A randomised clinical trial of high-intensity focused ultrasoundablation for the treatment of patients with localised breast cancerF Wu*,1, Z-B Wang1, Y-De Cao1, W-Z Chen1, J Bai1, J-Z Zou1 and H Zhu11Institute of Ultrasonic Engineering in Medicine, and Clinical Center for Tumor Therapy of 2nd Hospital, Chongqing University of Medical Sciences,Box 153, 1 Medical College Road, Chongqing 400016, ChinaHigh-intensity focused ultrasound (HIFU) is a noninvasive treatment that induces complete coagulative necrosis of a tumour at depththrough the intact skin. This study was to explore the possibility of using HIFU for the treatment of patients with localised breastcancer in a controlled clinical trial. A total of 48 women with biopsy-proven breast cancer (T1 – 2, N0 – 2, M0) were randomised to thecontrol group in which modified radical mastectomy was performed, and the HIFU group in which an extracorporeal HIFU ablationof breast cancer was followed by modified radical mastectomy. Short-term follow-up, pathologic and immunohistochemical stainswere performed to assess the therapeutic effects on tumour and complications of HIFU. The results showed that no severe sideeffect was observed in the HIFU-treated patients. Pathologic findings revealed that HIFU-treated tumour cells underwent completecoagulative necrosis, and tumour vascular vessels were severely damaged. Immunohistochemical staining showed that no expressionof PCNA, MMP-9, and CD44v6 was detected within the treated tumour cells in the HIFU group, indicating that the treated tumourcells lost the abilities of proliferation, invasion, and metastasis. It is concluded that, as a noninvasive therapy, HIFU could be effective,safe, and feasible in the extracorporeal treatment of localised breast cancer.British Journal of Cancer (2003) 89, 2227 – 2233. doi:10.1038/sj.bjc.6601411 www.bjcancer.com& 2003 Cancer Research UKKeywords: high-intensity focused ultrasound; focused ultrasound surgery; breast carcinomas; therapy; ablationBreast cancer is the most common malignancy in women and, eachyear, more than 1 million new cases of breast cancer are diagnosedworldwide (McPherson et al, 2000). Radical and modified radicalmastectomy including axillary lymph node dissection has long beenregarded as appropriate therapies. During the last two decades,significant advances have been made in the development of earlydetection modalities and therapeutic methods. Breast conservationsurgery, combined with radiotherapy, chemotherapy, and hormo-nal therapy, is performed with increasing frequency in patientswith early-stage breast cancer. The more from the mastectomytoward breast conservation therapy has not changed long-termsurvival rates of patients with breast cancer (Veronesi et al, 1993;Fisher et al, 1995; Jacobson et al, 1995; Curran et al, 1998).In spite of this progress, much remains to be achieved, andmajor new therapies will be required to keep on the fight againstbreast cancer. Recently, cryoablation (Staren et al, 1997), laser(Robinson et al, 1998; Dowlatshahi et al, 2000), radiofrequency(Bohm et al, 2000; McGahan et al, 2000), as minimally invasivemodalities, have been studied to explore the possibility of ablatingbreast lesions. However, these alternative treatments require atleast percutaneous access and only destroy small lesion in themanagement of breast tumours.High-intensity focused ultrasound (HIFU) is a noninvasivetechnique for the thermal ablation of solid tumours. Ultrasound(US) beam can be focused and transmitted through solid tissues within the body. This allows the possibility of using anextracorporeal source of US for therapeutic purpose. With real-time image guidance, HIFU could noninvasively induce completecoagulative necrosis of a target tumour, without requiring surgicalexposure or insertion of instruments into the lesion. Theseadvantages make it one of the most attractive potential therapiesfor the localised treatment of tumour. To our knowledge, most ofthe studies on HIFU have been dealing with animal experiments inwhich HIFU has induced target lesions at depth in tumour tissuesand normal organs, including liver, kidney, prostate, bladder, andsoft tissue (Chapelon et al, 1992; Chen et al, 1993; Sibille et al,1993; Prat et al, 1995; Rowland et al, 1997).Up to now, in the clinical application of HIFU ablation forhuman breast tumour, Hynynen et al (2001) reported that 11breast fibroadenomas in nine patients were treated with HIFU.Eight of the 11 lesions treated with HIFU demonstrated completeor partial ablation response. No adverse effects were detected,except for one case of transient oedema in the pectoralis muscle 2days after therapy. Huber et al (2001) used MRI-guided HIFU totreat a 56-year-old patient with breast cancer. Postproceduralpathologic examination indicated that HIFU induced lethal andsublethal tumour ablation without damage to the surroundinghealthy tissue or systemic effects. However, to our knowledge, norandomised clinical trial exists that explores the efficacy and safetyof HIFU in the treatment of patients with breast cancer, withemphasis on the assessment of histological changes in the treatedtumour. In this paper, we studied the efficacy and safty of HIFUtreatment in a randomised controlled trial for the ablation ofpatients with localised breast cancer. The purpose of this study wasto investigate the efficacy, safety, and feasibility of HIFU ablationReceived 24 April 2003; revised 26 August 2003; accepted 12September 2003*Correspondence: Dr F Wu; E-mail: mfengwu@yahoo.com British Journal of Cancer (2003) 89, 2227 – 2233& 2003 Cancer Research UK All rights reserved 0007 – 0920/03 $25.00www.bjcancer.com Clinical

in the treatment of patients with localised breast cancer. This paperreports the work in progress on this noninvasive modality, and it isnot our intention to compare it with other breast conservationtherapies for the treatment of patients with early breast cancer atpresent.MATERIALS AND METHODSHIFU therapeutic systemAs described in detail previously (Wu et al, 2001a, 2002), the HIFUtherapeutic system used in this clinical trial of breast cancertreatment principally consists of a diagnostic US device, units forcomputer automatic control, six-direction movement, and atherapeutic planning system, a US generator, integrated UStherapy transducers, and a degassed water circulation unit (seeFigure 1). The therapeutic US beam is produced by a 12-cmdiameter PZT-4 piezo-ceramic transducer with a focal length of90 mm, operating at a frequency of 1.6 MHz. The focal region isellipsoid, with dimensions of 3.3 mm along the beam axis and1.1 mm in the transverse direction.An AU3 US imaging device (Esaote, Genoa, Italy) was used asthe real-time imaging unit of the system. This imaging probe (3.5 –5.0 MHz) is situated at the centre of the HIFU transducer for real-time guidance during HIFU procedure. The US beams of thetherapeutic transducer and the imaging probe completely overlap,so that the longitudinal axis of HIFU beams is in the 2D USimaging plane. The integrated transducer is attached to electricmotors, and can be moved smoothly in the six directions withmillimetric precision. Through computer control, the imagingtransducer was placed either against the skin or at a distance fromthe skin in the water for pretreatment imaging. However, duringHIFU ablation, the imaging probe was usually situated in the waterapart from the skin.PatientsA total of 48 women with biopsy-proven breast cancer wereselected for this clinical trial phase II. Before the beginning of thisstudy, the protocol design was approved by the ethics committee atour university. In accordance with the specification stipulated bythe Helsinki Committee, an informed consent form was signed byeach patient at the time of enrollment after the principles of HIFUthermal ablation had been completely described. The selectioncriteria were as follows: histologically proven invasive breastcancer (T1-2, N0-2, M0); single palpable tumours no greater than6 cm in diameter; the lesion boundaries visualised with colorDoppler US imaging, circumscribed at least more than 0.5 cm fromskin or rib cage, and more than 2 cm from nipple. All patients wereolder than 18 years and had no breast implants. They had stablehaematogenic parameters, and no history of active myocardialinfarction within the past 6 months. They were randomised to two treatment groups: the control group (n ¼ 25), in which modifiedradical mastectomy was performed without any intervention priorto surgery; or the HIFU group (n ¼ 23), in which extracorporeal insitu ablation of the breast cancer with HIFU was followed bymodified radical mastectomy within 1 – 2 weeks. A 2-week follow-up after HIFU treatment was performed to evaluate the potentialside effects of HIFU such as skin burns, local pain or discomfort,mammary oedema, haemorrhage or infection, and fever. Aftersurgery, all patients received relevant doses of local radiationtherapy, conventional chemotherapy, and hormonal therapy, aspart of their adjuvant treatment.Preoperative clinical assessments included the patient’s history,a physical examination, haemotology tests, electrolytes, renal, andliver function tests. Color Doppler US imaging of the diseasedbreast, with a 10 MHz linear array probe (Q-2000, Siemens,Erlangen, Germany), was performed by a radiologist. Directvisualisation of blood flow within the lesion was imaged byadjusting color setting for optimal slow flow detection. If a patientwas recruited to receive the thermal ablation on the basis of theimages, US-guided core needle biopsy was performed to confirmhistological diagnosis. Technetium bone scanning and a chestradiograph were required for the each case. A comparison of theeligible patients’ characteristics in both groups, including agedistribution, tumour size, histologic diagnosis, node status, andTNM classification, is shown in Table 1. Three of 23 patients inHIFU group received one pre- and postprocedural magneticresonance imaging (1.0-T scanner, Impact, Siemens, Erlangen,Germany). The MRI protocol included transverse conventionalspin echo (SE) T1-weighted image (WI), FSE T2-WI with fatsaturation and dynamic contrast-enhanced (Gadolinium,0.2 ml kg1, Magnevist, Berlex Laboratories, Wayne, USA). Tworadiologists reviewed the pre- and postprocedural MR imaging,and reached a consensus in each patient.HIFU 3D conformal treatmentIn this study, real-time US imaging was utilised to monitor theHIFU ablation procedure. It can accomplish three separatefunctions, including real-time targeting of the tumour to betreated, guidance of US energy deposition within the treatedregion, and rapid real-time assessment of the volume ofcoagulation necrosis during therapy. High-intensity focusedultrasound treatment was performed in the patients underintravenous sedation (N ¼ 4) or general anaesthesia (N ¼ 19).During and following suitable anaesthesia, the patient wasmonitored to track the blood pressure, pulse, respiration rate,Figure 1 JC-HIFU therapeutic system for tumour (Chongqing HAIFUtTechnology Company, People’s Republic of China). Table 1 Characteristics of eligible patients in both groupsCharacteristics Control group HIFU groupNo. of patients 25 23Age (years) 45.571.2 46.571.7Tumour diameter (cm) 3.570.23 (1.8 – 5.6) 3.170.79 (2.0 – 4.7)HistologyInvasive 21 21Noninvasive 4 2Node statusN0 13 12N1 6 5N2 6 6TNMI 2 2II 22 21III 1 0HIFU ablation for human breast cancerF Wu et al2228 British Journal of Cancer (2003) 89(12), 2227 – 2233 & 2003 Cancer Research UKClinical

temperature, and peripheral oxygenation. Then, the patient wasplaced prone and carefully positioned, so that the skin overlayingto the lesion to be treated was easily in contact with degassedwater.The coaxial US imaging device was used to establish the 3Dimage of the whole tumour. For therapeutic purposes, the wholetumour was divided into slices with 5 mm separation using USimages. By scanning the HIFU beam in successive sweeps from thedeep to the shallow regions of the tumour, the targeted regions oneach slice were completely ablated. This process was repeated sliceby slice to achieve complete tumour ablation, in a mannerresembling that of cutting away slices of bread. In this study, theextent of HIFU treatment was larger than the tumour extent, inorder to ensure the ablation of any adjacent microsatellites and toobtain a sufficient tumour-free margin. It included the breastlesion and its marginal breast tissue about 1.5 – 2.0 cm around thevisible tumour. This is a principle that is routinely employed inconventional surgery. During the HIFU treatment, each slice of USimages taken before and immediately after the treatment of eachslice were compared. These differences between treated anduntreated areas were used for monitoring the therapeutic effect.In this study, the target tissue was exposed at acoustic focal peakintensities from 5000 to 15 000 W cm2. The scanning speed ranged1 – 3 mm s1, and the track length was 20 mm. High-intensityfocused ultrasound treatment time ranged from 45 min to 2.5 h(median, 1.3 h).Pathologic examinationIn the clinical trial, modified radical mastectomy was performed inall of the eligible patients. In all, 23 patients in HIFU groupunderwent the operation within 1 – 2 weeks after the thermalablation. All removed specimens including breast tumour andnormal breast tissue were evaluated by gross and histologicalobservations. They were serially sectioned at approximately 5 mmintervals and the slice thickness was measured. Gross observationswere recorded in a protocol containing the overall appearance,size, and shape. By using a calipers, the visible lesions on each slicewere measured in both parallel and perpendicular axes, andvolume was calculated by multiplying the area by slice thickness.Representative sections were prepared for routine microscopicanalysis. They were fixed in 10% phosphate-buffered formalin(pH ¼ 7), embedded in paraffin, cut at distances of 1 mm in 4 mm-thick slices, and stained with haematoxylin and eosin (H&E). Tostudy the possibility of tumour vascular wall destruction inducedby HIFU ablation, elasticity fibrin and collagen fibrin doublestaining (Victoria blue and ponceau’s histochemical staining) wereperformed.ImmunohistochemistryImmunohistochemistry was performed in all removed breast tissuefrom eligible patients. Proliferating cell nuclear antigen (PCNA),cell adhesion molecule CD44v6, and matrix metalloproteinase-9(MMP-9) are molecular indicators that represent malignantbehaviours of breast cancer cell such as proliferation, invasion,and metastasis, respectively. They were identified using thebiotin – streptavidin – peroxidase immunohistochemical technol-ogy in treated breast cancer in the HIFU group, breast cancer inthe control group, and normal breast tissue from the HIFU group,respectively. The paraffin-embedded breast specimens weresectioned in 5-mm-thick slices.The slices were deparaffinised in xylene, and rehydrated.Endogenous peroxidase was blocked with 0.3% hydrogen peroxidein methanol for 20 min. Antigen retrieval was performed usingmicrowave heating. Subsequently, sections were preincubated with1% bovine serum albumin. The primary antibodies used weremouse monoclonal anti-human PCNA, anti-CD44v6, and anti- MMP-9. All of them were obtained from Santa Cruz Biotechnology(Santa Cruz, CA, USA). Sections were incubated at roomtemperature with three kinds of the primary antibody, respectively,followed by biotinylated second antibody incubation. Thechromogen was 3,3-diaminobenzidine tetrahydrochloride (brown).Sections were incubated with phosphate-buffered saline, whichserved as a negative control instead of the primary antibody.The number of immunostained cells was counted in amicroscopic grid, 0.5  0.5 mm2 in size (0.25 mm2), using amicroscopic field of  200. A total of 10 areas with the mostabundant distribution were chosen in each case for this observa-tion. Two independent observers judged the distribution of thesethree variables and counted the positive cells. If the results wereinconsistent, the two observers discussed to reach a consensus onthe semi-quantitative scoring. For the PCNA, we used the labellingindex of PCNA in cancer cells to evaluate the result. The number ofPCNA þ cancer cells was counted among 100 cancer cells using a 200 microscopic field.Statistical analysisAll observed data are displayed as the mean value plus or minusthe standard deviation. The statistical significance of any observeddifference between the mean values of control and treatmentgroups was evaluated using an unpaired Student’s t-test. Thevariability in data was compared with the F test for variance. Thedifferences in percentage data were analysed by using the Fisher’sexact test. Statistical significance was defined as a P-value of lessthan 0.05.RESULTSShort-term follow-upAmong 23 patients, 16 cases underwent surgery at 7 – 10 days afterHIFU ablation, and the remaining seven patients were operated at11 – 14 days. After HIFU, oedema was noted in the mammary tissuesurrounding the treated tumour. In all cases, the treated areaincluded the tumour and a 1.5 – 2.0 cm margin of normal breasttissue around the tumour. At 7 – 10 days postoperatively, theoedema gradually disappeared. In total, 14 patients experiencedmild local pain, warmth, and sensation of heaviness in the diseasedbreast. But only four patients were given 3 – 5 days prescription fororal analgesics. One patient had minimal skin burn, which hadentirely recovered by 10 days post-HIFU. There was no incidenceof bleeding or infection of the treated breast requiring interven-tion. Three patients received one MRI examination after HIFUtreatment. Compared with preprocedural images, enhanced MRIrevealed the absence of contrast enhancement in the treated regionincluding tumour and 1.5 – 2.0 cm normal breast tissue surround-ing the tumour (see Figure 2), indicative of coagulative necrosis.Pathologic examinationAll treated tumours in the 23 patients were confirmed by grossobservation after removal of the diseased breast. Neither bleedingof the treated regions nor injury of intervening tissues wasidentified, indicating the safety of HIFU ablation. Macroscopicexamination showed that HIFU treatment induced completecoagulative necrosis of the target tissue, which included thetumour and a mean margin of 1.8070.58 cm (range from 1.5 to2.2 cm) of an obviously normal breast tissue surrounding thetumour. This felt firmer on palpation than normal tissue. At themargin between the treated and untreated regions, there was a rimof congestion that represented an inflammatory reaction tothermal ablation. Histological examination revealed homogeneouscoagulative necrosis, including the tumour and normal breasttissue within the target region. Figure 3 shows that the tumour cellsHIFU ablation for human breast cancerF Wu et al 2229British Journal of Cancer (2003) 89(12), 2227 – 2233& 2003 Cancer Research UK Clinical

of breast cancer treated with HIFU were very abnormal, with theappearance of irreversible cell death, as demonstrated by pyknoticnuclei, nuclear disruption, and disappearance. Variable amountsof granulation tissue were noted with the presence of immature fibroblasts, inflammatory cells, and new capillaries in theboundary region. Compared with blood vessels in the controlgroup, vascular structure in the HIFU group was seriouslydamaged. Appearances were of indistinct cellular margins,endothelial disruption, and nuclear disappearance, and destructionof tunica media, indicative of coagulative necrosis (Figure 4). Thesize of the destroyed vessels was less than 2 mm in diameter,including the arteries and veins. Scattered intravascular thrombiwere often observed in the destroyed vessels. Victoria blue andponceau’s histochemical staining showed that vascular elasticityfibrin and collagen fibrin were significantly collapsed anddisrupted, as shown in Figure 4.Immunohistochemical stainingFigure 5 showed immunohistochemical staining for PCNA,CD44v6, and MMP-9 in breast cancer in the control group, normalmammary tissue, and treated breast cancer in the HIFU group,respectively. The percentages of the PCNA labelling index in thenormal breast tissue, breast cancer, and HIFU-treated tumour were6, 44, and 0%. CD44v6-positive cells were present in 56% oftumour samples in the control group (n ¼ 14), and 4.5% of normalbreast tissue (n ¼ 1). But, in the HIFU group, CD44v6-positive cellsFigure 2 Contrast-enhanced MRI of a left breast cancer (arrowed) in a36-year-old woman before (A) and 1 week after (B) HIFU ablation. Notethe absence of contrast uptake in ablated volume, including tumour and amargin of treated normal breast tissue about 1.5 – 2.0 cm around thecancer. Oedema is observed in the mammary tissue surrounding thetreated region.Figure 3 Microscopic changes of tumour cells in patients with breastcancer treated by HIFU. (A) Untreated viable breast cancer cells. (B – D):Necrosed tumour cells represented by pyknotic nuclei (B); nucleardisruption (C); and nuclear disappearance (D). H&E staining  400. Figure 4 Microscopic changes of tumour vascularity in patients withbreast cancer treated by HIFU. (A) Untreated tumour blood vessels; (B)destruction of tumour vascularity with disruption of the endothelium andtunica media, H&E staining  400. (C) untreated tumour vessel wall; (D)destroyed tumour blood vessels with the collapse and disruption ofvascular elasticity fibrin and collagen fibrin, Victoria blue and ponceau’shistochemical staining  400.HIFU ablation for human breast cancerF Wu et al2230 British Journal of Cancer (2003) 89(12), 2227 – 2233 & 2003 Cancer Research UKClinical

were not detected within the treated breast cancer (n ¼ 0).CD44v6 þ cell density was significantly higher in tumour than inthe normal breast tissue (Po 0.01), but in the HIFU group theproportion of these membrane-positive cancer cells was 0%.Compared with 60% of MMP-9 within breast cancer in the controlgroup (n ¼ 15), no MMP-9 þ cancer cells were detected in eithernormal breast cancer or treated breast cancer (n ¼ 0), as shown inTable 2.DISCUSSIONThe energy deposition of focused US on the target tissue causescoagulative necrosis of tumours (Chapelon et al, 1992; Chen et al, 1993; Sibille et al, 1993; Prat et al, 1995; Rowland et al, 1997). Inthis study, pathological examination revealed that both breastcancer cells and normal tissues within the treated region presenteda typical appearance of coagulative necrosis with severe nucleardamage. Furthermore, tumour vessels were severely damaged inthe patients with breast cancer, resulting in significant vasculardisruption and vessel occlusion. By using immunohistochemicalstaining, we found that no expression of PCNA, CD44v6, andMMP-9 of the treated tumour cells was identified in the HIFUgroup, unlike those in the control group. These findings indicatedthat the treated breast cancer cells presented had not onlyundergone coagulative necrosis in histology, but also the absenceof malignant behaviours such as proliferation, invasion, andmetastasis.The goal of our clinical study is to explore the efficacy, safety,and feasibility of extracorporeal HIFU ablation for primary breastcancer. To achieve this purpose, we hypothesise that both tumourand some normal breast tissue surrounding the tumour should becompletely destroyed with focused US energy. This larger ablationextent would guarantee an adequate ‘surgical’ HIFU marginsurrounding the tumour, and would be required if HIFU therapyis considered as an alternative to surgical procedure. This principleis routinely used in conventional surgery in order to ensure theremoval of adjacent microsatellites, and to allow for uncertaintythat frequently exists concerning the exact location of definitetumour margins. In the clinical study, histological examinationrevealed that HIFU induced coagulative necrosis of the targettumour and its surrounding normal tissue by about 1.5 – 2.2 cm(mean 1.8070.58 cm), which corresponded closely with thepretreatment extent (1.5 – 2.0 cm around the tumour) determinedin the therapeutic protocol before HIFU procedure. This resultindicated that, with real-time US guidance, HIFU can preciselycause a complete necrosis of target tissue, including the tumourand normal tissue.In clinical applications, real-time imaging modalities includingboth US (Vallancien et al, 1996; Beerlage et al, 1999; Chapelon et al,1999; Gelet et al, 1999; Wu et al, 2001a, 2002; Uchida et al, 2002)and MRI (Huber et al, 2001; Hynynen et al, 2001) can guide HIFUablation of solid tumours. These imaging techniques have beenused for targeting the lesion to be treated, monitoring thetherapeutic procedure, assessing the thermal effects on targettissue, and controlling the ultrasonic energy. In this study, thereasons for choosing real-time US for the guidance of HIFUtreatment were as follows: low cost, flexibility, speed of imageprocessing, and extensive availability. During the HIFU procedure,obvious changes in grey scale within the target tissue wereimmediately detected on the US image after HIFU ablation. Ourprevious studies indicated that these hyperechoic zones corre-sponded well to the extent of coagulative necrosis, and that therewas a close relationship between the extent of necrosis as measuredby gross examination and the hyperechoic extent measured on theUS image immediately after FUS in both in vivo and in vitro tissues(Wang et al, 1997; Wu et al, 1998, 2000, 2001b; Bai et al, 1999). Inthe clinical treatment of patients with breast cancer, we also foundthat the hyperechogenic regions were regular in shape and size,and conformed to the extent of coagulative necrosis induced byHIFU ablation (Liu et al, 1998; Cao et al, 2001; Wu et al, 2001;Zhou et al, 2001). Although the mechanism for tissue grey-scalechange is still unclear, US effects on tissue such as heating andcavitation are involved in the formation of coagulative necrosis,and real-time diagnostic US guidance is seen to be useful in thedetection of the coagulative necrosis of target tissue during HIFUprocedures.As a noninvasive method, HIFU has a potential application inthe treatment of patients with breast cancer. However, someproblems still exist with this technique. First, to our knowledge,neither MRI nor diagnostic US have satisfactory sensitivity forprecise visualisation of tumour margins. A completely precisePCNACD44v6MMP-9 Untreated  breast Treated  breast Normal breast cancer cancer Figure 5 Immunohistochemistry for PCNA (the first row), CD44v6 (thesecond row), and MMP-9 (the third row) in the patients with breast cancertreated by HIFU (S-P staining,  400). The immunoreactivity is shown bythe color brown. From left to right, each separated morphology isrepresented by a normal breast tissue in the HIFU group, untreated cancercells in the control group, and treated cancer in the HIFU group,respectively. Compared with untreated cancer cells, no positive cells ofPCNA, CD44v6, and MMP-9 within the treated cancer cells were observedin the HIFU group.Table 2 Results of immunohistochemical staining in both groupsPCNA labellingindex (%) CD44v6(%) MMP-9(%)Normal breast tissue 6 4.5 0Untreated breast cancer 44 56 60Treated breast cancer 0 0 0 HIFU ablation for human breast cancerF Wu et al 2231British Journal of Cancer (2003) 89(12), 2227 – 2233& 2003 Cancer Research UK Clinical

imaging modality for pretreatment demonstration of tumourextent is not currently available. Therefore, in our clinical study,image-detected breast cancer and 1.5 – 2.0 cm normal tissuesurrounding the cancer were completely ablated to allow for theuncertainty that frequently exists concerning the exact location ofdefinite tumour margins. However, for some superficial lesionsclose to the skin, there is a possibility of causing skin burns, orleaving residual viable tumour cells if the overlying skin is leftundamaged. A second problem is that a palpable firm lump wouldpersist in the treated region for a period of time after HIFUtreatment. This lump may not be satisfactory or cosmeticallydesirable for the patient, although the tissue would revert to benignstatus. This could also cause potential problems with follow-upimages, in distinguishing whether there is local recurrencepostoperatively. Finally, a suitable indication of HIFU ablationfor breast cancer is essential if this technique can be used inclinical practice. Breast tumours with undefined margins, scatteredmultiple foci, and tumours close to the nipple should be excludedas potential targets for HIFU. In conclusion, this study demonstrates for the first time thatUS-guided HIFU is effective, safe, and feasible in the treatmentof localised breast cancer with well-demarcated margins.However, much supplementary investigation is necessary toexplore these above problems, and large controlled clinical trialsmust be performed before extracorporeal HIFU can be recom-mended as a conserving breast modality in the treatment of breastcancer.ACKNOWLEDGEMENTSThis work was supported by Ministry of Science and Technology ofChina (grant no. 96-905-02-01) and National Natural ScienceFoundation of China (grant nos. 39300125, 39630340, 39630340,39670749, 39770841, 39770712, 30070217, 30171060). We thank theHaifu Company (Chongqing, China) for technical support, and DrJames Kennedy at the Churchill Hospital of Oxford in UK for thehelp with linguistic revision of the manuscript.REFERENCESBai J, Wu F, Wang ZB, Ruan XY, Zou JZ, Wang ZL (1999) Localized lesionto normal miniswine liver with high intensity focused ultrasound and thedose – effect relation. Chin J Ultrasonography 8: 247 – 250Beerlage HP, Thuroff S, Debruyne FMJ, Chaussy C, de la Rosette JJ (1999)Transrectal high-intensity focused ultrasound using the Ablathermdevice in the treatment of localized prostate carcinoma. Urology 54:273 – 277Bohm T, Hilger I, Muller W, Reichenbach JR, Fleck M, Kaiser WA (2000)Saline-enhanced radiofrequency ablation of breast tissue: an in vitrofeasibility study. Invest Radiol 35: 149 – 157Cao YD, Zhu H, Chen WZ, Wu F, Wang ZB (2001) The pathologicalchanges of breast cancer after HIFU treatment. J Pract Oncol 16: 161 – 163Chapelon JY, Margonari J, Theillere Y, Gorry F, Vernier F, Blanc E, Gelet A(1992) Effects of high-intensity focused ultrasound on kidney tissue inthe rat and the dog. Eur Urol 22: 147 – 153Chapelon JY, Ribault M, Birer A, Vernier F, Soucheon R, Gelet A (1999)Treatment of localised prostate with transrectal high intensity focusedultrasound. Eur J Ultrasound 9: 31 – 38Chen L, Rivens I, ter Haar G, Riddler S, Hill CR Bensted JP, Bensted JP(1993) Histological changes in rat liver tumours treated with high-intensity focused ultrasound. Ultrasound Med Biol 19: 67 – 74Curran D, van Dongen JP, Aaronson NK, Kiebert G, Fentiman IS, MignoletF, Bartelink H (1998) Quality of life of early-stage breast cancer patientstreated with radical mastectomy or breast-conserving procedure: resultsof EORTC trial 10801. The European Organization for Research andTreatment of Cancer (EORTC), Breast Cancer Co-operative Group(BCCG). Eur J Cancer 34: 307 – 314Dowlatshahi K, Fan M, Gould VE, Bloom KJ, Ali A (2000) Stereotacticallyguided laser therapy of occult breast tumor: working-in-progress report.Arch Surg 135: 1345 – 1352Fisher B, Anderson S, Redmond CK, Wolmark N, Wickerham DL, CroninWM (1995) Reanalysis and results after 12 years of follow-up in arandomized clinical trial comparing total mastectomy with lumpectomywith or without lumpectomy with or without irradiation in the treatmentof breast cancer. N Engl J Med 333: 1456 – 1461Gelet A, Chapelon JY, Bouvier R, Pangaud C, Lasne Y (1999) Local controlof prostate cancer by transrectal high intensity focused ultrasoundtherapy: preliminary results. J Urol 161: 156 – 162Huber PE, Jenne JW, Rastert R, Simiantonakis I, Sinn HP, Strittmatter HJ,von Fournier D, Wannenmacher MF, Debus J (2001) A new noninvasiveapproach in breast cancer therapy using magnetic resonance imaging-guided focused ultrasound surgery. Cancer Res 61: 8441 – 8447Hynynen K, Pomeroy O, Smith DN, Huber PE, McDonnold NJ, KetterbachJ, Baum J, Singer S, Jolesz FA (2001) MR imaging-guided focusedultrasound surgery of fibroadenomas in the breast: a feasibility study.Radiology 219: 176 – 185Jacobson JA, Danforth DN, Cowan KH, d’Angelo T, Steinberg SM, Pierce L,Lippman ME, Lichter AS, Glatstein E, Okunieff P (1995) Ten-year results of a comparison of conservation with mastectomy in the treatment ofstage I and II breast cancer. N Engl J Med 332: 907 – 911Liu CA, Bai J, Wu F, Zou JZ, Du YH, Chen X, Wang ZB (1998) Primaryobservations of high-intensity focused ultrasound used in causinglocalized lesion to human breast tumor tissue. Chin J Ultrasonography7: 239 – 241McGahan JP, Griffey SM, Schneider PD, Brock JM, Jones CD, Zhan S (2000)Radio-frequency electrocautery ablation of mammary tissue in swine.Radiology 217: 471 – 476McPherson K, Steel CM, Dixon JM (2000) Breast cancer-epidemiology, riskfactors, and genetics. Br J Med 321: 624 – 628Prat F, Centarti M, Sibille A, Abou el Fadil FA, Henry L, Chapelon JY,Cathignol D (1995) Extracorporeal high-intensity focused ultrasound forVX2 liver tumors in the rabbit. Hepatology 21: 832 – 836Robinson DS, Parel JM, Denham DB, Gonzalez-Cirre X, Manns F,Milne PJ, Schachner RD, Herron AJ, Comander J, Hauptman G(1998) Interstitial laser hyperthermia model development forminimally invasive therapy of breast carcinoma. J Am Coll Surg 186:284 – 292Rowland IJ, Rivens I, Chen L, Lebozer CH, Collins DJ, ter Haar GR, LeachMO (1997) MRI study of hepatic tumours following high intensityfocused ultrasound surgery. Br J Radiol 70: 144 – 153Sibille A, Prat F, Chapelon JY, abou el Fadil F, Henry L, Theilliere Y,Ponchon T, Cathignol D (1993) Characterization of extracorporalablation of normal and tumor-bearing liver tissue by high-intensityfocused ultrasound. Ultrasound Med Biol 19: 803 – 813Staren ED, Sabel MS, Gianakakis LM, Wiener GA, Hart VM, Gorski M,Dowlastshahi K, Corning BF, Haklin MF, Koukoulis G (1997) Cryosur-gery of breast cancer. Arch Surg 132: 28 – 33Uchida T, Sanghvi NT, Gardner TA, Koch MO, Ishii D, Minei S, Satoh T,Hyodo T, Irie A, Baba S (2002) Transrectal high-intensity focusedultrasound for treatment of patients with stage T1b-2NoMo localizedprostate cancer: a preliminary report. Urology 59: 394 – 399Vallancien G, Harouni M, Guillonneau B, Veillon B, Bougaran J (1996)Ablation of superficial bladder tumors with focused extracorporealpyrotherapy. Urology 47: 204 – 207Veronesi U, Luini A, Del Vecchio M, Greco M, Galimberti V, Merson M,Rilke F, Saccozzi R, Savio T (1993) Radiotherapy after breast-preservingsurgery in women with localized cancer of the breast. N Engl J Med 328:1587 – 1591Wang ZB, Wu F, Wang ZL, Zhang Z, Zou JZ, Liu C, Liu YG, Cheng G, DuYH, He ZC, Gu ML, Wang ZG, Feng R (1997) Targeted damage effects ofhigh intensity focused ultrasound (HIFU) on liver tissues of GuizhouProvince miniswine. Ultrason Sonochem 4: 181 – 182Wu F, Chen WZ, Bai J, Zou JZ, Wang ZL, Zhu H, Wang ZB(2001a) Pathological changes in human malignant carcinoma treatedwith high-intensity focused ultrasound. Ultrasound Med Biol 27:1099 – 1106HIFU ablation for human breast cancerF Wu et al2232 British Journal of Cancer (2003) 89(12), 2227 – 2233 & 2003 Cancer Research UKClinical

Wu F, Chen WZ, Bai J, Zou JZ, Wang ZL, Zhu H, Wang ZB (2002) Tumorvessel destruction resulting from high-intensity focused ultrasound inpatients with solid malignancies. Ultrasound Med Biol 28: 535 – 542Wu CX, Liu CA, Wu F, Bai J, Chen WZ, Zou JZ, Qiao TY, Wang ZB (2001)Pathological observation on implanted rabbit breast tumor treated withhigh intensity focused ultrasound. Chin J Ultrasound Med 17: 481 – 484Wu F, Wang ZB, Wang ZL, Chang Z, Bai J (1998) Changes in ultrasonicimage of tissue damaged by high intensity focused ultrasound in vivo. JAcoust Soc Am 103: 2869Wu F, Wang ZB, Wang ZL, Chen WZ, Bai J, Zou JZ, Wang ZL, Ruan XY, DuYH, Kong FB, Zhu H (2000) Dosimetry study on the extracorporeal ablation of hepatic carcinoma with high-intensity focused ultrasound.Chin J Phys Med Rehabil 22: 267 – 269Wu F, Wang ZB, Wang ZL, Zhang Z, Bai J, Chen WZ, Zhou JZ(2001b) Extracorporeal simulating ablating on ox liver tissuewith 3D scanning high-intensity focused ultrasound. J Pract Oncol 16:379 – 381Zhou JZ, Wu F, Liao CR, Liao CR, Chen WZ, Zhu H, Xie ZP, Peng JP,Zhang YX, Lu M, Bai J, Wang ZL, Wang ZB (2001) CDFI surveillanceand evaluation of the efficiency of the high intensity focusedultrasound in treatment of breast carcinoma. Chin J Ultrasound Med17: 24 – 25HIFU ablation for human breast cancerF Wu et al 2233British Journal of Cancer (2003) 89(12), 2227 – 2233& 2003 Cancer Research UK Clinical