Electrophysiological Changes Correlated with Temperature Increases Induced by High-Intensity Focused Ultrasound Ablation

Electrophysiological Changes Correlated with Temperature Increases Induced by High-Intensity Focused Ultrasound Ablation
of 17
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  d  Original Contribution ELECTROPHYSIOLOGICAL CHANGES CORRELATED WITH TEMPERATUREINCREASES INDUCED BY HIGH-INTENSITY FOCUSED ULTRASOUNDABLATION Z IQI  W U ,* R ONALD  E. K  UMON ,* J ACOB  I. L AUGHNER , y I GOR  R. E FIMOV , y and C HERI  X. D ENG * *Department of Biomedical Engineering, University of Michigan, Ann Arbor, MI, USA; and  y Department of BiomedicalEngineering, Washington University at Saint Louis, MO, USA (  Received   25  November   2013;  revised   30  August   2014;  in final form  4  September   2014) Abstract—To gain better understanding of the detailed mechanisms of high-intensity focused ultrasound (HIFU)ablation for cardiac arrhythmias, we investigated how the cellular electrophysiological (EP) changes were corre-lated with temperature increases and thermal dose (cumulative equivalent minutes [ CEM   43 ]) during HIFU appli-cationusingLangendorff-perfusedrabbithearts.Employingvoltage-sensitivedyedi-4-ANEPPS,wemeasuredtheEP and temperature during HIFU using simultaneous optical mapping and infrared imaging. Both action poten-tial amplitude (APA) and action potential duration at 50% repolarization (APD 50 ) decreased with temperatureincreases, and APD 50  was more thermally sensitive than APA. EP and tissue changes were irreversible whenHIFU-induced temperature increased above 52.3 ± 1.4  C and log 10 ( CEM   43 ) above 2.16 ± 0.51 (n 5 5), but werereversible when temperature was below 50.1 ± 0.8  C and log 10 ( CEM   43 ) below –0.9 ± 0.3 (n 5 9). EP and temper-ature/thermal dose changes were spatially correlated with HIFU-induced tissue necrosis surrounded by a transi-tion zone. (E-mail:   2015 World Federation for Ultrasound in Medicine & Biology.  Key Words:  Bioeffects, Langendorff-perfused rabbit hearts, High-intensity focused ultrasound, Optical mapping,Infrared imaging, Cellular electrophysiology. INTRODUCTION Cardiac arrhythmias are the irregular electrical activitiesduring heart cycles. They greatly compromise the heartfunction and represent a major clinical problem affectingmillions of patients (Kannel et al. 1987). Ablation ther-apy, including radiofrequency (RF), cryothermal, micro-wave, and laser ablation (Aktas et al. 2008; Dewire andCalkins 2010; Lall and Damiano 2007; Scheinman andMorady 2001) has developed as an important treatmentoption. However, these ablation techniques requiredirect tissue contact and rely on thermal conduction,thus ablation depth is often limited and ablationvolumes are not well-confined, leading to collateral dam-ages (Lall and Damiano 2007) and unsuccessful out-comes (Nanthakumar et al. 2004).High intensity focused ultrasound (HIFU) ablationuses concentrated ultrasound energy to generate rapidtemperature increases and tissue modifications ( e.g.,  pro-tein denaturation and tissue coagulation) in a well-confined volume, without relying on heat conductionfrom tissue surface through interveningtissue. Itprovidesa promising technology for cardiac ablation especiallywhen intramural ablation is desired (Okumura et al.2008). For example, HIFU has been used for atrio-ventricular node ablation (Strickberger et al. 1999) andpulmonary vein isolation (PVI) to treat paroxysmal atrialfibrillation (AF) in multi-center trials (Aktas et al. 2008;Ninet et al. 2005). One study reported that 85% of 103patients treated with HIFU were free from AF at 6-mofollow-up (Aktas et al. 2008; Ninet et al. 2005).However, recent clinical trials using a HIFU balloonsystem (Neven et al. 2010) showed ablation-related com-plications including atrial-esophageal fistula, pulmonaryembolism and phrenic nerve injury, and 28% of 32 pa-tients showed electrical reconduction after initial PVIand underwent repeated procedures (Metzner et al.2010). These results indicate the need to better under-stand the detailed effects of HIFU in cardiac ablation toimprove the technology.As in other HIFU thermal applications,  in vitro studies have investigated the relationships between Ultrasound in Med. & Biol., Vol. 41, No. 2, pp. 432–448, 2015Copyright  2015 World Federation for Ultrasound in Medicine & BiologyPrinted in the USA. All rights reserved0301-5629/$ - see front matter Address correspondence to: Cheri X. Deng, Department of Biomedical Engineering, University of Michigan, 2200 BonisteelBlvd, Ann Arbor, MI 48109-2099. E-mail: xdeng@umich.edu432  HIFUconditions( e.g., acousticpowerandexposuretime)and tissue status changes (Engel et al. 2006; Fujikuraet al. 2006; Zimmer et al. 1995), while temperature andthermal dose were often used as feedbacks in HIFUablation and prediction of lesion formation (Rivenset al. 2007). It has been found that when tissue tempera-ture reaches 50  C or higher for a certain duration, tissuenecrosis occurs and becomes non-conductive (Haines1993). In particular, changes in the cellular conductivityand excitability are important in cardiac ablation andthe electrophysiological (EP) changes need to be moni-toredandultimatelyconfirmedtoensuresuccessfulisola-tion/elimination of arrhythmic foci. However, detailedspatiotemporal EP changes and their correlations withHIFU-induced temperature or thermal dose increaseshave not been specifically investigated previously.As the fundamental functional units, myocytes andtheir couplingsare able togeneratean electricaltransient,the action potential (AP), while the initiation (depolariza-tion) and termination (repolarization) of APs as well astheir propagations among myocytescan affect the electri-cal functions of the whole heart (Giridhar et al. 2012). Afew features such as resting membrane potentials(RMPs), AP amplitude (APA), AP duration (APD), and etc.  (Nath et al. 1993; Wood and Fuller 2002; Wu et al.1999) are often used to characterize AP morphologyand the functionalities of myocytes. Nath et al. (1993)first demonstrated the EP properties as well as cellularautomaticity and excitability were temperature depen-dent in hyperthermia providing useful information forthe development of RF and microwave ablation tech-niques. Haines (1993) and Wu et al. (1999) further illus- trated that RF energy had a combined thermal andelectronic effect on changing cellular EP. Wood andFuller (2002) later demonstrated thatRFcaninduceacutecellular EP changes surrounding RF lesions, which canrecover completely. Recently, three potential effects inHIFUcardiac ablation were identified, includingthermal,radiation force and cavitation, and the latter two factorscan interferewith HIFU thermal effect and lead to unsuc-cessful ablation (Laughner et al. 2012b). As the predom-inant mechanism of HIFU ablation was assumed to bethermal (Khokhlova et al. 2013), it is of interest to inves-tigate whether HIFU cardiac ablation induces similar EPchanges as in RF ablation and how HIFU-induced lesionsarecorrelatedwithcellularEPchanges.Suchinformationhas not been available before but can reveal the detailedcharacteristics and mechanisms of HIFU ablation of car-diac arrhythmias and possibly other thermal ablationtherapies.Previously, we demonstrated the feasibility of fluo-rescence optical mapping (Deng et al. 2005) formeasuring HIFU-induced cardiac EP changes usingLangendorff-perfused intact rabbit heart preparations.In the present study, by combining optical mappingwith infrared (IR) thermography, we obtained the firstsimultaneous measurements of EP changes and tempera-ture increases as well as thermal dose during HIFU appli-cation to help reveal their detailed spatiotemporalcorrelation in HIFU cardiac ablation. METHODS  Isolated heart preparations New Zealand white rabbits (n 5 11,    3 kg) wereused in the study according to a protocol approved bytheCommitteeontheUseandCareofAnimalsatUniver-sity of Michigan (Protocol #: PRO00003842). The ani-mals were anesthetized intramuscularly  via  ketamine(35 mg/kg) and xylazine (5 mg/kg), heparinized (1000U/kg) and euthanized intravenously by injection of so-dium pentobarbital (100 mg/kg). The heart was harvestedafter mid-sternotomy, and rapidly washed with warmedTyrode’s solution. As described before (Laughner et al.2012b), the heart was immediately placed on a Langen-dorff apparatus and retrogradely perfused withoxygenated Tyrode’s solution (pH  5  7.35  6  0.05;37  C; 95% O 2  /5% CO 2 )  via  aorta at constant pressure(60 2 80 mmHg). Peripheral tissue was cleaned and theheart was superfused in Tyrode’s solution at 35 2 37  C.Excitation-contraction decoupler, 2, 3-butanedione mon-oxime (BDM, 15 mM; Fisher Scientific, Pittsburgh, PA,USA) was administered to mechanically silence the heartbeat motion that can introduce artifacts during imaging.The electrocardiogram was recorded with two needleelectrodes and digitized (Powerlab 26 T, ADInstruments,Colorado Springs, CO, USA) while the aortic perfusionpressure was continuously monitored (BLPR2, WorldPrecision Instruments, Sarasota, FL, USA) to ensurephysiologic stability of the heart preparation. Experimental setup with simultaneous optical mappingand infrared imaging As shown in Figure 1, the heart preparation wasplaced in a custom-made holder with an acousticallytransparent window (4.5 3 4.5 cm 2 ) for ultrasound trans-mission from below. The heart was pinned to a siliconegel pad (Sylgard 184, Dow Corning, Midland, MI,USA) in the holder, which was controlled using a 3-Dpositioning stage (EPS300, Newport, Irvine, CA, USA)for precise (50  m m) spatial alignment.A concave single element HIFU transducer(2.0 MHz, H-148, Sonic Concepts, Bothell, WA, USA)(focal length 63.2 mm and F-number 0.95), driven by asignal generator (33250 A, Agilent, Santa Clara, CA,USA) and a 75 W power amplifier (75 A250, AmplifierResearch, Souderton, PA, USA), was used to generateHIFU exposures. The focal peak compressional/  HIFU-induced electrophysiological changes and temperature increases d Z. W U  et al . 433  rarefactional pressures were measured to be  p c  5  7.0–8.7 MPaand  p r  5 –4.4 to –5.2 MPainwater (in free field)using a custom fiber optic probe hydrophone system(Parsons et al. 2006) with 6% uncertainty (type B). The6-dB focal zone was determined to be 7.2  3  0.9 mm(length  3  width) for  p c  and 8.7 3  1.2 mm for  p r   when  I  sppa  was at 846 W/cm 2 . The corresponding spatial-peak pulse-average intensities (  I  sppa ) were calculated to be846–1266 W/cm 2 using the equation (Hsiao et al. 2013)  I  sppa 5 1 nT  ð  t  0 1 nT  t  0  p 2 ð t  Þ  p 0 c 0 dt   (1)where  p ð t  Þ  is the transient pressure value at time  t  ;  r 0  and c 0  are the density and speed of sound of the ambient wa-ter, respectively;  t  0  is the starting time of a full pressurewaveform;  nT   is integer periods of pressure waveformswhereas t   isthe time. Apower meter (PM-1, JJ&AInstru-ments,Duvall,WA,USA)wasinsertedtomonitorelectri-cal power output and to ensure stable power output aswell as detect impedance mismatch between amplifierandHIFUtransducer.The transducer was fixedatthe bot-tom of the tank, facing upward with its focus (maximumpeak positivepressure) placed on theepicardiumofeitherthe right or left ventricle of the heart preparation throughthe acoustic window (Fig. 1a). The HIFU focus was posi-tioned at the targeted region by employing a pulse/echomethod that detected the location of epicardium/air inter-facebasedonthetimeofflightoftheechoes.Consideringacoustic attenuation (  2 dB/cm $ MHz in heart tissue[Azhari 2010]) and an acoustic travelling distance inthe heart tissue (  1 cm), the  in situ  pressures for a trav-eling wave would be  p c  5  4.4–5.5 MPa and  p r   5  –2.8to –3.3 MPa at the HIFU focus. Because the heart wasplaced sideways with one side of the surface slightlyabove the solution to allow IR imaging, the tissue/airinterface in our experimental setup may result in reflec-tion of the HIFU beam and affect the actual pressure field in situ . Thus, we estimated that the  in situ I  sppa  were1344–2010 W/cm 2 .The isolated heart was perfused and stained with avoltage-sensitive dye (VSD) di-4-ANEPPS (Invitrogen,Carlsbad, CA, USA) (10  m M) for 10 2 15 min before op-tical mapping. The particular VSD are amphipathic mol-ecules, which can bond to the membranes of myocytesand become fluorescent whereas the fluorescent inten-sities are proportional to the membrane potentials(Loew 1992). Because of their fast response to thechanges of the ambient electrical field, the APs of myo-cytes can be recorded  via  optical mapping, which usestwo green-filtered light-emittingdiodes(5 GIlluminationsystem, SciMedia, Costa Mesa, CA, USA) for excitation(  531 nm) and corresponding emission. Fluorescencesignals were filtered ( . 617 nm) and recorded using a Fig. 1. Experimental setup and image registration. (a) Schematic diagram IR camera for temperature imaging and high-speed fluorescence CMOS camera for optical mapping of electrophysiology are confocally aligned focusing at the upperepicardium of the rabbit heart preparation. The HIFU transducer at the bottom of the tissue chamber is facing upwardswith its focus at the upper myocardium within the field-of-view of the two cameras. The heart is mostly submersed in theperfusion and superfusion system with only portion of the upper epicardium above the solution to permit IR imaging. (b)Photograph of a rabbit heart preparation. (c) Corresponding IR image during HIFU ablation ( t  5 14 s). (d) Backgroundimage of optical mapping. (e) Overlaid image of projective transformed IR and optical mapping image based on physicallandmarks ( e.g.,  atrial-ventricular groove). HIFU 5 high-intensity focused ultrasound; RA 5 right atrium; LA 5 leftatrium; RV 5 right ventricle; LV 5 left ventricle; AIVS 5 anterior inter-ventricular sulcus; IR 5 infrared; PA 5 pulmo-nary artery; CMOS 5 complementary metal-oxide-semiconductor. 434 Ultrasound in Medicine and Biology Volume 41, Number 2, 2015  complementary metal-oxide-semiconductor (CMOS)camera system (MiCAM Ultima-L, SciMedia, CostaMesa, CA, USA) at 1000 frames/s with a spatial resolu-tion of 420–460  m m/pixel (100 3 100 pixels). Since therecorded optical signals are weighted summation of emitted fluorescence from cells within a tissue depthless than 200  m m below the surface for rabbit heartsdue to tissue absorption and photonic scattering(Knisley 1995), optical mapping measures the EP of cellsin a superficial layer.The upper surface of the heart preparation waselevated slightly above the solution to allow IR imaging.The IR camera (Silver SC5600, FLIR, Wilsonville, OR,USA) was confocally aligned with the CMOS camerato measure temperature at 50 frames/s and with a focalresolution of 85  m m in the same region of interest(ROI) on the epicardium (Fig. 1b) with optical mapping(Fig. 1c). The emissivity of the hearttissuewas calibratedto be 0.86 before experiments using a black tape method(Madding, 1999), which used a vinyl electrical tape(Scotch Super33 1 , 3M, Silver Spring, MD, USA) witha known emissivity of 0.95. Because of the temperaturedifference between the perfusate within the heart(37  C) and air (23  C), 2  C temperature loss because of heat convection was observed and the baseline tempera-ture of the heart was measured as 35  C by IR imaging(Fig. 2b). However, no temperature compensation ( e.g., constant 1 2  C) was performed on IR imaging data as itmight be inaccurate during HIFU heating process.The CMOS and IR cameras were synchronized withHIFU  via  a FPGA board (Cyclone II, Altera, San Jose,CA, USA) to recode optical action potentials (OAPs)and surface temperature of the isolated heart duringHIFU application. Experimental procedure During experiment, the Langendorff-perfused rabbitheart was under its natural rhythm (sinus rhythm). Tomaximize the signal-noise-ratio for optical mapping,the excitationlightwasadjustedafter 10–15minofequil-ibration, by setting the highest pixel intensity slightlyhigher than 80% of the saturation level over several car-diac cycles. Ten second tone-burst HIFU with  in situ  in-tensity varying from 1344–2010 W/cm 2 was applied tothe heart preparation. Such intensities were high enoughto ensure lesion generation or EP changes while suffi-ciently low enough to avoid lifting epicardium out of the focus of optical mapping and IR imaging because of acoustic radiation force. After experiment, the heartwasphotographed(D5000,Nikon,Tokyo,Japan),stainedwith triphenyltetrazolium (TTC) (Sigma Aldrich, St.Louis, MO, USA) (Fishbein et al. 1981) where necrotictissue was stained as white and viable tissue was indark red. Gross tissue was then stored in 10% formalin Fig. 2. Changes of OAPs baseline and temperature, and their correlation. (a) Temperature map from infrared (IR) imag-ing at t  5  14.5 s. (b) A representative temperature trace at asterisk labeled location within HIFU focus in (a) withmaximum temperature rise in the region-of-interest (ROI). (c) Corresponding OAPs trace at the same location in (a)with each ‘‘spikes’’ corresponding to single cardiac cycle. The bold line highlights the baseline of fractional fluorescencesignal ( D F  ). HIFU was on from 4–14 s (shaded area). The inset shows an enlarged version of two OAPs. (d) Maps of temperature changes ( D T  ) and corresponding OAP baseline changes ( D  B ) in the ROI in white dashed box in (a). Ratio-metric map of   D  B  /  D T   at corresponding times (bottom). HIFU was applied from 4–14 s. (e) Linear regression slope( D  B = D T  ) for paired D  B – D T   frames and adjusted  R 2 as function of time (n 5 7). HIFU is applied  in situ  during the shadedperiod. OAP  5  optical action potential; IR  5  infrared; HIFU  5  high-intensity focused ultrasound; ROI  5  region of interest. HIFU-induced electrophysiological changes and temperature increases d Z. W U  et al . 435  solution for 48 h, paraffin embedded, and sectioned at100  m m step size across the region along the transmuraldirection. Masson’s trichrome (MT) staining was con-ducted on tissue sections and the stained slides werescanned with high resolution (CanoScan 8800F, Canon,Tokyo, Japan). HIFU-induced lesions were then identi-fied from the images.  Image processing and data analysis IR images were scaled to the size of VSD images  via bilinear interpolation. Each pair of optical-IR images wasregistered using a control point registration algorithm(Matlab v. 2011b, Mathworks, Natick, MA, USA)(Fig. 1b 2 e). Multiple common identifiable feature points(8–10)onbothimageswerepickedandaprojectivetrans-formation was performed to compensate the errorinduced by the angle differences between the twocameras.Spatial averaging (3  3  3) and a zero-phase band-pass (0–100 Hz) filter were applied for spatiotemporalsmoothing of the VSD images (Laughner et al. 2012a).The effects of photo bleaching were corrected usingexponential, 2nd order polynomial, and linear fitting of the VSD baseline values before, during and after HIFUapplication, respectively. OAPs were normalized from0 to unity and APA was defined as the value betweenOAPs peaks and resting potentials. Activation time wasidentified as the corresponding times when d( D F  )/dtreached maximum within an individual cardiac cycle,where fractional fluorescence change ( D F  ) was the ratiobetween change of fluorescence signal intensity and thebackground fluorescence level. Activation maps wereformedfromtheactivationtimeateachpixelandconduc-tion velocities (CVs) were derived as the spatial gradientof the activation map. AP duration at 50% repolarization(APD 50 ) was defined as the duration between activationtime and 50% peak APAvalue at repolarization. Changesin APA ( D APA), APD 50  ( D APD 50 ) and activation times( D activation) were calculated by subtracting the pre-HIFU baseline values.To evaluate the HIFU energy deposition, thermaldose or specifically the cumulative equivalent min at43  C (denoted as  CEM  43 ) (Sapareto and Dewey 1984) was calculated as CEM  43 5 X t   final t  5 0  R ð 43 2 T  Þ D t   (2)where  T   is the average temperature during time duration D t  ,  t   final  is the final time of exposure,  R 5 0.25 when thetemperature is below 43  C and  R 5 0.5 when the temper-ature is above 43  C.The HIFU lesion, as identified from images of TTCstained tissue specimens, was denoted as positive,whereas non-lesion region was denoted as negative.Receiver-operating characteristic (ROC) analysis wasperformed on temperature and EP maps at  t   5  14 s(HIFU application was from 4–14 s) and  CEM  43  mapsat  t  5 32 s to predict isotherms of lesion, APA changes,and APD 50  changes. ROC curves were formed bycomparing true-positive rate (sensitivity) against false-positive rate (1 2 specificity) at all running thresholds,and total area under curve (AUC) was used to assessthe overall prediction performance. The optimal temper-ature threshold for generating lesion (lethal isotherm)was decided based on the best detection accuracy.Leave-one-out cross-validation was used to determinethe variation of ROC AUC and optimal temperaturethresholds.Results were expressed as mean 6 standard error of mean (SEM). One-way analysis-of-variance (ANOVA)using the Tukey-Kramer test and paired Student -t   testwere performed for multiple group comparisons, withstatistical significance defined as  p , 0.05. Linear regres-sion and  F  -test were conducted for testing linear correla-tion between parameters, and goodness-of-fit wasassessed by the adjusted  R 2 and root mean square error(RMSE) of the residuals. RESULTS Spatiotemporally correlated IR imaging and opticalmapping Figure 1b 2 e show an example of spatiotemporallyregistered IR imaging and optical mapping duringHIFU ablation of the heart preparations. The IR image(Fig. 1c) was projective transformed and registered withthe VSD image (Fig. 1d). Figure 1e shows a composite image with IR image overlaid on the VSD image forROI analysis. Effect of HIFU on the OAPs baseline During HIFU ablation, IR imaging captured thespatiotemporal changes of tissue temperature (Fig. 2a,2b), while a baseline drift of OAPs (thick line, Fig. 2c) ap-pearedtobeproportionaltothetemperature increases.The2-D maps of OAPs baseline ( D  B ) within a ROI (17 3 17pixels) surrounding the HIFU focus and the correspondingtemperature changes ( D T  ) were spatiotemporally corre-lated, especially during HIFU application (4 2 14 s)(Fig. 2d). The ratiometric maps ( D  B  /  D T  ) (bottom row inFig. 2d) at given times were fairly uniform within theROIwhen D T  wasgreaterthan5  C,indicatingalinearcor-relation between D  B  and D T  in 2-D space. Such linear cor-relationwasconfirmedthrough F  -test,andtheoveralllinearregression slope during HIFU heating was(29.91  6  0.09)  3  10 2 2 %/   C (adj.  R 2 5  0.78, 16frames 3 7 heart preparations) with  RMSE   being 0.27%. 436 Ultrasound in Medicine and Biology Volume 41, Number 2, 2015
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks