Ultrasound-based thermometry

In ultrasound imaging, thermometry is usually carried out using the knowledge of the speed of sound at ambient temperature inside tissues (generally assumed to be 1540 m/s in soft tissues [13]). Here, similarly to MR imaging, fatty tissues introduce difficulties in temperature monitoring. In normal soft tissues speed of sound is increasing parallel with temperature up to 50-60 ^◦C, and the attenuation coefficient becoming lower simultaneously [51]. This trend turns to the reverse at higher temperatures, moreover, during the cooling back of the previously heated tissue, these values usually does not follow the course which was obtained during heating, due to the irreversible changes in treated tissues. Heat expansion here also plays an important role, as the distance between scatterers becoming bigger.

Active ultrasound

One approach using active ultrasound is temperature monitoring, using channel data delays. As it was mentioned previously, an increase in temperature causes an

increase in speed of sound of non-fatty tissues. This phenomena will decrease arrival time of the signal where this backscattered beam crosses the heated regions of tissue.

There are examples, where the used tissue model is simple; consist of regions, heated transversely to the imaging plane [13]. To build such models, knowledge of temperature dependence of imaged tissue is essential.

Having the knowledge of temperature dependence of tissue, these simple models can be built and practical to use, when considering a smaller region. In this specific example, monitoring of cardiac radiofrequency ablation was investigated using a transoesophageal (TEE) and an intra-cardiac (ICE) probe. Their aperture are in the range of the extent of heated area, which confirm the choice of the model.

If the aperture is bigger (e.g. a commonly used phased array is employed), there are several techniques, such as speckle tracking, to investigate smaller region of interests. Speckle is usually undesirable, and considered a kind of noise, which makes the images grainy. The phenomenon is arising from sub-wavelength size scatterers.

Thus, in thermometry one can benefit from speckle patterns, as it allows to estimate tissue motion and deformation with high resolution. [50] The estimation of echo-shift is obtained by determining the peak of cross-correlation between consecutive frames (att and t+ 1).

There are some artefacts, which are limiting the accuracy of such temperature monitoring systems. Usually these models are ignoring thermal lensing artefacts and assume tissue heterogeneity. Thermal lensing artefact occurs when the backscattered beam travels through the heated area. As temperature gradients are higher, this distortion is becoming stronger. Using spatial compounding, it is possible to reduce the effect of lensing at the cost of temporal resolution.

As mentioned earlier, the difference between tissues – in particular muscle and fat – can significantly distort the results of such algorithms as they rely on a spe-cific tissue model. In reality both temperature dependence of speed of sound and acoustic attenuation greatly varies from tissue to tissue. The solution is usually to introduce a post-processing algorithm taking into account some known constraints about temperature fields. [50]

Passive ultrasound

Passive ultrasound thermometry is based on receiving acoustic radiation from the tissues without any acoustic transmission from the imaging transducers. The main difficulty here is that the received signal has very low amplitude compared to ac-tive ultrasound methods. Therefore, it usually requires a more complex model, to exclude noise and exclude information outside from the region of interest and acqui-sition takes much more time usually as well. The basic difference is that in active ultrasound one knows the exact time of sound emission, meanwhile in passive ul-trasound one has only a receiver unit without the knowledge of the depth of the incoming signal explicitly.

In passive ultrasound, one can measure natural signals radiated by the human body, or in another hand, it can be used to monitor e.g. HIFU treatment. In the first setup, radiation is emitted by chaotic motions inside tissues, which system usually called as an acoustic black body [48, 49]. In the case of HIFU treatment, region of interest (ROI) is irradiated using a high Q number transducer – which means it is emitting long pulses with a very short bandwidth. Thus, cavitation is induced in the ROI. As the focal region is usually a small spatial region (several mm³), the transducer should be moved to destroy all the undesired tissue. For this reason, the time of treatment should be several hours long. During such long treatments, except CT all modalities mentioned in this review are suitable to detect the changes inside, however, it is reasonable to use a method, which does not introduce additional acoustic beam into the ROI. Hence, one convenient way to measure effects of HIFU treatment is using a transducer (-array), which passively detects these changes. The receiver array practically placed transversely to the HIFU beam. [44]

Optoacoustics

Optoacoustics is a phenomena, where a light source irradiates a boundary of the body (this can be the skin, or the intestinal wall using as transrectal probe [47] as well). When the light waves reach such a boundary, some part of it is converted into acoustic energy. During heating or cooling the tissue, the thermoacoustic effi-ciency (Gr¨uneisen parameter) changes in a linear fashion. Petrova et. al. [47] has

shown that heat monitoring in highly vascularized tissue can be done, due to the exclusive compartmentalisation of absorbing molecules in hemoglobin. Moreover, this temperature-dependent optoacoustic response was found to be independent of oxygen saturation of blood – so the same response could be obtained from both arterial and venal blood.