lable at ScienceDirect

Applied Thermal Engineering 62 (2014) 322e329

Contents lists avai

Applied Thermal Engineering

journal homepage: www.elsevier .com/locate/apthermeng

Acrylic acid controlled reusable temperature-sensitive hydrogelphantoms for thermal ablation therapy

Jay Shieh a,1, Shing-Ru Chen b,c,1, Gin-Shin Chen b, Chia-Wen Lo c, Chuin-Shan Chen d,Ben-Ting Chen e, Ming-Kuan Sun c, Chang-Wei Huang e,**, Wen-Shiang Chen b,c,*

aDepartment of Materials Science and Engineering, National Taiwan University, Taipei, TaiwanbDivision of Medical Engineering Research, National Health Research Institutes, Miaoli, TaiwancDepartment of Physical Medicine and Rehabilitation, National Taiwan University Hospital & College of Medicine, Taipei, TaiwandDepartment of Civil Engineering, National Taiwan University, Taipei, TaiwaneDepartment of Civil Engineering, Chung Yuan Christian University, Chung Li 32023, Taiwan

h i g h l i g h t s

� A method to fabricate a transparent reusable tissue-mimicking phantom was proposed.� The phantom changes color at a preselected threshold temperature.� The threshold temperature can be adjusted by the concentration of acrylic acid.� The phantom was designed for the real-time visualization of thermal lesions.

a r t i c l e i n f o

Article history:Received 4 April 2013Accepted 13 September 2013Available online 7 October 2013

Keywords:N-isopropylacrylamide (NIPAM)Acrylic acid (AAc)PhantomCloud pointThermal lesion

* Corresponding author. Department of Physical MNational Taiwan University Hospital, Taipei, Taiwan. Tfax: þ886 2 23832834.** Corresponding author. Department of Civil EnginUniversity, Chung Li , Taiwan. Tel.: þ886 3 2654206;

E-mail addresses: cwhuang@cycu.edu.tw (C.-W. Hu(W.-S. Chen).

1 Equal contribution.

1359-4311/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.applthermaleng.2013.09.02

a b s t r a c t

Polymerization of N-isopropylacrylamide (NIPAM) with acrylic acid (AAc) has been adopted to fabricatereusable tissue-mimicking hydrogel phantoms designed for the real-time visualization and examinationof thermal lesion formation in ablation and hyperthermia therapies. It is shown that the cloud pointtemperature of the NIPAM-based hydrogel phantoms can be adjusted by the concentration of AAc torepresent the threshold temperature of pain (42 �C) or tissue damage (52 �C). The mechanical, thermaland acoustic properties of the developed phantoms are similar to those of human soft tissues. The abilityof the phantoms to provide visualization of thermal lesions produced by either microwave or high-intensity focused ultrasound (HIFU) ablation was examined. Evolution of the optical transparency ofthe phantoms with temperature was found to be a stable hysteretic behavior and reproducible inconsecutive heatingecooling cycles, demonstrating the reusability of the phantoms. By processing theoptical images of the phantoms at different stages of the heating process, a thermal lesion can beconsidered formed (i.e., threshold temperature reached) when the grayscale value reaches the half-saturation point. The image processing method proposed for the NIPAM-based hydrogel phantoms isshown to be independent on the type of heating device used.

� 2013 Elsevier Ltd. All rights reserved.

edicine and Rehabilitation,el.: þ886 2 23123456x67087;

eering, Chung Yuan Christianfax: þ886 3 2654299.ang), wenshiang@gmail.com

All rights reserved.1

1. Introduction

Less-invasive ablative modalities using thermal energy, suchas laser ablation, focused ultrasound, microwave ablation, andradiofrequency ablation etc., have received considerable atten-tion in recent years, especially for localized tumor ablation [1e6].To extend potential applications and avoid in-vivo experimentsor human experimentations, a transparent tissue-mimickingphantom which is able to demonstrate the process and extentof thermal lesion formation in real time is critical important for

Table 1Constituents of NIPAM-based hydrogel phantoms.

NIPAM-42 NIPAM-52

Degassed water 150 ml 150 mlAAc 0.19 ml 0.44 mlNIPAM 9 g 9 gMBAm 0.375 g 0.375 gAPS 0.195 g 0.195 gTEMED 0.4 ml 0.4 ml

J. Shieh et al. / Applied Thermal Engineering 62 (2014) 322e329 323

all ablative devices during preclinical development and surgicalplanning.

Several temperature-sensitive tissue-mimicking phantomshave been reported as models for ablative therapy. For example,polyvinyl alcohol or agar-based phantoms were used to visualizethe effect of bubble-enhanced heating [7] by focused, MHz-frequency ultrasound. However, thermal lesions could not bewell visualized in such phantoms. Transparent polyacrylamide(PAA) gels containing bovine serum albumin (BSA) were thenproposed since BSA would turn white and optically opaque orcause a large reduction in the T2 signal on magnetic resonanceimaging upon reaching the threshold temperature of proteindenaturation [8,9]. Takegami et al. demonstrated a low-costversion by replacing BSA with egg white [10] for the study offocused ultrasound ablation. Although easily fabricated, themajor disadvantages associated with egg white or albumin-basedtissue-mimicking phantoms are the irreversible protein dena-turation and permanent color change above the threshold tem-perature, making them impossible to be reused. Moreover,precise identification and adjustment of the threshold tempera-ture of the phantoms are difficult to achieve.

To solve the above-mentioned disadvantages, new tempera-ture indicators were adopted to replace BSA for the constructionof tissue-mimicking phantoms. For example, a nonionic surface-active agent (NiSAA) which exhibited hydrophobic segregation attemperatures above the so called “cloud point” temperature wasproposed [11]. The cloud point, also known as the lower criticalsolution temperature (LCST), represents the temperature whenmacromolecules transform from a hydrophilic structure (belowthe cloud point) to a hydrophobic structure (above the cloudpoint) [12]. For example, polyacrylamide hydrogels containingNiSAA were shown to become opaque when heating above thecloud point, but gradually return to transparent upon cooling[11]. The cloud point of the polyacrylamide hydrogels could bealtered by the choice of NiSAA type and further finely adjustedwith the addition of methanol or butanol, or by changing the pHvalue [11].

Similar temperature-sensitive properties to NiSAA can also befound in N-isopropylacrylamide (NIPAM) and PolyNIPAM co-polymers which exhibit cloud points near physiological relevanttemperatures [12,13]. Existing studies have shown that heating anaqueous NIPAM solution past the cloud point would lead to thesegregation of NIPAM, resulting in the increase of opaqueness of thesolution [12,14]. Moreover, the acoustic properties of NIPAM orpolyNIPAM copolymer gels are similar to those of human tissue[15], and their temperature and optical properties, including LCSTand transparency, could be easily manipulated by the addition ofacrylic acid (AAc) [13,16].

This study reports a method based on polymerization ofNIPAM in the presence of AAc for the fabrication of reusabletissue-mimicking hydrogel phantoms, which are designed for thereal-time visualization and examination of thermal lesion for-mation in microwave and high-intensity focused ultrasound(HIFU) ablation therapies. In order to understand the similarity(or dissimilarity) between the phantoms developed and humantissue, key mechanical, thermal, acoustic and optical propertiesof the phantoms must also be characterized. Two types ofNIPAM-based reusable hydrogel phantoms with cloud pointtemperatures at 42 and 52 �C were prepared for the microwaveand HIFU ablation experiments; they are denoted as “NIPAM-42”and “NIPAM-52” phantoms, respectively, in this study. The cloudpoint was adjusted by the concentration of AAc and the twochosen threshold temperatures, 42 and 52 �C, represent thethresholds of pain and tissue damage [17], respectively. Pain istypically experienced at temperatures above 42 �C, while

irreversible tissue damage may occur above 52 �C. The NIPAM-based phantom systems demonstrated in this study are there-fore useful for the rapid characterization and calibration of anablative device/treatment to determine its efficacy and safetybefore animal or clinical studies.

2. Materials and methods

2.1. Preparation of hydrogel phantoms

The NIPAM-based reusable hydrogel phantoms were formed bycrosslinking copolymerization of NIPAM and N,N0-methyl-enebisacrylamide (MBAm) with the addition of AAc to adjust thecloud point so that it fell in the temperature range of biologicalsignificance. The constituents of the phantoms and their amountsare listed in Table 1. The fabrication process consisted of thefollowing steps: AAc (purity 99.5%) with an amount of either 0.19 or0.44 ml was dissolved in 150 ml degassed, distilled water first ethese two specific amounts of AAc gave rise to the cloud points of42 and 52 �C, respectively. Next, 9 g of NIPAM was added to theaqueous solution, which was gently stirred at room temperatureuntil complete dissolution of NIPAM. 0.375 g of MBAm (purity 97%)and 0.195 g of ammonium persulfate (APS), which acted as theinitiator for crosslinking, were then added consecutively into theaqueous solution. The mixture was gently stirred at room temper-ature until homogenized. Finally, 0.4 ml of polymerization agentN,N,N0,N0-tetramethyl-ethylene (TEMED; purity 99%) was added tothe mixture. The final aqueous mixture was immediately pouredinto molding containers and allowed to polymerize completely atroom temperature to form hydrogel phantoms. The hydrogelphantoms prepared were either tested in the ablation experimentswithin 24 h after complete polymerization, or stored in an airtightcontainer to avoid dehydration (if left in air) or swelling (if placed inwater) for later experimental usage. The NIPAM-based hydrogelphantoms were optically transparent, gelatin-like materials.

2.2. Measurement of density and elastic modulus

For the measurements of mechanical properties, the hydrogelphantoms were cut into cuboidal specimens measuring5 � 5 � 5 cm3. The apparent densities of the specimens weredetermined by the standard Archimedes’ immersion technique, i.e.,dividing the weight of the specimen (measured by an electronicbalance) by its volume (equaled to the volume of water displaced)when the specimen was suspended and completely immersed in awater displacement tank [18,19]. For the determination of elasticmodulus, the cuboidal specimen was mounted within an electro-mechanical load frame (Model 42) and compressed slowly by a flat-surface loading platen at a crosshead speed of 2.54 mm/min until25% deformation in the loading direction was reached. The attachedload cell had a resolution of 2.5 N. Strain was measured by thecrossheadmovementwith a position resolution of 5�10�5 mm. Theelastic modulus of the specimenwas then extracted from the initial,linear portion of the stressestrain curve obtained during the loading

Fig. 1. Schematic of the experimental heating apparatus: (a) microwave ablation and(b) HIFU ablation systems.

J. Shieh et al. / Applied Thermal Engineering 62 (2014) 322e329324

process. For each type of hydrogel phantom, the average of threeseparate measurements was used to represent its elastic modulus.

2.3. Measurement of thermal conductivity and specific heat

The thermal conductivities of the hydrogel phantoms (usingspecimens of size 6 � 6 � 7 cm3) were measured by a thermalconductivity analyzer (TPS 2500 S) with a measuring range of0.005e500W/m K and an accuracy within 2% or better. The specificheats of the hydrogel phantoms on the other hand were measuredby a differential scanning calorimeter (DSC; Q200) from 20 to 60 �Cwith a heating rate of 10 �C/min. The heat capacity was calculatedby dividing the heat absorbed by the phantom specimen by theresulting temperature increase, calibrated against a referencealuminum pan [20]. The DSCmeasurement was repeated five timesfor both NIPAM-42 and NIPAM-52 phantoms using small speci-mens with average weights of 8.5 and 5.9 g, respectively, to ensureuniform heating.

2.4. Measurement of acoustic properties

The speed of sound in the hydrogel phantomwas determined bymeasuring the time of flight of a single broadband ultrasound pulse(A392S) to traverse a certain length of the phantom. The acousticimpedance of the phantom was then determined by multiplyingthe density of the phantom and the speed of sound in the phantom.The acoustic attenuation coefficient of the phantom at 1 MHz(A392S) was obtained by comparing the logarithm of the spectralmagnitude of the ultrasound signals transmitted through 60 mm

long (thick) phantom specimen and water path [21]. All measure-ments of acoustic properties were conducted at room temperature.

2.5. Measurement of optical transparency

The NIPAM-42 and NIPAM-52 hydrogel phantoms were heatedabove their cloud points to examine the evolution of opticaltransparency with increasing temperature. The schematic of theheating apparatus is shown in Fig. 1(a). The cuboidal phantomspecimen measuring 6 � 6 � 7 cm3 was fixed in a lucite tank filledwith degassed, deionized water. A 0.5-mm diameter heating an-tenna of UMC-1 microwave ablation system was inserted into thespecimen and adopted as the heating source. Continuous micro-wave irradiation at 2450 MHz with a power of 35 W was suppliedby the microwave antenna. A thermocouple was also inserted intothe specimen, placed at 5 mm away from themicrowave antenna tomeasure temperature changes around the heating source. In orderto determine the threshold temperature for thermal lesion for-mation, the phantom specimen was heated until reaching opacitysaturation. Changes in temperature and specimen color throughoutthe heating process were monitored, respectively, by the insertedthermocouple and a high-definition video (HDV) camera (HDR-XR150) at 30 frames per second. For each image frame recorded, a5 � 5 pixel region of interest (ROI), which situated between thethermocouple tip and the microwave antenna and right next to thethermocouple tip, was selected for image processing. The averagegrayscale pixel value of the ROI at a particular temperature wascalculated as a measure of the optical transparency of the phantomspecimen at that temperature. The image frames recorded for theheating process were processed off-line by a custom MATLABprogram.

2.6. HIFU ablation

The formation of thermal lesions in the NIPAM-42 and NIPAM-52 hydrogel phantoms by HIFU ablation was investigated. Theexperimental HIFU system consisted of a function generator (Model33250), a radio-frequency (RF) power amplifier (1040L), a powermeter (Model 4025), and a 2 MHz focused piezoelectric transducer(SU-101). The schematic of the HIFU system is shown in Fig. 1(b).The HIFU transducer was made of a single element concavepiezoelectric ceramic with an aperture diameter of 35 mm and aradius of curvature of 55 mm. The ultrasound produced by thetransducer was aimed at the phantom specimen (5 � 5 � 5 cm3)fixed in a lucite tank filled with degassed, deionized water. Thefocal spot of the ultrasound was positioned inside the specimen.The spatial position of the transducer was controlled by a dual axisstepper motor to an accuracy of 0.1 mm. For the HIFU ablationexperiment, the transducer was operated in a continuous-wavemode for 30 s at an electrical power of 33 W (or 8.1 MPa peak-to-peak acoustic pressure), calibrated by a needle hydrophone(SPEH-S-0500). Changes in specimen color during the ablationexperiment were monitored by the HDV camera at 30 frames persecond. The average grayscale pixel values of a ROI in the imageframes were calculated to reveal the evolution of the opticaltransparency of the phantom specimen during HIFU ablation. TheROI had a size of 5� 5 pixels and as shown in Fig.1(b), situated rightat the ultrasound focal spot.

2.7. Statistical analysis

All material property data measured in this study are presentedas mean � standard deviation. The differences in the measuredvalues were analyzed using the Student’s t-test and a p-value <

0.05 was considered statistically significant.

Table 2Densities and elastic moduli of NIPAM-based hydrogel phantoms and selected hu-man tissues [15,21].

Density (g/cm3) Elastic modulus (kPa)

Brain 0.994 0.1e1Muscle 1.08 8e17Bone 1.912 50NIPAM-42 1.16 � 0.002 9.96 � 0.024NIPAM-52 1.19 � 0.007 13.63 � 0.767

Table 4Acoustic properties of NIPAM, egg white and BSA-based hydrogel phantoms, water,and selected human tissues [21].

Soundspeed (m/s)

Attenuationcoefficient (dB/cm MHz)

Acoustic impedance(MRayl)

Water 1480 0.0022 1.48NIPAM-42 1505 0.50 � 0.004 1.75NIPAM-52 1512 0.54 � 0.014 1.81Liver 1550 0.89 1.64Brain 994 0.85 1.55BSA 1540 0.17 1.6Egg white 1540 0.2 1.54

J. Shieh et al. / Applied Thermal Engineering 62 (2014) 322e329 325

3. Results

In order to evaluate the similarity (or dissimilarity) between theNIPAM-42 and NIPAM-52 hydrogel phantoms and human tissue,several important properties of the phantoms were characterized.A comparison of the densities and elastic moduli of the developedphantoms and several types of human tissues is given in Table 2.The densities and elastic moduli measured for the NIPAM-42 andNIPAM-52 hydrogel phantoms were 1.16 � 0.002 and1.19 � 0.007 g/cm3 and 9.96 � 0.024 and 13.63 � 0.767 kPa,respectively. These values are similar to those of human muscletissue (1.08 g/cm3 and 8e17 kPa) [15]. The thermal and acousticproperties of the NIPAM-42 and NIPAM-52 hydrogel phantoms arelisted in Tables 3 and 4, respectively. For comparison purpose, thesame properties of the BSA and egg white-based hydrogel phan-toms and of several types of pig or human tissues are also given inthe tables. In terms of acoustic attenuation, which is a key propertyfor HIFU propagation and heating, the acoustic attenuation co-efficients of the NIPAM-based phantoms are closer to, although stilllower than, those of human tissues (e.g., brain or liver) whencompared to those of the BSA and egg white-based phantoms. Interms of specific heat, the proposed NIPAM-based phantoms aresuperior since there is no statistical difference between humanliver tissue and the NIPAM-42 or NIPAM-52 phantom. The BSAphantoms show a significantly smaller specific heat than humanliver tissue, while, the reported specific heat of the egg white-basedphantom [22] is significantly larger than those of the NIPAM-42 andNIPAM-52 phantoms.

Fig. 2 shows the evolution of the optical transparencies of theNIPAM-42 and NIPAM-52 hydrogel phantoms when being heatedup by microwave ablation from room temperature (w25 �C) toabout 68 �C and then cooled naturally back down to room tem-perature. The NIPAM-based hydrogel phantoms were homoge-neous and optically transparent at room temperature. As shown inFig. 2(a) and (b), the average grayscale pixel values (i.e., adopted torepresent transparency) of both types of phantoms increasedrapidly as the temperature reached and increased above their

Table 3Thermal properties of NIPAM, egg white and BSA-based hydrogel phantoms, water,and selected pig and human tissues.

Thermal conductivity(W/m K)

Specific heat (J/g K)

Pig liver [24] 0.507 e

Pig meat [24] 0.464 e

Human liver [25] 0.572 � 0.009 3.628 � 0.078Brain [26] e 3.86Muscle [26] e 3.94Water 0.547 4.18NIPAM-42 0.60 � 0.02 3.365 � 0.160NIPAM-52 0.60 � 0.03 3.431 � 0.233Egg white [22] 0.59 � 0.06 4.27 � 0.365BSA [27] 0.473 � 0.030a 2.887 � 0.152a

BSA with glass beads [27] 0.654 � 0.033 3.597 � 0.540

a Significant difference from human liver.

Fig. 2. Evolution of optical transparency (represented by grayscale) of (a) NIPAM-42and (b) NIPAM-52 hydrogel phantoms during a microwave heating cycle from roomtemperature to about 68 �C and then cooled back down to room temperature. Evo-lution of optical transparency of NIPAM-42 hydrogel phantom during three microwaveheatingecooling cycles is shown in (c).

Fig. 3. Evolution of optical transparency (represented by grayscale) of (a) NIPAM-42and (b) NIPAM-52 hydrogel phantoms during three microwave heatingecooling cy-cles. For clarity purpose, (a) and (b) only show grayscale values of the heating part ofthe cycles.

Table 5Characteristic temperature parameters and half-saturation grayscale values (Y0) ofNIPAM-based hydrogel phantoms. T0 is corresponding to Y0 and represents thethreshold temperature (cloud point) of phantom.

Parameter NIPAM-42 NIPAM-52

Tmax (�C) 36 � 0.5 49.3 � 0.3T0 (�C) 42.4 � 0.5 52.6 � 0.3Tmin (�C) 47.2 � 0.5 55.8 � 0.3TR (s) 19.60 � 2.43 27.40 � 1.41Y0 (grayscale) 155.57 � 1.27 156.61 � 1.92

J. Shieh et al. / Applied Thermal Engineering 62 (2014) 322e329326

respective cloud points. The grayscale values then plateaued afteropacity saturation was reached (above 61 �C). The process ofchanging transparency was reversible but showed a marked hys-teresis during heatingecooling cycle. The reproducibility of thechange in transparency for the NIPAM-42 hydrogel phantom forthree heatingecooling cycles is shown in Figs. 2(c) and 3 (note:

Fig. 4. Optical transparency (represented by grayscale) of NIPAM-42 hydrogel phan-tom as a sigmoid function of temperature (or heating time). Second derivative of thissigmoid function is also shown.

Fig. 3 only shows the heating part of the hysteresis curves). Thestable cyclic behavior indicated a high degree of reusability.

Fig. 4 shows the optical transparency (in grayscale value of 0e255) of the NIPAM-42 hydrogel phantom as a sigmoid function oftemperature (or heating time). The sigmoid relationship could bedescribed as:

YðtÞ ¼ Bþ M

1þ exp�t0�tTR

� (1)

where Y is the optical transparency in grayscale value, B is thebaseline value of Y, t is the heating time, t0 is the heating time at themidpoint of grayscale change, TR is a time constant related to therate of grayscale change, and M is the maximum increment ofgrayscale value from the baseline value B. A larger TR gives rise to aslower rate (i.e., a longer time) to reach the maximum grayscalevalue BþM. When t¼ 0 (at the beginning of heating), the grayscaleis at its baseline value, i.e., Y(0) ¼ B since t0 >> TR. When t is muchlarger than t0 (at the end of heating), Y(t) ¼ B þ M. When t ¼ t0,Y(t0) ¼ B þ M/2; this particular Y value is denoted as Y0.

By fitting the experimental grayscale data with Eq. (1) andcalculating the second derivative of Eq. (1), which is expressed as:

Fig. 5. Series of optical images of thermal lesions formed within (a) NIPAM-42 and (b)NIPAM-52 hydrogel phantoms at different time points of microwave ablation experi-ment (microwave on at t ¼ 180 s and off at t ¼ 480 s). The left and right white lines inimages are the inserted thermocouple and microwave heater, respectively.

J. Shieh et al. / Applied Thermal Engineering 62 (2014) 322e329 327

Y 00ðtÞ ¼ MT2R

8>><>>:

2hexp

�t0�tTR

�i2h1þ exp

�t0�tTR

�i3 �exp

�t0�tTR

�

h1þ exp

�t0�tTR

�i2

9>>=>>;

(2)

three critical Y values: Ymax, Y0, and Ymin, which correspond to themaximum, zero, and minimum of Y00(t), respectively, can bedetermined (see Fig. 4). Temperatures correspond to these threecritical Y values are Tmax, T0, and Tmin, respectively. Upon heating,the NIPAM-42 hydrogel phantom started to change color at t ¼ tsand reached opacity maximum (i.e., the end of color change) att ¼ te. The corresponding temperatures and grayscale values for tsand te were Ts and Te and Ys and Ye, respectively. Fig. 4 shows that forthe heating part of a temperature cycle, Ts < Tmax < T0 < Tmin < Teand Ys < Ymax < Y0 < Ymin < Ye. Experimentally, Ts and Te, andtherefore Ys and Ye were difficult to identify immediately. Incontrast, Tmax, T0, and Tmin could be accurately determined from theanalysis of second derivative, and those values of the NIPAM-42 andNIPAM-52 hydrogel phantoms are listed in Table 5. Results from theanalytical calculations indicate that T0, which is the temperaturecorresponding to the half-saturation point of the grayscale curve,represents the threshold temperature (cloud point) of the phan-tom, i.e., 42 �C for NIPAM-42 and 52 �C for NIPAM-52. For the imageframes recorded during the heating process, a thermal lesion wasconsidered to be formed when the grayscale value of the ROIreached the half-saturation point.

Fig. 5 shows series of optical images of the thermal lesionsformed within the NIPAM-42 and NIPAM-52 hydrogel phantomsat different time points of the microwave ablation experiment.The experiment ran for 1000 s with the microwave systemturned on at t ¼ 180 s and off at t ¼ 480 s. The images were takenat 50 s interval from 180 to 630 s. As shown in Fig. 5, the oval-shaped thermal lesions appeared and grew with time duringthe process of microwave ablation, and then gradually dis-appeared after halting the microwave system. No permanentopaque regions were detected within the phantoms after theheatingecooling cycle. For both types of hydrogel phantoms, thesize and whiteness of the thermal lesions increased with timeduring the ablation process, and the largest and whitest lesionswere observed at t ¼ 480 s (i.e., the end of the ablation). Since theNIPAM-52 phantom contained a higher amount of AAc andexhibited a higher could point, the lesion size at 480 s was whiter(due to more prominent NIPAM segregation) but smaller in theNIPAM-52 phantom than in the NIPAM-42 phantom. Further-more, the decrease in opacity during the cooling phase (i.e., after480 s) was slower in the NIPAM-52 phantom.

Fig. 6 shows series of optical images of the thermal lesionsformed within the NIPAM-42 and NIPAM-52 hydrogel phantoms atdifferent time points of the HIFU ablation experiment. The exper-iment ran for 55 s with the HIFU system turned on at t ¼ 0 s and offat t ¼ 30 s. The images were taken at 5 s interval from 0 to 55 s. Asshown in Fig. 6, the needle-shaped thermal lesions appeared andgrew with time during the process of HIFU ablation, and thengradually disappeared after the HIFU system was turned off,completing a heatingecooling cycle without any residual lesions inthe phantoms. For both types of hydrogel phantoms, the largest andwhitest lesions were observed at t ¼ 30 s (i.e., the end of the

Fig. 6. Series of optical images of thermal lesions formed within (a) NIPAM-42 and (b)NIPAM-52 hydrogel phantoms at different time points of HIFU ablation experiment(HIFU on at t ¼ 0 s and off at t ¼ 30 s). Matching of Y0 ¼ 156 contour of NIPAM-52phantom (solid red line; obtained by data fitting) with actual clouding region (i.e.,thermal lesion) of the phantom after being heated by HIFU for 30 s is shown in (c). (Forinterpretation of the references to color in this figure legend, the reader is referred tothe web version of this article.)

J. Shieh et al. / Applied Thermal Engineering 62 (2014) 322e329328

ablation). The lesion size at 30 s was smaller in the NIPAM-52phantom than in the NIPAM-42 phantom. The Y0 ¼ 156 contourfor the NIPAM-52 phantom (obtained by fitting the microwaveablation data with Eqs. (1) and (2)) was used to match the actualclouding region (i.e., thermal lesion) of the phantom after beingheated by HIFU for 30 s. The result is shown in Fig. 6(c) e a good fitbetween the predicted contour (solid red line) and the actualclouding region is evident.

4. Discussion

This study describes the fabrication of two types of NIPAM-basedreusable tissue-mimicking hydrogel phantoms, which are designedfor the real-time visualization and examination of thermal lesionformationat specific threshold temperatures in ablative therapies. Incomparison to the existing tissue-mimicking phantoms containingtemperature-sensitive proteins (e.g., BSA), the proposed NIPAM-based phantoms have three major advantages: (1) the phantomsexhibit a high degree of optical transparency at room temperature,allowing the real-time observation of tiny opacity changes due toincreased temperature; (2) the reversible nature of the phantoms interms of thermal lesion formation enables them to be repeatedlyused in heatingecooling cycles; and (3) the threshold temperatureabove which the thermal lesions are formed, i.e., the cloud point ofthe phantom, canbe adjusted. TheNIPAM-basedhydrogel phantomsmimic the properties of human soft tissue (see Tables 2e4) andbecome opaque at temperatures above the cloud point due to thesegregation of NIPAM. Upon cooling below the cloud point, theNIPAM-based phantoms restore their optical transparency. Theprocess of changing transparency, although reversible, is thermallyhysteretic. The variation of opacity in consecutive heatingecoolingcycles is reproducible (see Figs. 2(c) and 3), demonstrating thereusability of the NIPAM-based hydrogel phantoms.

Fig. 3(b) shows that the room-temperature baseline opacity(grayscale value) of the NIPAM-52 phantom increases gradually asthe number of temperature cycle increases. However, suchbehavior is not observed for the NIPAM-42 phantom. The NIPAM-52 phantom contains a higher amount of AAc, leading to a moreprominent NIPAM segregation (i.e., whiter lesions) above the Cloudpoint. It is therefore believed that when a high degree of polymericsegregation takes place, a portion of the energy loss associated withthe thermal hysteresis is spent on the irreversible structural changeof the hydrogel phantom. This could explain the gradual increase inbaseline opacity of the NIPAM-52 phantom after each temperaturecycle.

The NIPAM-based phantoms developed in this study are shownto change color at the threshold temperature T0, which is thetemperature when the second derivative of the grayscale functionequals zero (see Fig. 4 and Table 5). Nevertheless, the visual abilityof human beings is quite different from that of a HDV camera. Animportant questionwould therefore be whether the same T0 can beused to represent the temperature of lesion formation from theview of human eyes. Park et al. [11] have proposed a visual methodbased on dividing the grayscale, with values between 0 (0% opacity)and 255 (100% opacity), into ten consecutive levels and comparingthem with the phantom images taken during the temperature cy-cle. According to their study, the grayscale difference noticeable tothe naked eyes, defined as the difference threshold or “justnoticeable difference (JND)”, is around levels 7e8 (grayscale valuesof 119.5e160.2) [11]. In this study, the Y0 values of the NIPAM-42and NIPAM-52 hydrogel phantoms fall inside the JND, and thus,their respective T0 values can be used to represent their thresholdtemperatures for thermal lesion formation.

The cloud point of the NIPAM-based phantom system developedin this study can be easily adjusted by the amount of AAcwithin the

phantom. This makes it highly flexible and useful in temperaturemonitoring applications. The cloud points of the NIPAM-42 andNIPAM-52 phantoms are designed to be at 42 and 52 �C, respec-tively; they represent the threshold of pain sensation before tissueinjury and the threshold of permanent tissue damage (i.e., onset ofprotein denaturation), respectively. In comparison, BSA-basedhydrogel phantoms typically undergo denaturation at about 58 �C[8]; while, the threshold temperature for thermal destruction ofhuman tissue is about 60 �C e irreversible tissue necrosis would beinduced when exceeding this limit [23]. If a higher temperaturemonitoring range is required by an ablative treatment, NIPAM-based hydrogel phantoms with higher cloud points could befabricated by adjusting the AAc amount.

Y0 values obtained from fitting the microwave ablation data ofthe NIPAM-52 phantomwith Eqs. (1) and (2) are adopted to predictthe size of the thermal lesion created within the same phantom atdifferent stages of HIFU ablation. For example, a good fit betweenthe Y0 ¼ 156 contour and the actual clouding area (i.e., thermallesion) after a 30-s HIFU application is achieved (see Fig. 6(c)). Thisshows the independence of the method on the type of heating de-vice used. Nevertheless, the predicted contour in Fig. 6(c) is slightlysmaller e the difference between the contour and actual cloudingarea is less than 2mm in both the vertical and horizontal axes. Suchsmall discrepancy would be undetectable by human eyes.

5. Conclusions

A method based on polymerization of NIPAM in the presence ofAAc for the fabrication of reusable tissue-mimicking hydrogelphantoms has been proposed. These phantoms are designed for thereal-time visualization and examination of thermal lesion forma-tion in ablation and hyperthermia therapies. The cloud pointtemperature of the NIPAM-based hydrogel phantoms can beadjusted by the concentration of AAc to represent the threshold ofpain (42 �C) or tissue damage (52 �C). The mechanical, thermal andacoustic properties of the phantoms are similar to those of humansoft tissues. The optical transparency of the phantoms decreasesrapidly as the temperature reaches and increases above the cloudpoint, but returns upon cooling. The process of changing trans-parency with temperature was found to be a stable hystereticbehavior and reproducible in consecutive heatingecooling cycles,demonstrating the reusability of the phantoms. By processing theoptical images of the phantoms at different stages of the heatingprocess, a thermal lesion can be considered formed (i.e., thresholdtemperature reached) when the grayscale value reaches the half-saturation point. The image processing method proposed for thephantoms is shown to be independent on the type of heating deviceused. The transparent NIPAM-based hydrogel phantoms developedin this study can serve a critical role in accessing ablative devicesduring their preclinical development and surgical planning.

Acknowledgements

The authors acknowledge the financial support of the NationalHealth Research Institutes (ME-101-PP-03), the Ministry of Eco-nomic Affairs (99-EC-17-A-19- S1-140), and the National ScienceCouncil (100-2923-B-002-003-MY3 or SonInCaRe) of the Republicof China (Taiwan).

References

[1] D.M. Agnese, W.E. Burak Jr., Ablative approaches to the minimally invasivetreatment of breast cancer, Cancer J. 11 (2005) 77e82.

[2] M. Ahmed, S.N. Goldberg, Thermal ablation therapy for hepatocellular carci-noma, J. Vasc. Interv. Radiol. 13 (2002) S231eS244.

J. Shieh et al. / Applied Thermal Engineering 62 (2014) 322e329 329

[3] H.W. Head, G.D. Dodd 3rd, Thermal ablation for hepatocellular carcinoma,Gastroenterology 127 (2004) S167eS178.

[4] F. Izzo, Other thermal ablation techniques: microwave and interstitial laserablation of liver tumors, Ann. Surg. Oncol. 10 (2003) 491e497.

[5] W.Y. Lau, T.W. Leung, S.C. Yu, S.K. Ho, Percutaneous local ablative therapy forhepatocellular carcinoma: a review and look into the future, Ann. Surg. 237(2003) 171e179.

[6] A. Veltri, P. Moretto, A. Doriguzzi, E. Pagano, G. Carrara, G. Gandini, Radio-frequency thermal ablation (RFA) after transarterial chemoembolization(TACE) as a combined therapy for unresectable non-early hepatocellularcarcinoma (HCC), Eur. Radiol. 16 (2006) 661e669.

[7] R.G. Holt, R.A. Roy, Measurements of bubble-enhanced heating from focused,MHz-frequency ultrasound in a tissue-mimicking material, Ultrasound Med.Biol. 27 (2001) 1399e1412.

[8] C. Lafon, V. Zderic, M.L. Noble, J.C. Yuen, P.J. Kaczkowski, O.A. Sapozhnikov,F. Chavrier, L.A. Crum, S. Vaezy, Gel phantom for use in high-intensity focusedultrasound dosimetry, Ultrasound Med. Biol. 31 (2005) 1383e1389.

[9] L.S. Bouchard, M.J. Bronskill, Magnetic resonance imaging of thermal coagu-lation effects in a phantom for calibrating thermal therapy devices, Med. Phys.27 (2000) 1141e1145.

[10] K. Takegami, Y. Kaneko, T. Watanabe, T. Maruyama, Y. Matsumoto, H. Nagawa,Polyacrylamide gel containing egg white as new model for irradiation ex-periments using focused ultrasound, Ultrasound Med. Biol. 30 (2004) 1419e1422.

[11] S.K. Park, S.R.A.R. Guntur, K. Il Lee, D.G. Paeng, M.J. Choi, Reusable ultrasonictissue mimicking hydrogels containing nonionic surface-active agents forvisualizing thermal lesions, IEEE Trans. Biomed. Eng. 57 (2010) 194e202.

[12] H.G. Schild, Poly (N-isopropylacrylamide) e experiment, theory and applica-tion, Prog. Polym. Sci. 17 (1992) 163e249.

[13] B. Vernon, S.W. Kim, Y.H. Bae, Thermoreversible copolymer gels for extra-cellular matrix, J. Biomed. Mater. Res. 51 (2000) 69e79.

[14] J.S. Scarpa, D.D. Mueller, I.M. Klotz, Slow hydrogenedeuterium exchange in anon-alpha-helical polyamide, J. Am. Chem. Soc. 89 (1967) 6024e6030.

[15] A.J. Engler, S. Sen, H.L. Sweeney, D.E. Discher, Matrix elasticity directs stem celllineage specification, Cell 126 (2006) 677e689.

[16] Y.Z. Xiao, The Application of Thermo-sensitive Hydrogels on UltrasoundThermal Therapy. Mechanical Engineering, Master, Tatung University, Taipei,2007, p. 50.

[17] J. Heisterkamp, R. van Hillegersberg, J.N. I.J, Critical temperature and heatingtime for coagulation damage: implications for interstitial laser coagulation(ILC) of tumors, Lasers Surg. Med. 25 (1999) 257e262.

[18] D. Halliday, R. Resnick, J. Walker, Fundamentals of Physics Extended, Wiley &Sons, New York, 1988, p. 371.

[19] W.J. Seong, U.K. Kim, J.Q. Swift, Y.C. Heo, J.S. Hodges, C.C. Ko, Elastic propertiesand apparent density of human edentulous maxilla and mandible, Int. J. OralMaxillofac. Surg. 38 (2009) 1088e1093.

[20] B. Wunderlich, Thermal Analysis, Academic Press, Boston, 1990.[21] W.R. Hedrick, D.L. Hykes, D.E. Starchman, Ultrasound Physics and Instru-

mentation, Mosby, St. Louis, USA, 1995, pp. 1e30.[22] G.W. Divkovic, M. Liebler, K. Braun, T. Dreyer, P.E. Huber, J.W. Jenne, Thermal

properties and changes of acoustic parameters in an egg white phantomduring heating and coagulation by high intensity focused ultrasound, Ultra-sound Med. Biol. 33 (2007) 981e986.

[23] C.R. Hill, G.R. terHaar, High intensity focused ultrasound-potential for cancertreatment, Br. J. Radiol. 68 (1995) 1296e1303.

[24] X.G. Liang, X.S. Ge, Y.P. Zhang, G.J. Wang, A convenient method of measuringthe thermal conductivity of biological tissue, Phys. Med. Biol. 36 (1991) 1599e1605.

[25] K. Giering, I. Lamprecht, O. Minet, A. Handke, Determination of the specific-heat capacity of healthy and tumorous human tissue, Thermochim. Acta251 (1995) 199e205.

[26] I.P. Herman, Physics of the Human Body, Springer, Berlin, New York, 2007.[27] M.J. Choi, S.R. Guntur, K.L. Lee, D.G. Paeng, A. Coleman, A tissue mimicking

polyacrylamide hydrogel phantom for visualizing thermal lesions generatedby high intensity focused ultrasound, Ultrasound Med. Biol. 39 (2013) 439e448.