A

BhMpbRosdC

K

Hbssmmsdcwthnctb

2

0d

Clinical surgery–International

Heat fixation of cancer cells ablated with high-intensity–focusedultrasound in patients with breast cancer

Feng Wu, M.D., Ph.D.a,*, Zhi-Biao Wang, M.D., Ph.D.b, You-De Cao, Ph.D.b,Zhong-Lin Xu, M.D.b, Qiang Zhou, M.D.b, Hui Zhu, M.D.b, Wen-Zhi Chen, M.D.b

aHIFU Unit, The Churchill Hospital, University of Oxford, Headington, Oxford OX3 7LJ, United KingdombClinical Center for Tumor Therapy of 2nd Affiliated Hospital and Institute of Ultrasonic Engineering in Medicine,

Chongqing University of Medical Sciences, Chongqing, China

Manuscript received July 29, 2005; revised manuscript March 22, 2006

bstract

ackground: High-intensity–focused ultrasound (HIFU) is a noninvasive thermal ablation technique. This study reports the use ofistological techniques for the pathological assessment of HIFU effects in patients with breast cancer.ethods: Twenty-three patients with biopsy-proven breast cancer underwent HIFU treatment for primary breast lesion. Mastectomy was

erformed on all patients after HIFU. By using histological examinations, the surgical specimens were assessed to explore HIFU effects onreast cancer.esults: Coagulation necrosis of targeted tumors was confirmed by microscopy in 23 patients. Tumor cells presented typical characteristicsf coagulation necrosis in the peripheral region of the ablated tumor in all patients. However, in 11 of 23 patients, hematoxylin and eosintaining showed normal cellular structure in the central ablated tumor. By using electronic microscopy and nicotinamide adenineinucleotide-diaphorase stain, those who had normal-appearing cancer cells were not viable.onclusions: HIFU can cause the heat fixation of ablated tumor through thermal effect. © 2006 Excerpta Medica Inc. All rights reserved.

The American Journal of Surgery 192 (2006) 179–184

eywords: High-intensity–focused ultrasound; Focused ultrasound surgery; Breast cancer; Thermal ablation; Ultrasound; Therapy; Breast; Neoplasm

smimettts

M

P

tos

igh-intensity–focused ultrasound (HIFU) has recentlyeen used as a noninvasive therapy to treat patients witholid malignancy [1–4]. No therapeutic instrument is in-erted into a targeted lesion during the procedure; thisakes HIFU potentially more attractive than other mini-ally invasive therapies for the local treatment of tumor,

uch as radiofrequency, laser, and microwave. In previousecades, animal studies showed that HIFU could induceoagulation necrosis in vivo of a targeted volume at depth,ithout damaging overlying tissue [5–10]. Three clinical

rials designed for assessing the histopathologic changesave revealed obvious coagulation necrosis of solid malig-ancy treated with HIFU [11–13]. However, light micros-opy alone may not always show coagulation necrosis of thereated tumor clearly [14,15], and the results may be limitedy means of using conventional histological examination,

* Corresponding author. Tel.: �44-1865-763-100; fax: �44-1865-751-22.

sE-mail address: mfengwu@yahoo.com

002-9610/06/$ – see front matter © 2006 Excerpta Medica Inc. All rights reservoi:10.1016/j.amjsurg.2006.03.014

uch as hematoxylin and eosin (H&E) stain, for the assess-ent of the tumor destruction. Therefore, several histolog-

cal examinations, including H&E, histochemical, and im-unohistochemical stains, were performed in this study to

valuate therapeutic effects of HIFU on both cellular struc-ures and viability of human breast cancer. The purpose ofhis study was to explore reliable histological methods forhe assessment of HIFU effect on the tumor in patients witholid malignancy.

ethods

atients

A total of 23 women with breast cancer were enrolled inhis study, which was approved by the ethics committee atur university. At the time of enrollment, each patientigned an informed consent form, in accordance with the

pecification stipulated by the Helsinki Committee.

ed.

ptufnaol

rsHatt

H

[spbc1amm

e(psdsuagat

psWftstt

H

mttiwtsc

paFdwe(

�apgwMawoeEmi

pdHht(pCtCt

TC

N

ATH

N

T

180 F. Wu et al. / The American Journal of Surgery 192 (2006) 179–184

The selection criteria were as follows: histologicallyroven breast cancer (T1-2, N0-2, M0), single palpable mass,umor size �6 cm, the lesion boundaries visualized withltrasound imaging, and tumors separated by at least 0.5 cmrom skin or rib cage and by more than 2 cm from theipple. On these criteria, 27 patients were initially evalu-ted, of whom 4 (15%) were excluded in this study becausef a lesion less than 2 cm from the nipple (n � 2), and theesion margin was unclear on ultrasound imaging (n � 2).

The characteristics of 23 enrolled patients are summa-ized in Table 1. Breast cancer was confirmed by ultra-ound-guided fine-needle aspiration in all patients beforeIFU treatment. Each patient was treated with HIFU ther-

py before modified radical mastectomy, adjuvant chemo-herapy, and radiation. They did not receive any interven-ion at the time of enrolment.

IFU device and treatment

Mode-JC HIFU therapy system (Chongqing HaifuHIFU] Tech Co, Ltd, Chongqing, China) is used in thistudy, and the technical details of this device have beenreviously described [16–21]. Briefly, this device is guidedy real-time ultrasound imaging during the therapeutic pro-edure. The therapeutic ultrasound beam is produced by a2-cm diameter piezoelectric ceramic transducer. It oper-tes at a frequency of 1.6 MHz, and the focal length is 90m. The focal region is 3.3 mm along the beam axis and 1.1m in the transverse direction.HIFU treatment was performed in all patients under

ither general anesthesia (n � 19) or intravenous sedationn � 4). After suitable anesthesia, the patient was placedrone and immobilized on the treatment table so that thekin overlying the tumor was easily in contact with theegassed water. By using ultrasound imaging, a planningession was performed to identify the targeted tumor vol-me, which was subsequently divided into parallel slices of5-mm separation. To achieve a sufficient tumor-free mar-in, the HIFU-targeted volume included the breast lesionnd marginal breast tissue 1.5 to 2.0 cm around the visible

able 1haracteristics of patients treated with HIFU

o. of patients 23

ge (y) 46.5 � 1.7umor size in diameter (cm) 3.1 � 0.79 (2.0–4.7)istologyInvasive 21Noninvasive 2

ode statusN0 12N1 5N2 6

NMI 2II 21

umor. i

Therapeutic parameters were delineated based on theosition and depth of the target. Acoustic focal peak inten-ities for tissue exposure ranged from 5000 to 15,000

/cm2. By scanning the HIFU beam in successive sweepsrom the deep to the shallow regions of the tumor, theargeted regions on each slice were completely ablated. Thecanning spread of the therapeutic transducer ranged from 1o 3 mm/s, and the track length was 20 mm. Median therapyime was 1.3 hours, ranging from 45 minutes to 2.5 hours.

istopathologic examinations

By using standard surgical technique, modified radicalastectomy was performed in all patients 7 to 14 days after

he HIFU treatment. Each breast specimen was kept at a lowemperature and immediately submitted to pathology wheret was examined grossly and histologically. Tissue blocksere sampled from the central and peripheral edges of the

reated region and surrounding untreated normal breast tis-ue for assessing the effect of HIFU ablation on the breastancer.

For electron microscopic examination, fresh tissue sam-les were put immediately into 2.5% glutaraldehyde, fixedt 4°C for 2 hours, and postfixed in 2% osmium tetroxide.ixed samples were treated with a graded series of dehy-ration and then embedded in epon. Sections 0.1 �m thickere cut, stained with uranyl acetate and lead citrate, and

xamined with a transmission electron microscope at 60 kVHitachi H-600, Tokyo, Japan).

The fresh samples were frozen, batched, and stored at70°C for enzyme histochemical examination. Nicotin-

mide adenine dinucleotide-diaphorase (NADH) stain waserformed on 8-�m cryostat-cut unfixed sections placed onlass slides for the analysis of tumor viability. Chemicalsere obtained from Sigma-Aldrich Corporation (St Louis,O). Each tissue section slide was incubated for 15 minutes

t room temperature with 100 �L of incubation media,hich consisted of 2.5 mg/mL of a-NADH 1 mL, 2 mg/mLf nitroblue tetrazolium chloride 2.5 mL, phosphate-buff-red saline (pH 7.4) 1 mL, and Ringer solution 0.5 mL.ach slide was subsequently washed in distilled water for 2inutes. Slides were assessed for characterization of stain-

ng during 24 hours of processing.Finally, the remaining specimens were fixed in 10%

hosphate-buffered formalin (pH � 7). They were embed-ed in paraffin, cut in 4-�m thick slices, and stained with&E. By using the biotin-streptavidin-peroxidase immuno-istochemical staining, the expression of carbohydrate an-igen 15-3 (CA15-3) and vascular endothelial growth factorVEGF) was determined in each removed breast cancer. Therimary antibodies used were mouse monoclonal antihumanA15-3 (Maxim Biotech Inc, San Francisco, CA) and an-

ihuman VEGF (Santa Cruz Biotechnology, Santa Cruz,A). The tumor cells were considered as positive when

here was a homogeneous and clearly visible signal present

n cytoplasm (brown) and negative if the signal was absent.

E

aupws

R

H

wpbrshibn

Ctt2fttth

ithbtccecsetrructotmB

ca

E

arrabbtmcpmcltvsai

H

wtcvcTta

Fcct(ilcca

181F. Wu et al. / The American Journal of Surgery 192 (2006) 179–184

valuation and analysis

The gross and microscopic characteristics of the centralnd peripheral edges of the treated region and surroundingntreated normal breast tissue were recorded. Rates of com-lete and incomplete coagulation induced by HIFU ablationere descriptively documented. In immunohistochemistry

taining, the positivities were assessed qualitatively.

esults

istological examination

Immediately after HIFU treatment, local mammary edemaas noted in all patients. Fourteen patients had mild localain, warmth, and sensation of heaviness in the treatedreast. Among them, 4 patients required a 3- to 5-dayegimen of oral analgesics for pain relief. At the time ofurgery, physical examination revealed that the local edemaad almost disappeared, and a larger palpable lump was feltn all patients because the treated region consisted of thereast lesion and a 1.5- to 2.0-cm margin of adjacentormal breast tissue.

The ablated tumors were clearly defined in each patient.oagulation necrosis of the targeted tissue, including the

umor and 1.5 to 2.0 cm of normal breast tissue surroundinghe tumor, was confirmed by macroscopic observation in the3 patients. It was gray-white or gray-yellow in color andelt firm on palpation. A red hemorrhagic ring was seen athe margin between the treated and untreated breast tissue;his might represent an inflammatary reaction of normalissue to the ablation. However, there was no evidence ofemorrhage in the central region of the treated lesion.

Microscopic examination showed 2 aspects of histolog-cal change in the peripheral and central tissue of the ablatedumors, respectively. In the peripheral region, tumor cellsad typical characteristics of coagulation necrosis inducedy thermal energy, such as pyknotic nuclei, nuclear disrup-ion, and disappearance, indicative of lethal and irreversibleell damage. Along the margin of the ablation, a narrowellular band of fibrous tissue was identified, with the pres-nce of fibroblasts, inflammatory cells, collagen fibrin, andapillary network (Fig. 1A). These changes were only ob-erved at the peripheral region in all treated patients. How-ver, the same tumor cell changes shown histologically inhe peripheral region were only detected within the centralegion of the ablated tumor in 12 of 23 patients. In theemaining 11 patients, the central ablated tumors showed anique set of additional features different from classicaloagulation necrosis. H&E stain showed that cellular struc-ure looked normal in some cancer cells, without any signsf tumor cell breakdown. The tissues maintained their cy-ologic staining characteristics and preserved nuclear chro-atin, giving an appearance similar to viable cells (Fig. 1B).

y using conventional histological techniques, it was diffi- b

ult to identify whether the treated tumor cells were dead orlive, in this region, under light microscope.

lectron microscopy

Electron microscopy showed markedly varied cellularrchitectures of breast cancer cells between the peripheralegion and the central region of treated tumor. In the pe-ipheral region, the treated breast cancer had unrecognizedmorphous electron-dense material (Fig. 2A). Plasma mem-rane, intracytoplastic organelles, and nucleonic mem-rane were totally destroyed. The breast cancer architec-ure within the center of the treated tumor at low-poweragnification was abnormal. Compared with breast can-

er cells without any intervention (Fig. 2D), at high-ower magnification the breast cancer cells revealed sub-icroscopic cellular damage that showed no morphologic

oagulation necrosis by light microscopy. The tumor cellsacked nuclear membranes, and chromatin was clumped athe periphery of the nuclei. The cytoplasm contained someacuoles, cell membranes were disintegrated, and organelletructures were not identified (Figs. 2B and C), suggestingn irreversible cell death occurred in these normal-appear-ng cancer cells.

istochemical examination

NADH-diaphorase analysis for cell viability showedell-demarcated regions of negative staining within the

reated region in all patients (Fig. 3C), consistent withomplete loss of cellular viability. No pockets or foci ofiable tumor cells were identified in the treated regionompared with breast cancer cells untreated (Fig. 3D).here was no difference in this histochemical staining be-

ween the peripheral region and the central region of theblated tumor, including the 11 samples that looked viable

ig. 1. Histopathologic changes of tumor cells in patients with breastancer after HIFU treatment. (A) Ablated breast cancer presents typicalharacteristics of coagulation necrosis in the peripheral part of ablatedumor, including nuclear disruption and disappearance of (a) tumor cells,b) fat cells, and (c) normal breast tissue. New-growth granulation tissue isdentified along the margin of the ablation (arrow). (B) Cellular structureooks normal in some cancer cells at the central part of ablated breastancer after HIFU treatment, without any signs of destruction. The tumorells maintain their cytological and nuclear staining characteristics, givingn appearance similar to viable cells (H&E staining �100).

y H&E staining in the central region that showed nonvi-

awu

I

tls(clssIomc

C

ai

Fccmtwdccb

Fanctsst(((Na

FaccdsV

182 F. Wu et al. / The American Journal of Surgery 192 (2006) 179–184

bility by NADH-diaphorase staining. However, the marginas very clear between the treated region (dead cells) andntreated region (alive cells), as shown in Figure 3A.

mmunohistochemical staining

Both CA15-3 and VEGF positivities were confined toumor cell cytoplasm. Twelve (52%) of 23 primary breastesions treated with HIFU had a positive CA15-3 expres-ion, and 7 (30%) of 23 expressed positive VEGF stainingFig. 4). Among them, a CA15-3–positive expression oc-urred in 7 patients who had no morphologic necrosis byight microscopy and VEGF in 4 patients, respectively, ashown in Figure 4. Most of the positive tumors had multiplemall foci of tumor cells exhibiting cytoplasm reactivity.nterestingly, the preserved and damaged cellular structuref breast cancer showed intense cytoplasmic positivity si-ultaneously, regardless of the histological changes under

ig. 2. Changes in cellular structure of tumor cells in patients with breastancer after HIFU treatment. (A) At the peripheral region, breast cancerells treated with HIFU present unrecognized amorphous electron-denseaterial, and it is difficult to identify plasma membrane and intracytoplas-

ic organelles (�6000). (B) At the central tumor, the cancer cells treatedith HIFU contain some vacuoles in cytoplasm, cell membranes areisintegrated, and organelle structures are not clear (�3500). (C) Theontinuity of nuclear membrane is interrupted (arrow) in the HIFU-treatedancer cells (�12,000). (D) Subcellular structures in a removal sample ofreast cancer without any intervention (a control) (�10,000).

onventional light microscope. a

omments

Ultrasound is a form of mechanical energy. It is gener-lly believed that thermal effect and cavitation are directlynvolved in the tissue damage during HIFU exposure. The

ig. 3. Histochemical changes in NADH activity at the marginal region ofblated tumor after HIFU treatment. (A) The treated tissue (a) presentsegative staining, and the untreated tissue (b) shows positive staining (blueolor). A clear margin (arrows) is detected between the treated and un-reated regions (NADH-diaphorase histochemical stain �400). (B) H&E-tained adjacent section of the same sample treated with NADH-diaphorasetaining. The treated tissue (a) presents typical characteristics of coagula-ion necrosis in the peripheral part of ablated tumor, and granulation tissueb) is observed along the margin of the ablation (arrows) (H&E stain �100).C) The negative staining of nonviable breast cancer treated with HIFUNADH-diaphorase histochemical stain �400). (D) Positive expression ofADH-diaphorase staining in a removal sample of breast cancer without

ny intervention (a control) (NADH-diaphorase histochemical stain �400).

ig. 4. Immunohistochemical changes in the expression of tumor associ-ted antigens VEGF and CA15-3 after HIFU treatment (SP immunohisto-hemical stain �400). (A and B) CA15-3–positive expression (brownolor) in the treated cancer cells (arrow) that have nuclear disruption andisappearance (A), or thermally ablated tumor cells with an appearanceimilar to viable cells (B). (C and D) Positive expression (brown color) ofEGF in the treated cancer cells that have pyknotic nuclei (C) or thermally

blated tumor cells with an appearance similar to viable cells (D).

aqraaeicncssocfimroccddtHicfithattfic

tmzndmitwtssaaasdtobw

dasmgcbagpbaelft[dtadittnpttsH

tatvdwbttHs

camHttuawiHm

183F. Wu et al. / The American Journal of Surgery 192 (2006) 179–184

ssessment of the HIFU efficiency in tumor ablation re-uires pathological confirmation. Previous animal studiesevealed that H&E stain, a routine histological method, wasstandard method to verify the viability of the treated cells

fter HIFU ablation [5–10]. By using this staining, clearvidence of complete coagulation necrosis could be clearlydentified in the treated tissue under light microscope. In thislinical study, H&E staining showed complete coagulationecrosis of the treated tumor in 12 patients (52%) with breastancer, the same histological results as found in the animaltudies. But in the remaining 11 patients (48%), the cellulartructures of breast cancer were preserved in the central partf the ablated tumor, and the tissues also maintained theirytological and nuclear staining characteristics. Similarndings have been previously reported in a patient withetastatic liver cancer after HIFU treatment [22]. This

aised the question of whether incomplete destruction hadccurred in the ablated tumors. However, electron micros-opy revealed that those who had normal-appearing cancerells, with light microscopy, showed an irreversible celleath, indicating the preservation of cellular structure in-uced by thermal fixation, instead of incomplete coagula-ion necrosis. The high temperature, which was caused byIFU exposures in the treated region, appeared to result

n either cell death or the denaturation of enzyme proteinonstituents, similar to the preservation seen with formalinxation. As a result of thermal fixation, the central part of

he ablated tumor resisted degradation because the wound-ealing process could not extend this region immediatelyfter HIFU treatment. In contrast, in the peripheral region,he local response resulted in the formation of granulationissue, neutrophil and monocyte infiltration, and activatedbroblasts that can breakdown, resolve, and repair sites ofoagulation necrosis.

Furthermore, demonstration of cytological enzyme ac-ivity was used in this study to develop an alternativeethod for identifying cell viability more easily. This en-

yme histochemical analysis is based on the reduction ofitroblue tetrazolium chloride, a redox indicator, by NADH-iaphorase, resulting in an intense blue cytoplasmic pig-ent. The activity of this enzyme has been shown to subside

mmediately on cell death [23,24]. In this study, we foundhat H&E staining of cell necrosis was associated directlyith cell nonviability by NADH-diaphorase. Also, within

he ablated tissue, areas that seemed to be normal by H&Etaining were deemed nonviable by NADH-diaphorasetaining. This suggests that NADH-diaphorase stain is moreccurate and more objective than H&E staining in assessingcute cell death because it is based on the presence orbsence of enzyme function instead of changes in cellulartructure. Therefore, the combination of H&E and NADH-iaphorase stains should be performed to confirm whetherhe treated tumor is viable or nonviable in the tissue samplesbtained from core biopsy after HIFU ablation. Follow-upiopsy provides scanty tissue making interpretation as to

hether there has been complete ablation of the tumor very d

ifficult. As such, follow-up radiologic examinations, suchs contrast ultrasound, enhanced computed tomographycan, or magnetic resonance imaging, provide alternativeodalities to assess the whole ablative appearance of tar-

eted tumors after HIFU ablation. They can show thehanges in the vascular perfusion of either tumor or normalreast tissue and distinguish the difference between nonvi-ble and residual viable tumor directly in the treated re-ions. CA15-3 is a high–molecular-weight (300–450 kDa)olymorphic epithelial mucin. It has a high expression inreast cancer cells, which is usually related to clinical stagend progression [25,26]. VEGF is a heparin-binding mol-cule (46 kDa), which plays an active role as an endothe-ial cell-specific mitogen and as a vascular permeabilityactor. Expression of VEGF in breast cancer has been showno increase tumor growth, angiogenesis, and metastases27,28]. Although high temperatures could induce proteinenaturation during HIFU exposure, it was unclear whetherhis change would have an effect on the expression ofssociated tumor antigens such as CA15-3 and VEGF in theenatured breast cancer after HIFU treatment. Therefore,mmunohistochemical staining was performed in this studyo detect subtle phenotypic changes in the breast cancer ando explore the feasibility of using cellular markers for theecrosis induced by HIFU. However, CA15-3–positive ex-ression was confirmed in 52% patients receiving HIFUreatment and VEGF in 30% patients, respectively, despitehe histological changes. As a result, this method is notufficient to confirm the coagulation necrosis caused byIFU ablation.There are several problems encountered in HIFU abla-

ion for breast cancer. General anesthesia may be problem-tic in an HIFU procedure. It can be explained by the factshat general anesthesia could be necessary at the early de-elopment of a new technique. However, other studies in-icated that either local anesthesia or intravenous sedationas feasible during the procedure for the thermal ablation ofreast cancer [12,29], and this result may change our selec-ion of anesthetic methods in the future. Patients whoseumor is close to the skin are not suitable candidates forIFU treatment because there is a possibility of causing

kin burn or leaving residual cancer cells.The lack of reliable histological techniques for assessing

oagulation necrosis induced by HIFU may complicate thepplication of this noninvasive treatment for patients withalignancy. This study reveals that pathological changes inIFU-ablated tumor are comprised of not only obvious

issue destruction detected by light microscopy but alsohermal fixation of the tumor. The thermally fixed tissue issually located in the center of the tumor, and cytologicalnd nuclear staining characteristics are normally maintainedith a similarity to viable cells. When using H&E staining,

t is difficult to identify fixed cancer cells from dead cells.owever, histochemical demonstration of cytological enzy-atic activity can be used as an alternative method for

istinguishing morphologic difference between classical ne-

ccetha

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

184 F. Wu et al. / The American Journal of Surgery 192 (2006) 179–184

rosis and fixed tissue. Thus, it is concluded that HIFU canause the heat fixation of ablated tumor through thermalffect, and the combination of both H&E and NADH his-ochemical stains could provide a reliable assessment foristological analysis of therapeutic effects on tumor cellsfter HIFU treatment.

eferences

[1] ter Haar GR. High intensity ultrasound. Semin Laparosc Surg 2001;8:77–89.

[2] Wu F, Wang ZB, Chen WZ, et al. Extracorporeal high intensityfocused ultrasound ablation in the treatment of 1038 patients withsolid carcinomas in China: an overview. Ultrason Sonochem 2004;11:149–154.

[3] Clement GT. Perspectives in clinical uses of high-intensity focusedultrasound. Ultrasonics 2004;42:1087–1093.

[4] Kennedy JE. High-intensity focused ultrasound in the treatment ofsolid tumours. Nat Rev Cancer 2005;5:321–325.

[5] Adams JB, Moore RG, Anderson JH, et al. High-intensity focusedultrasound ablation of rabbit kidney tumors. J Endourol 1996;l10:71–75.

[6] Prat F, Centarti M, Sibille A, et al. Extracorporeal high-intensityfocused ultrasound for VX2 liver tumors in the rabbit. Hepatology1995;21:832–836.

[7] Rowland IJ, Rivens I, Chen L, et al. MRI study of hepatic tumoursfollowing high intensity focused ultrasound surgery. Br J Radiol1997;70:144–153.

[8] Yang R, Reilly CR, Rescorla FJ, et al. High-intensity ultrasound inthe treatment of experimental liver cancer. Arch Surg 1991;126:1002–1010.

[9] Wang ZB, Wu F, Wang ZL, et al. Targeted damage effects of highintensity focused ultrasound (HIFU) on liver tissues of Guizhouprovince miniswine. Ultrason Sonochem 1997;4:181–182.

10] Daum DR, Smith NB, King R, Hynynen K. In vivo demonstration ofnoninvasive thermal surgery of the liver and kidney using an ultra-sonic phased array. Ultrasound Med Biol 1999;25:1087–1098.

11] Wu F, Chen WZ, Bai J, et al. Pathological changes in human malig-nant carcinoma treated with high-intensity focused ultrasound. Ultra-sound Med Biol 2001;27:1099–1106.

12] Gianfelice D, Khiat A, Amara M, et al. MR imaging-guided focusedUS ablation of breast cancer: histopathologic assessment of effective-ness—initial experience. Radiology 2003;227:849–855.

13] Wu F, Wang ZB, Cao YD, et al. A randomised clinical trial ofhigh-intensity focused ultrasound ablation for the treatment of pa-tients with localised breast cancer. Br J Cancer 2003;89:2227–2233.

14] ter Haar G. Acoustic surgery. Phys Today 2001;54:29–34.

15] Wu F, Wang ZB, Chen WZ, et al. Advanced hepatocellular carci-noma: treatment with high-intensity focused ultrasound ablation com-bined with transcatheter arterial embolization. Radiology 2005;235:659–667.

16] Wu F, Wang ZB, Zhu H, et al. Extracorporeal high intensity focusedultrasound treatment for patients with breast cancer. Breast CancerRes Treat 2005;92:51–60.

17] Wu F, Wang ZB, Chen WZ, et al. Extracorporeal focused ultrasoundsurgery for treatment of human solid carcinomas: early Chineseclinical experience. Ultrasound Med Biol 2004;30:245–260.

18] Wu F, Chen WZ, Bai J, et al. Tumor vessel destruction resulting fromhigh-intensity focused ultrasound in patients with solid malignancies.Ultrasound Med Biol 2002;28:535–542.

19] Kennedy JE, Wu F, ter Haar GR, et al. High-intensity focusedultrasound for the treatment of liver tumours. Ultrasonics 2004;42:931–935.

20] Wu F, Wang ZB, Chen WZ, et al. Extracorporeal high intensityfocused ultrasound ablation in the treatment of patients with largehepatocellular carcinoma. Ann Surg Oncol 2004;11:1061–1069.

21] Wu F, Wang ZB, Chen WZ, et al. Preliminary experience using highintensity focused ultrasound for the treatment of patients with ad-vanced stage renal malignancy. J Urol 2003;170:2237–2240.

22] Kennedy JE, ter Haar GR, Wu F, et al. Contrast-enhanced ultrasoundassessment of tissue response to high-intensity focused ultrasound.Ultrasound Med Biol 2004;30:851–854.

23] Neumann RA, Knobler RM, Pieczkowski F, Gebhart W. Enzymehistochemical analysis of cell viability after argon laser-induced co-agulation necrosis of the skin. J Am Acad Dermatol 1991;25:991–998.

24] Solbiati L, Ierace T, Goldberg SN, et al. Percutaneous US-guidedradio-frequency tissue ablation of liver metastases: treatment andfollow-up in 16 patients. Radiology 1997;202:195–203.

25] Kerin MJ, McAnena OJ, O’Malley VP, et al. CA15-3: its relationshipto clinical stage and progression to metastatic disease in breast cancer.Br J Surg 1989;76:838–839.

26] Gion M, Mione R, Nascimben O, et al. The tumour associated antigenCA15.3 in primary breast cancer. Evaluation of 667 cases. Br JCancer 1991;63:809–813.

27] Adams J, Carder PJ, Downey S, et al. Vascular endothelial growthfactor (VEGF) cancer: comparison of plasma, serum, and tissueVEGF and microvessel density and effects of tamoxifen. Cancer Res2000;60:2898–2905.

28] Toi M, Inada K, Suzuki H, Tominaga T. Tumor angiogenesis inbreast cancer: its importance as a prognostic indicator and the asso-ciation with vascular endothelial growth factor expression. BreastCancer Res Treat 1995;36:193–204.

29] Hynynen K, Pomeroy O, Smith DN, et al. MR imaging–guidedfocused ultrasound surgery of fibroadenomas in the breast: a feasi-

bility study. Radiology 2001;219:176–183.