. no. 03CH37476, 2003). 8. Gen try KL, Smith SW. In te grated cath e ter for 3-D intracardiac echocardiography and ul tra sound ab la tion, IEEE Trans Ultrason Ferroelec Freq Contr 51, (2004). 9. Olgin JE, Kalman JM, Chin M, et al. Electrophysiological ef fects of long, lin ear atrial le sions placed un der intracardiac ultrasound guidance, Circulation 96, (1997). 10. Gen try KL, Palmeri ML, Sachedina N, Smith SW. Fi nite el e ment anal y sis of tem per a ture rise and le sion for - mation from catheter ultrasound ablation transducers, IEEE Trans Ultrason Ferroelec Freq Contr (in press). 11. Lee W, Idriss SF, Wolf PD, Smith SW. Dual lu men trans ducer probes for real-time 3-D interventional car - diac ultrasound, Ul tra sound Med Biol 29, (2003). 12. Damianou CA, Hynynen K, Fan X. Eval u a tion of ac cu racy of a the o ret i cal model for pre dict ing the necrosed tissue volume during focused ultrasound surgery, IEEE Trans Ultrason Ferroelec Freq Contr 42, (1995). 13. Sapareto SA, Dewey WC. Ther mal dose de ter mi na tion in can cer ther apy, Int J Radiat Oncol 10, (1984). 14. Smith SW, Pavy HG, von Ramm OT. High-speed ul tra sound vol u met ric im ag ing sys tem part I: trans ducer de sign and beam steer ing, IEEE Trans Ultrason Ferroelec Freq Contr 38, (1991). 15. von Ramm OT, Smith SW, Pavy HG. High-speed ul tra sound vol u met ric im ag ing sys tem part II: par al lel processing and image display, IEEE Trans Ultrason Ferroelec Freq Contr 38, (1991). 16. Light ED, Fiering JO, Hultman PA, Lee W, Smith SW. Up date of two di men sional ar rays for real time vol u - met ric im ag ing and real time intracardiac im ag ing, in Proc IEEE Ultrasonics Symp, pp (IEEE cat. no. 99CH37027, 1999). 17. Krimholtz R, Leedom DA, Matthaei GL. New equiv a lent cir cuits for el e men tary pi ezo elec tric trans duc ers, Elec tron Lett 6, pp (1970). 18. Center for Devices and Radiological Health. 510(k) Guide for Measuring and Reporting Acoustic Output of Di ag nos tic Ul trasound Med i cal De vices (FDA, Rockville, MD, 1985). 19. Jensen JA, Svendsen NB. Cal cu la tion of pres sure fields from ar bi trarily shaped, apodized, and ex cited ul tra - sound transducers, IEEE Trans Ultrason Ferroelec Freq Contr 39 (1992).