High-intensity focused ultrasound (HIFU) is a rapidly maturing technology with diverse clinical applications. In the field of oncology, the use of HIFU to non-invasively cause tissue necrosis in a defined target, a technique known as focused ultrasound surgery (FUS), has considerable potential for tumour ablation. In this article, we outline the development and underlying principles of HIFU, overview the limitations and commercially available equipment for FUS, then summarise some of the recent technological advances and experimental clinical trials that we predict will have a positive impact on extending the role of FUS in cancer therapy.
Focused ultrasound surgery (FUS), using high-intensity focused ultrasound (HIFU) technology in combination with modern imaging methods, has the potential to ablate internal tumour target tissue with great precision, giving it all the benefits of minimally invasive surgery . Damage to adjacent or intervening tissues may be minimised with careful image-based treatment planning and the tumour target may be visualised during treatment. As it does not involve ionising radiation, it is low risk and repeat treatments are possible. The non-invasiveness of FUS reduces toxicity compared with other ablation techniques and adjacent blood vessels may be less vulnerable to damage compared with surgical risks [2,3]. FUS therefore holds great promise as a single or part of a multimodal approach for cancer treatment, especially for patients with cancers unsuitable for other established therapeutic options.
We describe how recent technical developments in HIFU equipment design, electronic control, ablation focusing and target imaging have made rapid advances that are overcoming previous limitations of HIFU for destroying target tumour tissue, especially in shortening FUS treatment times. Together with ongoing worldwide trials exploring oncology applications, this is strengthening confidence in FUS and broadening its scope. As a result, we believe that it is evolving into an increasingly more useful alternative or complementary treatment option and have continued expectation that FUS will be successfully integrated into routine future clinical practice.
PRINCIPLES OF FUS
HIFU transducers are made from piezoelectric materials that oscillate upon application of an alternating voltage, resulting in the generation of ultrasound waves in the receiving medium. They are capable of handling relatively high levels of power and focus the resulting ultrasound beam to a small “cigar”-shaped volume, typically of a few cubic millimetres. Focusing can be achieved geometrically, either by using a curved (spherical section) transducer or by using a plane transducer and a curved lens (Figure 1a). In devices that use an array of small transducers, beam focusing may also be achieved by electronic control (Figure 1b). Modern transducers can create acoustic intensities in a target tissue of ∼100–10 000 W cm−2 and peak compression pressures of up to 30 MPa. In comparison, diagnostic ultrasound transducers deliver intensities of ∼0.0001–0.1000 W cm−2 and a compression of 0.001–0.003 MPa .
Diagram illustrating focusing principles of high-intensity focused ultrasound in single (a) and array (b) transducers. Reproduced with permission from Pioneer Bioscience Publishing Company, from Khokhlova and Hwang .
Rapid elevation of the local tissue temperature is the main causative mechanism of tissue destruction. Coagulative necrosis occurs as a high amount of acoustic energy is deposited in a short period of time—a function of both tissue temperature and exposure time [3–7]. This thermal effect was the preferred mode of targeted ablation in early clinical applications of HIFU as it was most predictable and understood . Mechanical tissue effects also occur at very high ultrasound intensities [3,8–10]. Cavitation, i.e. bubble formation, occurs as microscopic gas bodies are drawn out of solution because of alternating rarefaction and compression and local temperature elevations. A low-pressure acoustic field results in stable cavitation, where microbubbles oscillate. In turn, fluid movement leads to the production of shear forces that cause cell membrane disruption and resulting cell damage—a phenomenon known as microstreaming. With high acoustic pressures, vibration-induced changes in microbubble volume result in inertial cavitation, i.e. violent bubble collapse. If this happens near the cell membrane, destruction of the cell may occur [8–13]. Radiation forces are also created in tissues owing to the absorption and reflection of the ultrasound wave energy . These cause additional destructive bioeffects, including cell membrane deformation, microstreaming and organelle rotation [8,14]. Mechanical destructive effects have been increasingly exploited as HIFU understanding, experience and technological developments have advanced. Harnessing mechanical bioeffects can result in larger treatment volumes, and hence shorter treatment times, as well as achieving very sharply demarcated precise lesions. This latter effect forms the basis of “histotripsy”—a development of HIFU tissue ablation, which uses short pulses of very high-intensity ultrasound to specifically induce mechanical bioeffects for tissue destruction [3,15].
KEY LIMITATIONS OF HIFU FOR FUS
Since ultrasound is reflected at interfaces between soft tissues and air–gas and is rapidly attenuated in bone, the presence of lung, ribs or gaseous bowel in front of the FUS target region can be problematic. Sonication through the cranium is particularly challenging owing to high attenuation and variable thickness and density of the skull. In addition, non-uniform soft tissues cause the ultrasound beam to propagate variably. Therefore, an appropriate “acoustic window” may be required for an ultrasound beam to propagate through the body to the target volume, restricting the application of FUS to specific patients/tumours. Beam scattering and diffraction may also occur. Unwanted high-energy deposition to tissues in the ultrasound pathway, resulting from energy reflected from acoustically resistant media, such as air, bowel gas or bone, to tissues with strong acoustic absorbance, such as skin, muscle or the gastrointestinal tract, can lead to complications like skin burns or serious side effects like bowel perforation owing to thermal injury [3,16]. Beam scattering, diffraction and reflection therefore need to be prevented or carefully accounted for in planning and during delivery, and acoustic coupling of the transducer to the skin throughout treatment is necessary to avoid skin burns.
Compared with HIFU transducer focal volume, clinically relevant tissue target volumes may be very large. This means that the HIFU focus may have to be moved within the target volume to achieve sufficient tissue ablation, either by shooting the beam continuously while moving the transducer or by interrupting the beam and moving the HIFU focus. When combined with the need to frequently verify the location of the focus within the body by means of imaging and ensure unwanted energy deposition to avoid side effects, excessively long treatment times can result.
ADVANCES IN TRANSDUCER DESIGN AND BEAM FOCUS THAT COUNTER LIMITATIONS
HIFU transducers need to be optimally designed for specific clinical applications. The development of piezoactive materials with specific acoustic properties, e.g. lead zirconate titanate-type ceramics and composites of piezoactive elements, that are capable of being driven at high power and can be tailored for the specific clinical application, has been an important step. For curved transducers, the radius of curvature, which determines the distance at which the focal volume is located, and transducer diameter, which determines the surface area, are important parameters. In the case of arrays, the size and number of individual elements required to achieve appropriate acoustic power, their spatial distribution and their relationship to operating frequency are evaluated. Often elements are placed on a curved surface to achieve some geometric focusing [17,18]. The advantage of arrays is that the electrical signals applied to each element can be varied . Using multichannel electronics, the acoustic fields produced by individual elements can be used coherently to produce a single focus that can be adjusted in size, shape and position and manoeuvred through a clinically relevant volume, or several foci can be created simultaneously. This increases the overall volume that can be ablated and achieves faster treatment times. An important factor in transducer array design is the compromise between performance, which favours a large number of elements, and cost and complexity, which favours a small number of elements. Many specific designs for arrays have now been reported. For example, ablation of large deep-seated tissue volumes has been reported with a 256-element phased array ; high-power beam steering through human skull was demonstrated with a 200-element sparse phased array ; high-power acoustic fields were achieved with an intracavitary 57-element aperiodic array device designed for prostate treatment ; and an endorectal transducer with 1000 elements for high-resolution treatment of prostate conditions has been clinically approved . One drawback associated with arrays is that of “grating lobes” caused by sound energy spreading out from the transducer in undesired directions, which occurs when the element spacing is greater than a half wavelength. Several methods to minimise this have now been reported [21,22,24,25], including a patented random array design (Figure 2), which further reduces the time taken to deliver therapy and avoids delivering significant acoustic energy to non-targeted tissues, even when multiple simultaneous foci located off axis are produced .
The spherical surface of a patented array transducer with randomly distributed elements that allows multiple simultaneous foci and minimises off-target energy delivery. Reproduced with permission from IOP Publishing Ltd, from Hand et al .
Design and testing of a HIFU system with flexible and controllable multifocus pattern ability is another important advance. Using a 256-element spherical section phased array system capable of producing “fit-to-shape” multifocus patterns, e.g. X, S, C, square and Q shapes, simulation and phantom experiments showed that treatment volumes could be up to 6.6 times greater in one sonication. Further, by using three-dimensional (3D) focus steering, it was feasible for other subarrays to operate if some of the elements were blocked by ribs, providing the device with the ability to avoid obstacles . Advanced phased array systems with up to 20 000 elements, allowing 3D multiple foci sonication and rapid beam steering, are currently under further technological research and development.
The new transducer design has also allowed increased exploitation of mechanical effects to enhance ablation. Controlled use of cavitation can induce larger target lesions—thus achieving reduced treatment times—and research in this area is ongoing. A new approach using an endocavitary plane transducer showed that cavitation effects were induced beyond a threshold dose of acoustic intensity in ex vivo studies. Further, when the cavitation effect was combined with the thermal effect, it was possible to necrose cylindrical target volumes up to 31 cm3 in 4 min .
In transcranial HIFU, where skull ultrasound wave refraction can cause severe beam degradation, there have been recent important developments to improve focusing, including validation of an in vitro 3D CT adaptive correction method. A specifically designed 300-element spherical array therapeutic transducer was used in conjunction with CT scan acquisitions to deduce acoustic properties of the skull. Precise beam refocusing was achieved through ex vivo human and monkey skulls with a positioning error < .7mm. A later development by the same group, which used MR acoustic radiation force imaging for energy-based adaptive focusing in the human cadaver head, showed greater enhancement of transcranial ultrasound beam focusing , paving the way for in vivo human trans-skull FUS.
COMMERCIAL FUS DEVICES IN CLINICAL USE
There are currently two commercially available intracavitary FUS clinical devices: the Ablatherm® (EDAP TMS, Lyon, France), which was jointly developed with the French Institute of Medical Research in the early 1990s, and Sonablate® 500 (SonaCare Medical, Charloltte, NC, previously Focus Surgery Inc., Indianapolis, IN), which was developed in the USA in 1994 (Figure 3a,c). Both use a single moveable probe, are guided by ultrasound imaging, and have been employed in trials for treating prostate cancer . They also have potential application in other pelvic malignancies. Ablatherm has a robotically controlled treatment probe with dual ultrasound transducers. Sonablate 500 has a single transducer and uses a split beam technology that increases the size of the focal zone and allows near simultaneous treatment and imaging. It is more operator dependent but has the practical advantage of being fully portable.
Examples of high-intensity focused ultrasound devices currently in Western clinical use or research. (a) Ablatherm® (EDAP TMS, Lyon, France), (b) ExAblate® OR (InSightec, Haifa, Israel), (c) Sonablate® 500 (SonaCare Medical, Charlotte, NC, previously Focus Surgery Inc., Indianapolis, IN), (d) Sonalleve MR-HIFU (Philips Healthcare, Guildford, UK), (e) ExAblate Neuro (InSightec). Figures are reproduced with permission from the manufacturers.
Extracorporeal FUS devices offer a longer focal length than intracavitary devices and are more versatile overall. MRI-guided FUS (MRgFUS) extracorporeal machines include the ExAblate® system (InSightec, Haifa, Israel), which uses real-time thermometry MRI guidance (Figure 3b) and is currently used worldwide to treat uterine fibroids and in Europe to treat breast cancer, adenomyosis and bone metastasis, and is being investigated in clinical trials for other uses. The Sonalleve MR-HIFU is an alternative system (Philips Healthcare, Guildford, UK), which combines an extracorporeal HIFU system and MR coil elements integrated into a patient table compatible with Philips MRI platforms (Figure 3d). A novel electronic concentric circle beam path method is used to increase ablation volumes. Sonalleve has largely been used to treat uterine fibroids in countries other than the UK but is now under investigation for oncology applications. Extracorporeal ultrasound-guided FUS (USgFUS) machines are more popular in Asia. The Model JC focused ultrasound system (Haifu Technology Co. Ltd, Chongquing, China) originated in China. It can be operated using a choice of transducers with varying focal length and has been used to treat several cancer types including liver and renal cancer . Alternative USgFUS machines include the HIFU-2001 (Sumo Corporation Ltd, Kowloon, Hong Kong) machine, which has been used since 2001 to treat cancer patients in China, Hong Kong and Korea, the HIFUNIT-9000 tumour therapy system (Shanghai Aishen Technology, Shanghai, China) and the FEP-BYTM system (Yuande Biomedical Engineering Lim. Co., Beijing, China). Extracorporeal devices specifically designed for transcranial FUS include the ExAblate Neuro hemispheric phased array HIFU system, which is currently used only for neurosurgery research purposes in brain disorders (Figure 3e) .
CLINICAL APPLICATIONS OF HIFU IN ONCOLOGY
During the last decade, many trials have assessed intracavitary FUS as a non-invasive alternative to prostatectomy and radiotherapy for localised prostate cancer. The UK National Institute for Health and Clinical Excellence initially supported the use of intracavitary HIFU ablation in the management of prostate cancer . Although it is currently clinically used in other parts of the world, in the UK its use in the National Health Service has been recommended to be confined to clinical trials [34,35] and it is presently under interventional procedure consultation . However, because of its organ-sparing and tumour control ability, retreatment potential, recent technical advances in delivery and imaging and recent promising trial results, HIFU is strengthening as a viable alternative treatment for tumour control—particularly for patients in whom localised cancer control with minimal morbidity or effective salvage are priorities . For example, a review in 2009 on salvage HIFU following recurrent disease after radiotherapy reported biochemical disease-free rates, negative biopsy rates and complication rates similar to other salvage methods . Similar results were reported in a 2011 study of 19 males treated with HIFU for locally recurrent prostate cancer following radical prostatectomy when good cancer control with acceptable morbidity was shown . Both studies indicated better outcome for males with pre-treatment lower risk disease. A study of HIFU as salvage therapy in 22 Tokyo patients in 2011 also reported a good biochemical disease-free rate at 5 years of 52% . The use of HIFU for focal salvage therapy following radiotherapy failure was also recently indicated to reduce the harms of whole-gland salvage therapies . Moreover, recent encouraging results from a trial of HIFU as primary treatment in localised prostate cancer showed no histological evidence of cancer in 30 of 39 males biopsied at 6 months and a low rate of treatment-related genito-urinary side effects . Non-invasive MRgFUS has also been used for prostate cancer ablation and has the advantage of improved targeting and real-time temperature monitoring, but only a few studies have been conducted with human patients .
Following surgery for rectal tumours, locally recurrent disease is a major concern that is often accompanied by severe pain and incapacitating complications. There is therefore an unmet clinical need for new treatments, especially for patients with residual or progressing disease in whom all current therapies have failed.
Recently, we reported the first case exploring the feasibility of intracavitary HIFU as a therapeutic option for tumour ablation in advanced rectal cancer. The patient had originally undergone surgical resection but developed recurrent local and liver metastatic disease with debilitating symptoms and was not fit for any conventional adjuvant options. Using the Sonablate 500 HIFU device, adjusted to deliver about 50% of the intensity per pulse used for prostate cancer treatment, the exophytic part of the tumour was targeted over 29 min. Symptoms improved within 24 h, there were no complications and repeat MRI at Day 7 showed tumour necrosis of the targeted area. Furthermore, the patient’s overall physical condition improved to the extent that palliative radiotherapy became possible . A UK Phase I/II trial has since been initiated to further investigate the feasibility and efficacy of transrectal HIFU in patients with locally advanced rectal cancer (09/H0808/43).
Surgical resection or transplantation has been the gold standard treatment for both primary and metastatic liver tumours. Since the first successful HIFU liver ablation in a male in 1993 , extracorporeal HIFU approaches have been investigated and developed, concentrating on patients with unresectable hepatocellular carcinoma (HCC) or in whom comorbidity prevents surgery. Particular challenges include beam propagation through the ribs, respiratory movement of the liver and long ablation times owing to large tumour size and small focal volume [46, 47]. The high prevalence of HCC in China has driven HIFU technology to overcome the associated challenges, with emerging encouraging results.
A large randomised study in China in 2005 using the Model JC Haifu system in patients with stage IVA HCC reported median survival time to be significantly longer in patients who received combined HIFU and transcatheter arterial chemoembolisation (TACE) therapy (11.3 months vs 4 months; p=0.004) . A 2011 Chinese study of unresectable HCC showed slightly longer median survival of 12 months after combined HIFU+TACE treatment. 45% of patients achieved complete ablation, with ablation response reported as a significant prognostic factor . For HIFU treatment alone, a report in 2011 of 49 patients from a Hong Kong cancer centre who received single HIFU treatment for unresectable HCC concluded that HIFU was an effective treatment modality with a high effectiveness rate and favourable survival outcome: complete tumour ablation was reported in 80% and local tumour control was 67% at 24 months . However, serious complications have recently been reported in a minority of HCC patients, including rib fractures, diaphragmatic rupture, biliary obstruction, pleural effusion, pneumothorax and fistula formation . These have arisen from unwanted thermal damage, indicating the need for caution and improved targeting of beam energy to lower risk.
Many malignant renal lesions are small, so a non-invasive nephron-sparing therapeutic method is attractive. Initial studies, which used either multiple elements in a concave disc or the Storz investigational HIFU prototype device (Storz Medical, Schaffhausen, Switzerland), showed skin burns and problems with tissue ablation, inhibiting clinical use [45,52]. More contemporary extracorporeal and laparoscopic HIFU systems have produced smaller but better defined lesions and thus better results: however, they remain as investigative procedures, requiring improvements in order to compete with other ablative techniques . A preliminary trial in patients with advanced renal cancers was carried out in 2003 using the Model JC Haifu device. A decrease in both flank pain (90%) and haematuria (89%) were reported with no adverse events . A later study using the same device reported stable lesions in two-thirds of patients with minimal morbidity . A Phase I study of laparoscopic HIFU in 2008 showed feasibility and demonstrated that this more invasive method helped to resolve the limitations caused by bowel, rib cage and abdominal wall obstruction and respiratory motion, although technological and methodological refinements were necessary to improve targeted ablation . Feasibility, good tumour ablation and low morbidity with laparascopic HIFU was also shown in 2011 . Methods in development, such as photoacoustic real-time monitoring  and respiration-induced movement correlation modelling , or the application of MR image guidance to monitor temperature changes for optimal heat deposition and safety  may improve future non-invasive renal FUS.
Most patients with pancreatic cancer present with inoperable disease, such that palliative treatment for local tumour control and pain relief are the main aims of treatment for which HIFU may have significant benefits. The long treatment times previously required owing to large target volumes are being addressed by the development of new multi-array devices as well as methodology harnessing mechanical tissue effects to enhance tissue ablation. Increased clinical experience is further enabling its development . Early clinical studies in China supported HIFU as a primary therapy for pain relief, with no major adverse events reported [61–63]. Recent studies have confirmed pain palliation and have also indicated efficacy. A report in 2009 of all stage unresectable patients in Peking, China, treated with an FEP-BY device showed pain improvement in 80.6% of patients, an overall median survival of 8.6 months and no complications . A Phase II trial in 2010 of concurrent gemcitabine and HIFU in locally advanced pancreatic cancer using a HIFUNIT-9000 system also showed promising activity, with 78% pain relief rate, 43% response rate and a median survival rate of 12.6 months . In a 2011 report of mixed stage inoperable patients treated with HIFU alone, an 87.5% pain relief rate, no complications and an 8-month median survival was shown . In a recent European study in 2010, all six patients with tumours in difficult to treat locations showed pain relief and full tumour ablation, with one experiencing a serious complication . A minority incidence of serious complications, including third-degree burns and fistula formation, has been separately reported .
The breast is suited to HIFU treatment as it offers a soft-tissue acoustic window and can easily be immobilised. In a 2001 feasibility study of MRgFUS of 11 breast fibroadenomas using a custom-made device, 8 lesions indicated complete or partial tissue devascularisation and necrosis . In 2003, MRgFUS using ExAblate as an adjunct to tamoxifen in patients with breast carcinoma reported negative biopsies in 19 of 24 patients at 6 months . A 2007 study of MRgFUS using ExAblate in Japanese females with ductal carcinoma showed only 1 case of recurrence in 21 patients over a median follow-up of 14 months . Similar favourable results have been reported in China using the Model JC USgFUS device, with a 95% 5-year disease-free survival rate  in one study and pathology confirming ablation in all cases in another . However, limitations have included the risk of tissue damage to proximate skin, rib and lungs, which are currently being addressed by technological improvements, for example in device design , focal aberration correction  and novel contrast enhancement agents .
As ultrasound is commonly used as a first-line imaging method for investigation of urinary tract symptoms, HIFU offers an attractive means to visualise and treat bladder cancers at the same time. Encouraging results were reported in the first study of extracorporeal HIFU in superficial low-grade transitional cell bladder carcinoma, with no recurrence seen in 67% of treated patients . However, the drawbacks of long treatment time and the need for regional anaesthesia require more research. Current interest is largely placed on ultrasound-based combination therapy .
The first successful targeting of bone lesions using HIFU in animal models, causing necrosis of osteocytes, was reported in 2001 . A key potential advantage for primary bone tumours is limb sparing. A recent study using the Model JC Haifu device showed that USgFUS was feasible and effective in primary bone malignancy. Complete tumour ablation was seen in 69 of 80 patients. Further, for patients whose tumours were completely ablated with HIFU and who completed systemic chemotherapy, the 5-year survival rate was greater than reported for other treatments . Encouraging results have also been achieved for pain control of bone metastasis. MRgFUS for pain palliation in patients for whom other treatments were ineffective or not feasible showed HIFU was a safe and effective treatment option; 72% of patients reported significant pain improvement and a 67% reduction in opioid usage was recorded . Supported by clinical studies, the ExAblate MRgFUS system received the European CE mark and US Food and Drug Administration approval for palliative treatment of bone metastasis in 2007 and 2012, respectively. The first UK trial testing of the Sonalleve MRgFUS system for bone metastasis is currently under way.
There is great interest and potential use of HIFU in brain tumours. Enhancement of drug delivery across the blood–brain barrier (BBB) is a key active area of research, enabled by targeting BBB disruption [81,82]. However, for the development of effective and highly focused transcranial HIFU tissue ablation, physical problems caused by the skull have created significant technical hurdles (see section “Advances in transducer design and beam focus that counter limitations”). A recent pilot study in three glioblastoma patients using transcranial ExAblate showed focal heating was achieved, but greater device power was required to produce focal coagulative necrosis .
Non-invasive techniques that utilise HIFU to ablate tumours will enable improvements in future healthcare provision as patient morbidity can be minimised while potentially saving costs. The limitations of HIFU that have delayed its potential use in clinical practice are being overcome through advances in technology and design, ongoing research is enabling improvements and reducing risk, and experimental clinical trials for various types of tumours are showing considerable promise: for some tumour types, e.g. prostate and pancreatic cancer, randomised controlled trials are now required to compare FUS with standard treatments. Clinical applications of FUS are thus continuing to expand and improve and we predict that its benefits along with its increasingly clinically relevant fast treatment times will rapidly result in its adoption as a routine part of multimodal therapy for many cancers.