1. INTRODUCTION Thermal models are used in hyperthermia to predict temperature distributions for treatment and applicator optimization. It is known that blood flow can significantly influence temperature profiles but an accurate description of this effect is unknown. Two models that have been used to model microvascular effects are the Pennes Bioheat Transfer Equation (BHTE) and the Effective Thermal Conductivity Equation (ETCE) [1], while an advection term is used in combination with the above to model large vessel effects [2]. The purpose of this work is to compare model predictions in an experimental system and to critically examine the effects of thermally significant vessels.

Spatial Correlation of Flow Induced Temperature Gradients During Tissue Heating with Vascular Geometry using CT Angiography: Implications for Thermal Therapy

A study was conducted to compare the ability of conventional ultrasound imaging, transmission ultrasound imaging, and photoacoustic imaging to discriminate lesions created by high intensity therapeutic ultrasound. These lesions resulted in thermally and mechanically induced tissue destruction. High intensity focused ultrasound (HIFU) was used to create thermal lesions by coagulating the target tissue. A SONIX RP® clinical imaging system was used to monitor lesion growth in real time pre- during post-exposure through its endocavity 6MHz convex array probe that was coaxially mounted in the center of the therapy transducer. The coaxial transducer showed the location of the focal spot during the exposure. It was observed that two different high intensity ultrasound lesions were formed and three modalities adopted in assessing the lesions.

It has previously been demonstrated in tissue-mimicking phantoms and in tissue that envelope statistics of US backscatter are affected by changes in the scatterer properties [1–4, 32, 37]. At higher frequencies the wavelength of the US begins to approach the size of cells and cellular components and at this scale the envelope statistics of HFUS backscatter become more sensitive to structural changes within cells. To investigate the relation between the envelope statistics and cell structure, experiments were performed in vitro. The physical meaning of the fit parameters was evaluated by investigating HFUS backscatter from suspensions of various concentrations of two different cell lines of different sizes.

Recent ultrasound (US) experiments on packed myeloid leukaemia cells have shown that, at frequencies from 32 to 40 MHz, significant increases of signal amplitude were observed during apoptosis. This paper is an attempt to explain these signal increases based upon a simulation of the backscattered signals from the cells nuclei. The simulation is an expansion of work in which a condensed sample of cells, with fairly regular sizes, could be considered as an imperfect crystal. Thus, destructive interference could occur and this would be observed as a large reduced value of backscattered signals compared with the values obtained from a similar, but random, scattering source. This current paper explores the possibility that simple changes in the nuclei, such as their observed condensation or the small loss of nuclei scatterers from cells, could cause a significant increase in the observed backscattered signals. This model indicates that the greater backscattered signals can be explained by further randomisation of the average positions of the scattering sources in each cell. When these "microechoes" are added together, so that the destructive interference is reduced, a large increase in the signal is predicted. The simplified model strongly suggests that much of observed large increases of the backscattered signals could be simply explained by the randomisation of the position of the condensed nuclei during apoptosis, and the destruction of the nuclei could produce further signal amplitude changes due to disruption of the cloud of backscattered waves.

Changes in the ultrasound (US) properties of tissue during beating affect the delivery of US thermal therapy and may provide a basis for US image monitoring of thermal therapy. The US attenuation coefficient and backscatter power of fresh human prostate tissue were measured as the tissue was heated. Samples of human prostate were obtained directly from autopsies and heated rapidly to final temperatures of 45degreesC, 50degreesC, 55degreesC, 60degreesC and 65degreesC. A 5.0-MHz transducer was scanned in a raster pattern over the tissue and radiofrequency (RF) data were collected at 36 uncorrelated positions. Both attenuation and backscatter were measured over the frequency range 3.5 to 7.0 MHz at each min of a 30-min heating. Little change was observed in attenuation or backscatter at 55degreesC or less. The attenuation coefficient and backscatter power increased by factors of 1.25 and 5, respectively, during the 60degreesC heating. During the 65degreesC heating, the same properties showed increases by factors of 2.7 and 9. (E-mail: sherar@oci.utoronto.ca) (C) 2002 World Federation for Ultrasound in Medicine Biology.

Changes in the ultrasound (US) attenuation and backscatter of fresh pig kidney were measured as the tissue was heated. The objective was to use these changes to predict how an US image mould change in real-time with a view to its use as a monitoring tool for minimally invasive thermal therapy (MTTT), Separate samples of fresh pig kidney were heated from 37 degreesC to temperatures of 45 degrees, 50 degrees, 55 degrees, 60 degrees and 65 degrees with warm water. Measurements were made over the frequency range from 3.5 MHz to 7.0 MHz during 30-min heating experiments. A general increase in attenuation magnitude (dB/cm) and slope (dB/cm-MHz) was observed at temperatures of 55 degreesC or greater. Little change in backscatter power was observed during heating to 45 degreesC. At higher temperatures, the changes in backscatter showed a more complex pattern throughout the experiments, but still showed a trend of increase to a greater value at the end of heating than at the start. This backscatter increase was greater at higher temperatures, The net effect of the changes in US properties suggests that it may be possible to use diagnostic US to monitor, in real-time, MITT in kidney. (E-mail: sherar@oci.utoronto.ca) (C) 2001 World Federation for Ultrasound in Medicine & Biology.

Temperature distributions measured during thermal therapy are a major prognostic factor of the efficacy and success of the procedure. Thermal models are used to predict the temperature elevation of tissues during heating. Theoretical work has shown that blood flow through large blood vessels plays an important role in determining temperature profiles of heated tissues. In this paper, an experimental investigation of the effects of large vessels on the temperature distribution of heated tissue is performed. The blood flow dependence of steady state and transient temperature profiles created by a cylindrical conductive heat source and an ultrasound transducer were examined using a fixed porcine kidney as a flow model. In the transient experiments, a 20 s pulse of hot water, 30 degrees C above ambient, heated the tissues. Temperatures were measured at selected locations in steps of 0.1 mm. It was observed that vessels could either heat or cool tissues depending on the orientation of the vascular geometry with respect to the heat source and that these effects are a function of flow rate through the vessels. Temperature gradients of 6 degrees C mm(-1) close to large vessels were routinely measured. Furthermore, it was observed that the temperature gradients caused by large vessels depended on whether the heating source was highly localized (i.e. a hot needle) or more distributed (i.e. external ultrasound). The gradients measured near large vessels during localized heating were between two and three times greater than the gradients measured during ultrasound heating at the same location, for comparable flows. Moreover, these gradients were more sensitive to flow variations for the localized needle heating. X-ray computed tomography data of the kidney vasculature were in good spatial agreement with the locations of all of the temperature variations measured. The three dimensional vessel path observed could account for the complex features of the temperature profiles. The flow dependences of the transient temperature profiles near large vessels during the pulsed experiments were consistent with the temperature distributions measured in the steady state experiments and provided unique insights into the process of convective heat transfer in tissues. Finally, it was shown that even for very short treatment times (3-20 s), large vessels had significant effects on the tissue temperature distributions.

Thermal models are used to predict temperature distributions of heated tissues during thermal therapies. Recent interest in short duration high temperature therapeutic procedures necessitates the accurate modelling of transient temperature profiles in heated tissues. Blood flow plays an important role in tissue heat transfer and the resultant temperature distribution. This work examines the transient predictions of two simple mathematical models of heat transfer by blood flow (the bioheat transfer equation model and the effective thermal conductivity equation model) and compares their predictions to measured transient temperature data. Large differences between the two models are predicted in the tissue temperature distribution as a function of blood flow for a short heat pulse. In the experiments a hot water needle, approximately 30 degrees C above ambient, delivered a 20 s heating pulse to an excised fixed porcine kidney that was used as a flow model. Temperature profiles of a thermocouple that primarily traversed the kidney cortex were examined. Kidney locations with large vessels were avoided in the temperature profile analysis by examination of the vessel geometry using high resolution computed tomography angiography and the detection of the characteristic large vessel localized cooling or heating patterns in steady-state temperature profiles. It was found that for regions without large vessels, predictions of the Pennes bioheat transfer equation were in much better agreement with the experimental data when compared to predictions of the scalar effective thermal conductivity equation model. For example, at a location r approximately 2 mm away from the source, the measured delay time was 10.6 +/- 0.5 s compared to predictions of 9.4 s and 5.4 s of the BHTE and ETCE models, respectively. However, for the majority of measured locations, localized cooling and heating effects were detected close to large vessels when the kidney was perfused. Finally, it is shown that increasing flow in regions without large vessels minimally perturbs temperature profiles for short exposure times; regions with large vessels still have a significant effect.

The objective of this work was to image liver tissue heated to temperatures below the vaporization threshold as a function of time, to test the feasibility of real-time ultrasound monitoring to control lesion size during minimally invasive thermal therapy (MITT). Two experiments were devised. In one experiment, a thermal gradient was established in a rectangular volume of tissue to correlate changes in ultrasound image echogenicity (B-mode image brightness) with tissue temperature. In the other, a thermal lesion was produced in a rectangular volume of tissue by an interstitial microwave antenna, and the progression of the lesion was monitored by ultrasound. In both experiments, the echogenicity of the tissue increased slightly for tissue temperatures up to 40°C, but became lower than that of unheated tissue for temperatures above 40°C. In the second experiment, images of the lesion were compared with a photograph of the lesion taken after the experiment was complete. The final lesion was composed of two concentric regions—an inner region of heavily coagulated tissue and an outer region of less-damaged tissue. These two damaged regions indicated that increased ultrasound attenuation was largely responsible for the decreased echogenicity observed in the ultrasound images, and the increase in echogenicity of tissue heated to temperatures up to 40°C is thought to be due to decreased ultrasound attenuation at these temperatures.

Minimally invasive surgery by intense focused ultrasound beams producing defined lesions is being studied extensively by different groups. Lesion formation from a single pulse, depending on treatment time, tissue temperature, and pulse repetition of about 1 minute, should produce little damage near the skin. However, this scheme results in unacceptably long treatment times when used on larger tumors. A possible solution is to generate more rapid treatment times, or larger lesion volumes per pulse. However, hyperthermic temperatures in the overlying normal tissues including the skin may limit these treatments. In a previous presentation, simulations using an "ideal" transducer, pulses as short as 4 s and rapid stirring of the coupling bolus would reduce the temperature rise near the skin. Thus pulse repetitions as short of 10 s would be acceptable. However, real transducer beams show large aberrations which can greatly increase the near-field intensities, and make them unacceptable for hyperthermia therapy. Some artifacts are be caused by clamping of the transducer, others are related to thickness variations of the transducers which generate heterogenous phase shifts from different parts of the transducer which produce unwanted spreads at beam's focus. The authors present detailed amplitude and phase scans near different transducers demonstrating the artifacts, and confirm them using novel ultrasound/magnetic-resonance phantoms showing the measured temperatures at the focus, and at 1 cm depth from the "skin" where the heating is considerably larger than that predicted by theory. Finally, we will discuss solutions for problems in the near field by improving the transducer mounting and reducing the unwanted phase shifts.

To investigate the role of arterial occlusion on temperature homogeneity during hyperthermia for deep seated tissue, a renal hyperthermia animal model has been established using New Zealand white rabbits. The effects of ultrasound-induced renal hyperthermia, with or without continuous and intermittent renal artery occlusion, were compared and analysed. Both continuous and intermittent occlusion showed certain protection of surrounding:tissue and demonstrated improved temperature homogeneity and heating efficiency. The benefits of continuous vs. intermittent occlusion are compared and discussed as well.

Intense ultrasound beams have the potential for treating malignant tumours when combined with sonodynamic sensitizers. Some of these agents, e.g. porphyrins, are also used for photodynamic therapy. However, the experimental evidence is inconsistent. This work attempts to discover if the yield of free-radicals such as .OH and .H which are produced by transient cavitation could explain the killing of Chinese hamster ovary (CHO) cells in vitro with and without sonodynamic agents. CHO cells were irradiated with ultrasound beams in phosphate buffered saline [PBS] or in growth medium, and the immediate cell lysis and loss of cell colony forming ability measured. Under the authors' specific conditions, in which the standing wave patterns were minimized, a general correlation was observed between the transient cavitation, free-radical production, and cytotoxicity. However, the yield of free-radicals was much too small to explain the cell killing observed. The authors conclude that cytotoxicity is not linked to attack from free-radicals formed outside the cells. In their experiments, immediate cell lysis is closely linked to the transient cavitation which is known to produce shear forces which disrupt cellular membranes. The authors hypothesize that the loss of cell colony forming ability is also linked to damage of cellular membranes. In further experiments, CHO cells were irradiated with ultrasound beams in different media with and without hematoporphyrin (Hp) to ask the question: is there enough toxic singlet oxygen, formed from energy transfer from the excited Hp to oxygen, to explain the cytotoxicity? The addition of Hp to CHO cell insonations did not change the amount of immediate lysis or cytotoxicity as measured by colony growth. During Hp experiments under conditions which mimic a real PDT treatment, singlet oxygen fluorescence was produced by laser excitation but none was produced by the ultrasound beam. The authors conclude that the cytotoxicity is not linked to free-radical attack, or singlet oxygen, but hypothesize that the cytotoxicity is due to the transient cavitation producing shear forces which disrupt cellular membranes.

Significant differences in the backscatter amplitudes which are correlated with different tissue morphology have been observed in ultrasound images of tissue. While many factors could be linked to subtle changes in the images, the purpose of this paper is to explore the possibility that backscatter signals are linked to the organization of the spatial distribution of individual cells that produce an ensemble of scattering sources. Simple one- and two-dimensional simulations of backscatter signals produced by weak scatters separated by << lambda to < lambda in regular, random, and pseudo-random distributions in a "sample" are performed. Both regular and pseudo-random distributions produce large boundary signals, and in the central regions of the sample, the square root of the backscatter power is directly related to the amount of randomization, R, over a large range. Large changes in backscattering intensities are predicted for the same density of scatterers with differing R in different regions of the same sample. Thus, the subtle differences in the scattering distribution should show significant changes in the backscatter images.

Actual thermal gradients in perfused tissues are difficult to observe using thermocouples because of thermal conduction along the probes. We have used fine type-K (chromel-alumel) probes, which have a much lower thermal conductivity than equivalent-sized type-T (copper-constantan) thermocouples, to examine thermal gradients in two mouse tumour systems during water bath heating. The results indicate substantial heterogeneity in temperature distribution even in tumours transplanted in the foot and immersed to a depth of 2 cm in a 44 degrees C water bath for 20 min, i.e. thermal gradients greater than 1 degree C/mm were observed in KHT fibrosarcomas. The temperature heterogeneity for water bath heating is primarily a result of blood flow and appears to be tumour-specific. Temperature measurements using an excised perfused canine kidney demonstrate that increased perfusate volume flow increases the range of tissue temperatures. Consistent with theory, an artifactual improvement in temperature homogeneity resulted when temperature was measured using type-T thermocouples instead of type-K probes. These results emphasize the difficulties in obtaining accurate temperature measurements during experimental and clinical hyperthermia. Even extensive measurements of temperature in tissues may underestimate the true range of heterogeneity unless factors such as thermal smearing are controlled.