The present invention relates to a device for therapeutic treat ^¬ ment and a method for controlling a therapeutic treatment device according to the independent claims.

In particular the invention concerns treatments by focused high intensity ultrasounds (HIFU) .

Conventionally, in the HIFU (High intensity focused ultrasound) technology, an acoustic treatment transducer emits concentrated acoustic waves into a target tissue. These waves are absorbed by the tissue, which provokes a temperature rise in the tissue in the focal region. This temperature elevation in turn allows the creation of a necrosis and thereby allows destruction of living tissue at a distance without any direct contact.

One of the main challenges associated with the HIFU technique is the control of the temperature increase and of the size of the lesion created with this technique. Among the possibilities for control some device implement MRI imaging which allows a direct visualisation of the temperature in the treated area. Those MRI systems are however expensive and lead to high costs of treat ^¬ ment .

Furthermore, indirect measurements using ultrasound imaging as an indication of temperature or lesion size have been proposed e.g. in WO 2010/020730.

Some HIFU systems that are guided by ultrasound use constant in ^¬ tensity values to obtain a correct treatment efficacy. Such an approach is only possible when the target to be treated is very close to the HIFU source. In cases when the HIFU beam has to go through skin and potentially other layers of tissue before reaching the targeted area the use of constant values to adapt the treatment efficacy would not allow to compensate for the different configurations of the ultrasound path which would lead to very large variations of the acoustical intensities in the targeted area.

Another approach is to use the occurrence of hyperechoic marks (HEM) as an indicator for the local changes of tissue properties. Such an approach has for example been described in WO 2010/020730 or WO 2011/036485. Due to the fact that the detec ^¬ tion of the apparition of hyperechoic marks is not easy and hy ^¬ perechoic marks occur in a temperature range of approximately 80 °C to 100 °C, the detection of hyperechoic marks is not consid ^¬ ered sufficiently reliable for the control of HIFU treatments.

It is an object of the present invention to prevent the disad ^¬ vantages of the prior art and in particular to create a device and a method that create a reliable way to control the effect of the treatment on the target.

The object is achieved by a device and a method according to the independent claims.

In particular, the object is achieved by a device for therapeu ^¬ tic treatment comprising at least an acoustic treatment trans ^¬ ducer, at least one detector and at least one means for time measurement or means for detecting the frequency of appearance of changes of tissue properties. The acoustic treatment trans ^¬ ducer is able to emit high intensity waves (HIFU) toward a tar ^¬ get in order to treat the target, wherein the high intensity waves have a focal point. The detector is able to detect a change of tissue properties caused by the high intensity waves at the target in a manner basically known to the skilled person. The means for time measurement is able to measure the time t _mark from the beginning of emission of a pulse of HIFU until changes of tissue properties are detected.

In the present context changes of tissue properties are defined to be any sign of a change in the acoustic properties of tissue under treatment. In particular, the invention refers to observa ^¬ tions of changes of tissue per se. In particular, the change in the properties of the tissue without any changes of the environ ^¬ ment within the tissue is observed, in particular, e.g. without the presence of any additional agents such as contrast agents. Preferably, the changes in acoustic properties are hyperechoic marks. The change in the acoustic properties of the tissue can be broadband emissions, which can occur due to cavitation. In particular hyperechoic marks are a manifestation of increased acoustic reflectivity of the tissue in the treated area, notably originating from newly created gas bubbles. Furthermore, such changes can for example stem from temperature increase and tis ^¬ sue boiling, inertial cavitation or tissue hardening. In particular, the temperature increase and tissue boiling caused thereby can be considered as a phase change of the tissue per se which can be observed in accordance with the present invention. Hyperechoic marks may reflect the sudden occurrences of bubbles in the tissue, which will locally increase the reflection of in ^¬ coming ultrasonic waves back to the transducer. Those reflected waves may in turn be detected by whitening on the ultrasonic im ^¬ age of the treated area or by a surge in the reflected electri ^¬ cal signal to the HIFU transducer. The device allows for an indirect control of the HIFU treatment by means of measuring a time which consequently leads to a reli ^¬ able but nevertheless cost efficient device.

Based on the measured value of t _mark , the power P _e and/or the pulse duration t _on of the HIFU transducer is adapted to obtain an optimal treatment result.

It is observed in real treatment situations that t _Ma r _k values may vary from one pulse to the other. Since HIFU pulses are also limited in time (typically a few seconds) , this leads to situa ^¬ tions where changes in tissue properties do only appear for some pulses. An example of such a situation is a case where t _on =4s, while the median value of t _Ma r _k is 3s and t _Ma r _k Interquartile Range (IQR) is 2s. In such a situation, only ¾ of the pulses will ex ^¬ hibit changes of tissue properties. For this reason, the rate of apparition of changes of tissue properties can also be taken into account for the control of HIFU treatment.

According to the invention P _e represents either a constant value of electrical power for the duration of the pulse, or a variable power law P _e (t) which by definition is not null over the pulse duration t _on -

The detector can comprise a acoustic imaging transducer, preferably a second acoustic imaging transducer, able to emit waves to provide an imaged representation of the target and of its en ^¬ vironment .

The imaged representation of the target can be provided before, during, or after the emission of the power waves from the high intensity transducer. By means of such a device the HIFU impact at the target can be controlled in a cost-effective way. The focal point of the emitted HIFU waves can be in an ultra ^¬ sound imaging plane of the second acoustic imaging transducer.

By such an arrangement the imaging transducer enables to detect the occurrence of changes of tissue properties directly at the target without any delay.

The detector can comprise a device for detection of variation of pattern of interference during HIFU pulse as for example de ^¬ scribed in WO 2010/020730.

The device detects changes of tissue properties, in particular hyperechoic marks, by means of the detection of a change of structure of the interference image, wherein the interference image is created due to an interference of the HIFU transducer waves with the imaging transducer waves.

The use of the variation of pattern of interference during HIFU pulse enables uninterrupted firings of the HIFU transducer. This leads to a more efficient treatment.

According to another aspect of the invention, there is provided a device for therapeutic treatment, preferably as described above, comprising an acoustic treatment transducer a radio fre ^¬ quency generator generating a radiofrequency signal to power the acoustic treatment transducer and a detector. The detector includes a radiofrequency directional coupler able to detect the characteristic features of the reflected radiofrequency signal. The radiofrequency coupler is arranged on the power line of the HIFU transducer, which will allow the separation between the forward radiofrequency power (used for therapeutic ultrasound generation by the transducer) and its reflected part. The characteristics features of interest in the reflected signal may be its envelope, or its spectral content along with their respective evolution with time or other characteristic features of a radiofrequency signal. A change in the reflected part of the radiofrequency signal is indicative of a local change of the tissue properties so that a monitoring of the reflected part of the radiofrequency signal is equivalent to monitoring changes of tissue properties, in particular hyperechoic marks. Likewise, in case the detection of changes of tissue properties is done via the reflected radiofrequency signal, t _mark is represented by the delay between the onset of the pulse and the detected change in the radiofrequency signal.

An example of a radiofrequency coupler which can be applied for the above mentioned purpose is the Werlatone 40dB Dual Direc ^¬ tional Coupler, Model C5571. For this application only the reverse power coupling port will be used.

Such a measurement of the occurrence of changes of tissue prop ^¬ erties allows for a cost-efficient and easy way to either only detect the occurrence of changes of tissue properties or pref ^¬ erably to determine the time until hyperechoic marks occurs, i.e. the determination of t _mark .

The devices as can further comprise means for storage of a ref ^¬ erence curve and means for comparing measured values with a ref ^¬ erence curve.

A reference curve can be based on experience, on theoretical de ^¬ duction, on statistical evolutions or a combination thereof. Typically, one reference curve per clinical application is stored. Preferably, the reference curve contains expectancy val ^¬ ues for t _mar k and/or t _on in combination with an electrical power of the HIFU transducer. The reference curve may be calculated for different target depth or may include a derating factor al ^¬ lowing taking target depth into account

The device can comprise a display for displaying the time t _mark to an operator. The display enables the operator to adjust the power and the pulse duration of the HIFU transducer to obtain an optimal treatment result.

The device can comprise means for determining a power P _e for sub ^¬ sequent HIFU pulses based on t _mark . Preferably, the power P _e for subsequent pulses is determined based on t _mark by means of a ref ^¬ erence curve as described above.

Means for determining the power P _e for subsequent pulses allow for an automatic adjustment based on the previous measurement and hence allow for an optimal treatment result.

The device can comprise several detectors, preferably at least two detectors .

The use of several detectors eliminates the random error of one detection means. Hence, a measurement is more reliable and accu ^¬ rate. Preferably, different methods for detection of changes of tissue properties are used, e.g. one based on ultrasound imagery of the therapy beam and one based on the characteristic features of the reflected radiofrequency signal or passive cavitation de ^¬ tection. As a matter of course, even other methods and various combinations are applicable. This also eliminates method- immanent measuring errors. The values of t _mark determined by the different methods are preferably used as a basis for a resulting tmar _k -value that is used for the determination of the power P _e and/or pulse duration t _on - The device can comprise several means for time measurement and means for determining an average t _mark or for determining a weighted average time from different t _mark values.

The detector preferably provides an indication of occurrence of t _mark as well as a confidence value of the accuracy of the detec ^¬ tion. The confidence value is taken into account for determina ^¬ tion of the weighted average in a manner basically known to the skilled person.

The determination of a weighted average t _mark enhances the accu ^¬ racy of the adjustment of the treatment power.

According to another aspect of the invention there is provided a method for controlling a therapeutic treatment device, prefera ^¬ bly a device as described above. The method comprises a step of monitoring a HIFU treatment by a HIFU transducer of a target by means of measuring a time t _mark , wherein the time t _mark is the time from the beginning of emission of a HIFU pulse to detection of changes of tissue properties or by measuring the frequency of appearance of changes of tissue properties f _mark - In this case t _mark can be substituted by f _mark -

Based on t _mark , or f _mark several types of decisions can be made by the operator of the device or automatically by the device itself in order to better regulate the treatment.

The regulation of the treatment can be conducted by changing the power P _e or the pulse duration t _on - Furthermore, the time between pulses t _0ff and/or the special position of the pulses can be adapted . Since the time t _mark is correlated to the intensity of the HIFU beam in the target zone, this method provides for a cost- efficient and easy control of a HIFU treatment.

The analytical derivation of t _mark can be done by the model by Parker, K.J., Effects of heat conduction and sample size on ultrasonic absorption measurements . J Acoust Soc Am, 1985.

77(2): p. 719-25. The model expresses the temperature rise in the focus of a HIFU pulse by solving the bio-heat transfer equa ^¬ tion. During the pulse the increase in temperature ΔΤ at time t is given by: a 4*k,

AT ^: exp(-r ² /a ² ) * ln(l + ^■ 0 (1)

Where :

r = distance from the beam axis,

I = in-situ intensity of the HIFU pulse which is roughly proportional to P _e

a = half width of the focus in tissue

Ko = thermal conductivity of tissue,

Co = thermal capacity of tissue

Po = density of tissue,

Simplifying the notation gives:

ΔΓ = I * B ₀ *t ₀ * exp(-r ² /a ² ) * ln(l +— )

t ^o (2)

=

with: 4 ^H *k ^Λ ° (3)

Where to represents the time required for heat to diffuse through the focus; and only depends on the tissue.

Equation 2 can be inverted to obtain the time t at which a certain temperature increase ΔΤ is obtained.

AT

t * exp(

(5) t _mark is the time when the boiling temperature T _boi i is reached and hyperechoic marks can be detected. In this case a tissue tem ^¬ perature of 37°C before application of the pulse is assumed.

Hyperechoic marks can be detected by measuring the variation of a pattern of interference during HIFU pulse. The interference occurs between ultrasound waves from an imaging ultrasound transducer and an acoustic treatment transducer.

A measurement of the occurrence of hyperechoic marks by measur ^¬ ing the variation of a pattern of interference during a HIFU pulse does not interrupt the treatment and hence enhances the efficiency of the treatment. Furthermore, the measurement by an ^¬ other ultrasound transducer is cost-efficient.

According to another aspect of the invention, there is provided a method for controlling a therapeutic treatment device prefera ^¬ bly as described above. The method comprises the step of deter ^¬ mining changes of tissue properties by detecting characteristic features of a reflected radiofrequency (RF) signal by means of an radiofrequency (RF) coupler on a power line of the HIFU transducer .

Such a method is reliable and simple and does not require a sup ^¬ plemental transducer to achieve the goal of observing a HIFU pulse .

The changes of tissue properties can be detected by a acoustic imaging transducer, preferably a second acoustic imaging trans ^¬ ducer, able to emit waves to provide an imaged representation of the target and of its environment.

The detection of changes of tissue properties by a second acous ^¬ tic imaging transducer is cost-efficient and leads to reliable determination of t _mark .

The therapeutic treatment device can be initialized, wherein the initialization is directed to the values of pulse duration t _on and/or pulse power P _e , whereby the initialization includes the steps of emitting at least one initial calibration HIFU pulse from an acoustic treatment transducer onto a target to be treated at an electrical power Pi. The time t _mark between the be ^¬ ginning of emission of the initial calibration HIFU pulse and detection of changes of tissue properties can be measured and the electrical power P _e for subsequent pulses can be determined in order to obtain a predefined intensity of HIFU at the target.

The initialization can further be based on a theoretical value of Pi or a value based on experience or a combination thereof in ^¬ stead of or in addition to the initial calibration HIFU pulse.. The electrical power P _e is determined based on the known vari ^¬ ables after initialization Pi and t _mar k- In simple cases this could be a linear transformation or a proportional integral de ^¬ rivative (PID) . More generally a curve such as the one repre ^¬ senting equation (6) may be used. Since the bio-heat transfer equation is a differential equation in most of the cases, the adaptation of P _e can be based on predefined curves for the spe ^¬ cific clinical application to get an accurate result.

The power P _e of the HIFU pulse emitted from the acoustic treat ^¬ ment transducer can be adapted on the value of t _mar k measured for the previous pulse or based on the evolution of t _mark over several previous pulses and consequent adjustment of the electrical power P _e and/or the time t _on -

The adaption of the power P _e based on one or more previous pulses enables the control of the treatment and hence leads to a safe and efficient treatment.

Furthermore, the power P _e or t _on of the HIFU-pulse emitted from the acoustic treatment transducer can be adapted based on the frequency of appearance of changes of tissue properties.

The adaption of the power P _e or t _on based on the frequency of ap ^¬ pearance of changes of tissue properties also enables the con ^¬ trol of the treatment and leads to a safe and efficient treat ^¬ ment .

The adjustment of the power based on t _mar k or on the evolution of t _mar k can be conducted while taking into account its inherent dis ^¬ persion by applying techniques inspired by statistical process control or filtering. The method can comprise the following steps:

- determining a time T _mark i from the beginning of the

emission of a HIFU pulse until hyperechoic marks are detected with a first method of detection hyperechoic marks ,

- determining a second time T _mark 2 with a second method different from the first,

- determining t _mark on the basis of T _mark i and T _mark2 .

This approach leads to a more reliable determination of t _mark since random errors are reduced.

The invention is described in the following by means of embodi ^¬ ments .

Schematic view of a treatment device

Schematic view of a radiofrequency coupler for determi ^¬ nation of the reflected HIFU signal

graphic view of an example of a reflected signal enve ^¬ lope

graphic view of calculated times to HEM over intensi ^¬ ties for various boiling temperatures

graphic view of a determination of P _er ef in a given sam ^¬ ple

graphic view of t _mark over a standardized P _e of a com ^¬ plete sample

graphic view of t _mark error distribution

usage of the statistical model over the course of a treatment Fig. 9 temperature profiles and intensity profiles. The con ^¬ stant line, where the intensity is kept constant is shown for comparison.

Figure 1 shows a schematic overview of a treatment device 1 ac ^¬ cording to the invention.

The device 1 is controlled by the user through a graphical user interface 7 (for instance a touchscreen) . The processing unit 6 allows control of the interaction between the different subsys ^¬ tems .

An acoustic treatment transducer 2 emits high intensity acoustic waves (HIFU Beam) 2b onto a target 3. The HIFU waves are focused onto the target 3 within a target area. A radiofrequency genera ^¬ tor 8 provides the necessary radiofrequency power for generation of the HIFU beam. This radiofrequency generator 8 is controlled both in terms of power and timing by the processing unit 6.

An ultrasound imagery transducer 4 allows the observation of the target area and of the beam propagation, provided that its field of view 4b encompasses the HIFU beam focal point. An ultrasound scanner 5 generates an image stream that is fed to the process ^¬ ing unit 6.

The reflected power signal is tapped on the radiofrequency power line by a radiofrequency directional coupler 9, and then fed to the processing unit 6 through an Analog to Digital Conversion (ADC) unit 10. The processing unit 6 contains a circuit / algo ^¬ rithm to detect a change in the RF signal. A counter / timer measures the time between the onset of the pulse and the instant the RF signal changes. Figure 2 shows a preferred embodiment of the processing occur ^¬ ring in the processing unit 6. The processing unit 6 provides a common timebase 11 that will allow synchronization of HIFU emission by radiofrequency generator 8 with detection of HEM. This feature allows deriving t _Ma r _k from the knowledge of HIFU pulse timing and detection of HEM.

The ultrasound image stream coming from ultrasound scanner 5 is fed to image processor 12 where a technique such as the one de ^¬ scribed in WO 2010/020730 allows retrieving information on the presence of interference patterns. The output of this processor is fed to detector 13 which infers T _Ma r _k ^Imag along with detection quality information from the output of 13 and from the timing information provided by timebase 11.

The reflected radiofrequency signal provided by the analog to digital conversion unit 10 is processed by signal processor 14 in order to retrieve features such as its envelope and/or its spectral content. These features are then provided to HEM detec ^¬ tor 15, along with timing information from timebase 11 to obtain T _Ma r _k ^RF along with detection quality information.

Both T _Mark ^Image and T _Mark ^RF are then used by processor 16 to provide a best estimate of T _Ma r _k to the user through the graphic user in ^¬ terface 7.

As a complement, direct outputs from processors 12 & 14 may be directly provided to the user for supplemental information. An example of a reflected power signal is provided in Figure 3. This example yields a T _mark ^RF of 2.67s. On the basis of this information, the user is able to adapt treatment parameters (electrical power and t _on ) and provide these to the control timebase for the next HIFU pulse.

As a complement, T _Ma rk information may be provided to a model and statistical processor 17, that will compute the best estimate of electrical power to use for the next pulse as described above.

An example for deciding on a value for P _e , a decision table can be used. The following table corresponds to a case were a treat ^¬ ment is carried out with the objective of having hyperechoic marks appearing at t _Ma rkG = 3s. For the sake of simplicity, we consider that the intrinsic variability of t _Mar k, o = Is. For each HIFU pulse, we use the model t _Ma rk =A (eB . PeRef/Pe-1 ) as a refer ^¬ ence. This allows to define - when a mark appears - a value for PeRef (n) as:

F¾ _e/ C«) =—ln{ )

Pe _Xt/ r ⁽ n ⁾ j_ _s interpreted as the necessary electrical power that would have been necessary in the conditions of HIFU pulse n to

Obtain t{ _a rk ⁼ t{ _a rkG-

In a typical implementation, the case where no marks appear dur ^¬ ing a pulse, an arbitrary value such as 1.5 * t _0n instead of t _Mar k is used to obtain a value for ^j P%->f ^« 5. This provides a value for each pulse.

Inputs Output tMark (n) t _M ark (n- 1 ) Pe (n+1 )

t _mark ^G -o < tM<

any Pe (n)

t _ma rk +0 ^"

t _mark ^G +o < tM<

t _mark ^G +o < tM< (Pe _Ref (n)+ Pe _Ref (n-1) ) /2

t _ma rk +2o

t _ma rk +2o

otherwise Pe (n)

t _mark ^G +2o < tM

any (Pe _Ref (n)+ Pe _Ref (n-1) ) /2 or no HEM t _mark ^G -2o < tM<

t _mark ^G -2o < tM< (Pe _Ref (n)+ Pe _Ref (n-1) ) /2

tmark ^— 0 ^~

tiriark Q

otherwise Pe (n)

tM < t _mark ^G -2o any (Pe _Ref (n)+ Pe _Ref (n-1) ) /2

This information is also provided to the user through graphic user interface 7.

Figure 3 shows an example of a reflected signal envelope. This example yields a t _mark of 2.67s.

Figure 4 shows the calculated time t _mark to hyperechoic marks as a function of intensity for various boiling temperatures. This curve links in situ acoustical intensity with T _{Ma r} k , and thus al ^¬ lows an indirect estimation of the in situ tissue characteris ^¬ tics. As can be seen in figure 4 and this is confirmed by ex ^¬ perience, t _mark is sharply dependent upon the intensity. Therefore if the intensity is too low, no mark will appear (t _mark > tow ) · Even if the intensity is constant, the appearance of hyperechoic marks will depend much upon the local characteristics of the tissue (these are translated as "boiling temperature" in the model) and it will be difficult to regulate the power with effi ^¬ ciency because so many t _mark would be missing during the course of the treatment.

Figure 5 shows the measured values of t _mark over the power of the HIFU-device 1 (see Figure 1) for one given sample with homogene ^¬ ous coupling conditions. On each sample, a total of 15 to 25 HIFU pulses were delivered. In order to assess repeatability of T _{Ma r} k in homogeneous situations, the following analysis was car ^¬ ried out:

By hypothesis, acoustical conditions are considered to be iden ^¬ tical in a given sample: Coupling and pre-focal attenuation are the same for the whole sample. For a given sample, retrieve all values of T _Ma rk and corresponding electrical power P _e . Obtain for the considered sample an estimate of the electrical power value P _e for which t _Ma rk would have been equal to t _Ma rkRef value. Provide the best fit for the data with a mathematic model, in this case: t _mark =A(e ^B/Pe -l) Define PeRef for the considered sample as the value for which the mathematical model intercepts t _Mar kRef- Some samples are rejected at this stage because of the lack of dif ^¬ ferentiation in terms of power (typically all lesions have been created with the same electrical power) . For each lesion, calcu ^¬ late a standardized power, which is defined as P _e /P _e Ref- The proc ^¬ ess is repeated for each sample, allowing to define a full data- set of t _Mar k over standardized power. The measurement of these de ^¬ lays has been carried in various coupling conditions (skin, various attenuations, etc..) . The dataset consist of 55 samples. The corresponding result is presented in Figure 6.

The data presented in Figure 6 allows determining a model for tmark vs. relative power as an exponential with form: T _Mark ^est = t _mark =A (e ^{B ' PeRef/Pe} - 1 ) . The numerical expression for a t _ma rk reference of 2.8s gives A = 0.77s and B=96.98

On the basis of this model, it is possible to control treatment power P _e By providing an assessment for P _e Ref- This is obtained through the following steps:

1· TMark (1) is measured through a first pulse with a given electrical power P _e (1). By inversion of the model t _ma rk =A(e ^{B PeRef Pe} -l), an initial value of P _eRe f is calculated.

2. Once PeRef is calculated, it is possible to compute the electrical power to apply in order to obtain a given expected value for T _Ma rk

3. The process is iterated over the successive pulses in order to improve the estimation of P _eRe f on the basis of all data pairs {Pe(n), T _Mar k(n)}, including the cases where no hyperechoic marks have been detected. In a typical implementation, these steps are carried out by Sta ^¬ tistical processor 17 of fig 2.

Figure 7 shows a residual error distribution of T _Ma rk ^est with the aforementioned model. Variance is less than Is, even though the data set includes a wide variety of coupling situations.

Figure 8 shows the usage of the statistical model over the course of a treatment. In step 18 the value is set to be n=l . The initial, application based proposed values for P _eP (n=l), T _onP (n=l) are determined. In step 19 the operator decides on the values of P _e (n) , T _on (n) for pulse number n. In step 20 the HIFU Pulse n at Pe (n) is applied. In step 21 T _Mark (n) is measured dur ^¬ ing the application of the pulse. In step 22 new proposed value for next pulse are computed based on the measurement of T _Ma rk(n) and the knowledge of P _e (n) . For the following pulse in step 23 n = n+1 the operator again decides as in step 19. Step 24 is the end of the treatment. The values P _e (n) , T _Mark (n) are collected to further improve the statistical model described in Figure 6.

In an attempt to always obtain hyperechoic marks during each pulse, it is proposed to expose the tissue to pulses with in ^¬ creasing power over time. For example the electrical power could start with 80 We and ramp up to 200 We over a t _on =4 s pulse dura ^¬ tion. If a change in tissue properties is detected, the pulse would be aborted, thus impeding over exposure.

A mathematical derivation yields the temperature rise as a func ^¬ tion of time in the form of a differential equation:

ΔΓ

^: exp(-r ² / a ² ) I exp(- ( -) (6)

at Figure 9 shows examples of profiles of intensity vs. time. The process is carried out numerically. The figure shows examples of temperature profiles calculated with variable power. With pro ^¬ file "Lin-" the intensity is decreased linearly with time, y- ielding an essentially flat temperature after a 4 s warm-up pe ^¬ riod. This type of profile could be used if hyperechoic marks are detected at the beginning of the pulse (for example during the scanning phase of the treatment) and the operator wants to maintain an essentially constant temperature throughout the pulse .

With profile "Lin+" the intensity is increased linearly with time, yielding an essentially linear temperature rise with time, after a 2 s warm-up period. This type of profile would be best if no hyperechoic marks are detected during the previous pulses (for example at the beginning of the treatment) . Profile "Lin++" is similar to "Lin+" but the intensity rises more sharply with time. The result is an almost linear increase of temperature during the pulse.

If the intensity profile is sharp (starts low ends high, Lin ++ in the graph) then T (t) is quickly (almost) linear: this means that there is high probability that T will reach T _b0 ii during the pulse, meaning that hyperechoic marks will always be visible. This will help in regulating the power.

In general, the example treatment procedure comprises two phases: An initialization phase and a treatment monitoring phase .

The initialization phase comprises the steps of: An initial calibration HIFU pulse at a given electrical power is emitted from the HIFU treatment transducer. The initial power chosen is based on the knowledge of the specific clinical indi ^¬ cation and will make HEM appear within a time typically 1 to 2 seconds. t _mark is measured by device 1 as shown in Figure 1 & 2. Afterwards, the electrical power to use in order to obtain a given value of t _mark is chosen. This is done by the application of the theoretical model for t _mark as described in Figures 4-6. The treatment is then conducted based on the calculated value of the power of the HIFU transducer.

Within the treatment monitoring the values for t _Ma r _k are measured for each pulse. During the monitoring phase the power can be adjusted based on the evolution of t _mark . Preferably, the inherent dispersion is taken into account, for example by statistical process control. This can be done by the following decisions:

If t _mark is within lo of the goal value, keep current power P _e . If T _Mark remains twice in a row in the l-2o range (on the same side) , increase or decrease the power P _e by 1 power ^x step' . If t _Mark is over 2o range, the power P _e is increase or decrease by 1 power 'step' . In a typical application a power ^x step' can be defined as 20% of the current value of electrical power: this provides a logarithmic scale of power values that is well adapted to the proposed exponential model linking T _Mark and P _e .