A Data-Driven Approach for Estimating Temperature Variations Based on B-mode Ultrasound Images and Changes in Backscattered Energy.

Thermal treatments that use ultrasound devices as a tool have as a key point the temperature control to be applied in a specific region of the patient's body. This kind of procedure requires caution because the wrong regulation can either limit the treatment or aggravate an existing injury. Therefore, determining the temperature in a region of interest in real-time is a subject of high interest. Although this is still an open problem, in the field of ultrasound analysis, the use of machine learning as a tool for both imaging and automated diagnostics are application trends. In this work, a data-driven approach is proposed to address the problem of estimating the temperature in regions of a B-mode ultrasound image as a supervised learning problem. The proposal consists in presenting a novel data modeling for the problem that includes information retrieved from conventional B-mode ultrasound images and a parametric image built based on changes in backscattered energy (CBE). Then, we compare the performance of classic models in the literature. The computational results presented that, in a simulated scenario, the proposed approach that a Gradient Boosting model would be able to estimate the temperature with a mean absolute error of around 0.5°C, which is acceptable in practical environments both in physiotherapic treatments and high intensity focused ultrasound (HIFU).

Quadratic versus linear models to estimate the mean scattering spacing as a function of temperature in ex-vivo tissue.

Previous works have shown the feasibility of temperature estimation during ultrasonic therapy using pulse-echo diagnostic ultrasound. These methods are based on the measurement of thermally induced changes in backscattered RF echoes due to thermal expansion and changes in ultrasonic velocity. They assume a joint contribution of these two parameters and a linear dependence with temperature. In this work, the contributions of velocity changes and thermal expansion to the evolution of the mean scatterer spacing of ex vivo bovine skeletal muscle tissue samples were decoupled. This was achieved by employing an experimental setup which allows measuring the absolute velocity value, using the through-transmission technique in a direct transmission configuration. The mean-scatterer spacing was estimated from spectral analysis of the backscattered signals obtained in pulse-echo mode. We propose a quadratic model of the thermal expansion coefficient to fit the evolution of the mean-scatterer spacing with temperature. The temperature increase estimated by the linear model, in the range of 29.5-47 °C, presents a percentage error (mean square error) of 11 %, while for the quadratic model the error is 4.8 %.

Method for estimating average grey-level's measurement uncertainty from ultrasound images for non-invasive estimation of temperature in different tissue types.

The objective of this work is to assess, on metrological basis, the average grey-levels (AVGL) calculated from B-Mode images for estimating temperature variations non-invasively in different kinds of tissues. Thermal medicine includes several thermal therapies, being hyperthermia the most noted and well known. Recently, efforts have been made to understand the benefits of ultrasound hyperthermia at mild temperature levels, i.e., between 39 °C and 41 °C. Moreover, the best practices on ultrasound bio-effects research have been encouraged by recommending that temperature rise in the region of interest should be measured even when a thermal mechanism is not being tested. In this work, the average grey-levels (AVGL) calculated from B-Mode images were assessed for non-invasive temperature estimation in a porcine tissue sample containing two different tissue types, fat and muscle, with temperature varying from 35 °C to 41 °C. The sample was continuously imaged with an ultrasound scanner, and simultaneously the temperature was measured. The achieved results were assessed under the light of the measurement uncertainty in order to allow comparability among different ultrasound thermometry methods. The highest expanded uncertainty of estimating temperature variation using AVGL was determined as 0.68 °C.

Durability study of a gellan gum-based tissue-mimicking phantom for ultrasonic thermal therapy.

Stability and duration of ultrasonic phantoms are still subjects of research. This work presents a tissue-mimicking material (TMM) to evaluate high-intensity therapeutic ultrasound (HITU) devices, composed of gellan gum (matrix), microparticles (scatterers), and chemicals. The ultrasonic velocity and attenuation coefficient were characterized as a function of temperature (range 20 °C-85 °C). The nonlinear parameter B/A was determined by the finite amplitude insertion substitution (FAIS) method, and the shear modulus was determined by a transient elastography technique. The thermal conductivity and specific heat were determined by the line source method. The attenuation was stable for 60 days, and in an almost linear frequency dependence (0.51f0.96 dB cm-1), at 20 °C (1-10 MHz). All other evaluated physical parameters are also close to typical soft tissue values. Longitudinal ultrasonic velocities were between 1.49 and 1.75 mm µs-1, the B/A parameter was 7.8 at 30 °C, and Young's modulus was 23.4 kPa. The thermal conductivity and specific heat values were 0.7 W(m K)-1 and 4.7 kJ(kg K)-1, respectively. Consistent temperature increases and thermal doses occurred under identical HITU exposures. Low cost, longevity, thermal stability, and thermal repeatability make TMM an excellent material for ultrasonic thermal applications. The TMM developed has the potential to assess the efficacy of hyperthermia devices and could be used to adjust the ultrasonic emission of HITU devices.

Ex vivo determined experimental correction factor for the ultrasonic source term in the bioheat equation.

The objective of this work is to propose an effective absorption coefficient (αeffec) as an empirical correction factor in the source term of the bioheat equation. The temperature rise in biological tissue due to ultrasound insonification is produced by energy absorption. Usually, the ultrasonic absorption coefficient (αA) is used as a source term in the bioheat equation to quantify the temperature rise, and the effect of scattering is disregarded. The coefficient αeffec includes the scattering contribution as an additional absorption term and should allow us to make a better estimation of the thermal dose (TD), which is important for clinical applications. We simulated the bioheat equation with the source term considering αA or αeffec, and with heating provided by therapeutic ultrasound (1MHz, 2.0Wcm-2) for about 5.5min (temperature range 36-46°C). Experimental data were obtained in similar heating conditions for a bovine muscle tissue (ex vivo) and temperature curves were measured for depths 7, 30, 35, 40 and 45mm. The TD values from the experimental temperature curves at each depth were compared with the numerical solution of the bioheat equation with the classical and corrected source terms. The highest percentual difference between simulated and experimental TD was 42.5% when assuming the classical αA, and 8.7% for the corrected αeffec. The results show that the effective absorption coefficient is a feasible parameter to improve the classical bioheat transfer model, especially for depths larger than the mean free propagation path.

Assessing heating distribution by therapeutic ultrasound on bone phantoms and in vitro human samples using infrared thermography.

BACKGROUND: Bioheat models have been proposed to predict heat distribution in multilayered biological tissues after therapeutic ultrasound (TUS) stimulation. However, evidence on its therapeutic benefit is still controversial for many clinical conditions. The aim of this study was to evaluate and to compare the TUS heating distribution on commercially available bone phantoms and in vitro femur and tibia human samples, at 1 MHz and several ultrasonic pulse regimens, by means of a thermographic image processing technique. METHODS: An infrared camera was used to capture an image after each 5-min 1-MHz TUS stimulation on bone phantoms, as well as in vitro femur and tibia samples (N = 10). An intensity-based processing algorithm was applied to estimate temperature distribution. Sections of five femurs in the coronal plane were also used for the evaluation of heat distribution inside the medullar canal. RESULTS: Temperature increased up to 8.2 and 9.8 °C for the femur and tibia, respectively. Moreover, the temperature increased up to 10.8 °C inside the medullar canal. Although temperature distributions inside the region of interest (ROI) were significantly different (p < 0.001), the average and standard deviation values for bone phantoms were more similar to the femur than to the tibia samples. Pulsed regimens caused lower increments in temperature than continuous sonication, as expected. CONCLUSIONS: Commercially available bone phantoms could be used in research focusing on thermal effects of ultrasound. Small differences in mean and standard deviation temperatures were observed between bone samples and phantoms. Temperature can reach more than 10 °C inside the medullar canal on a fixed probe position which may lead to severe cellular damage.

Thermochromic Phantom and Measurement Protocol for Qualitative Analysis of Ultrasound Physiotherapy Systems.

Thermochromic test bodies are promising tools for qualitatively evaluating the acoustic output of ultrasound physiotherapy systems. Here, a novel phantom, made of silicone mixed with thermochromic powder material, was developed. Additionally, a procedure was developed to evaluate the stability and homogeneity of the phantom in a metrologic and statistical base. Twelve phantoms were divided into three groups. Each group was insonated by a different transducer. An effective intensity of 1.0 W/cm(2) was applied to each phantom; two operators performed the procedure three times in all phantoms. The heated area was measured after image processing. No statistical difference was observed in the heated areas for different samples or in the results for different operators. The heated areas obtained using each transducer were statistically different, indicating that the thermochromic phantom samples had sufficient sensitivity to represent the heated areas of different ultrasonic transducers. Combined with the evaluation procedure, the phantom provides an approach not previously described in the literature. The proposed approach can be used to quickly assess changes in ultrasonic beam cross-sectional shape during the lifetime of ultrasound physiotherapy systems.

Influence of ultrasonic scattering in the calculation of thermal dose in ex-vivo bovine muscular tissues.

This study explores the effect of ultrasound scattering on the temperature increase in phantoms and in samples of ex-vivo biological tissue through the calculation of the thermal dose (TD). Phantoms with different weight percentages of graphite powder (0-1%w/w, different scattering mean free paths, ℓS) and ex-vivo bovine muscle tissue were isonified by therapeutic ultrasound (1 MHz). The TD values were calculated from the first 4 min of experimental temperature curves obtained at several depths and were compared with those acquired from the numerical solution of the bio-heat transfer equation (simulated with 1 MHz and 0.5-2.0 W cm(-2)). The temperature curves suggested that scattering had an important role because the temperature increments were found to be higher for higher percentages of graphite powder (lower ℓS). For example, at a 30-mm depth and a 4-min therapeutic ultrasound application (0.5 W cm(-2)), the TDs (in equivalent minutes at 43 °C) were 7.2, 17.8, and 58.3 for the phantom with ℓS of 4.35, 3.85, and 3.03 mm, respectively. In tissue, the inclusion of only absorption or full attenuation in the bio-heat transfer equation (BHTE) heat source term of the simulation leads to under- or overestimation of the TD, respectively, as compared to the TD calculated from experimental data. The experiments with phantoms (with different scatterer concentrations) and ex-vivo samples show that the high values of TD were caused by the increase of energy absorption due to the lengthening of the propagation path caused by the changing in the propagation regime.

Evaluating geodesic active contours in microcalcifications segmentation on mammograms.

Breast cancer is the most commonly occurring type of cancer among women, and it is the major cause of female cancer-related deaths worldwide. Its incidence is increasing in developed as well as developing countries. Efficient strategies to reduce the high death rates due to breast cancer include early detection and tumor removal in the initial stages of the disease. Clinical and mammographic examinations are considered the best methods for detecting the early signs of breast cancer; however, these techniques are highly dependent on breast characteristics, equipment quality, and physician experience. Computer-aided diagnosis (CADx) systems have been developed to improve the accuracy of mammographic diagnosis; usually such systems may involve three steps: (i) segmentation; (ii) parameter extraction and selection of the segmented lesions and (iii) lesions classification. Literature considers the first step as the most important of them, as it has a direct impact on the lesions characteristics that will be used in the further steps. In this study, the original contribution is a microcalcification segmentation method based on the geodesic active contours (GAC) technique associated with anisotropic texture filtering as well as the radiologists' knowledge. Radiologists actively participate on the final step of the method, selecting the final segmentation that allows elaborating an adequate diagnosis hypothesis with the segmented microcalcifications presented in a region of interest (ROI). The proposed method was assessed by employing 1000 ROIs extracted from images of the Digital Database for Screening Mammography (DDSM). For the selected ROIs, the rate of adequately segmented microcalcifications to establish a diagnosis hypothesis was at least 86.9%, according to the radiologists. The quantitative test, based on the area overlap measure (AOM), yielded a mean of 0.52±0.20 for the segmented images, when all 2136 segmented microcalcifications were considered. Moreover, a statistical difference was observed between the AOM values for large and small microcalcifications. The proposed method had better or similar performance as compared to literature for microcalcifications with maximum diameters larger than 460µm. For smaller microcalcifications the performance was limited.

Assessing the combined performance of texture and morphological parameters in distinguishing breast tumors in ultrasound images.

PURPOSE: This work aims to investigate the combination of morphological and texture parameters in distinguishing between malignant and benign breast tumors in ultrasound images. METHODS: Linear discriminant analysis was applied to sets of up to five parameters, and then the performances were assessed using the area A(z) (± standard error) under the receiver operator characteristic curve, accuracy (Ac), sensitivity (Se), specificity (Sp), positive predictive value, and negative predictive value. RESULTS: The most relevant individual parameter was the normalized residual value (nrv), calculated from the convex polygon technique. The best performance among all studied combinations was achieved by two morphological and three texture parameters (nrv, con, std, R, and asm(i)), which correctly distinguished nearly 85% of the breast tumors. CONCLUSIONS: This result indicates that the combination of morphological and texture parameters may be useful to assist physicians in the diagnostic process, especially if it is associated with an automatic classification tool.

Study of superficial basal cell carcinomas and Bowen disease by qualitative and quantitative ultrasound biomicroscopy approach.

In the present work the ultrasound biomicroscopy (UBM) technique is applied in the study of cutaneous cell carcinomas in vitro, including superficial basal cell carcinomas (BCC) and Bowen disease (BD) cases. The evaluation was made by qualitative observation of UBM images, and by quantitative computation of integrated backscatter coefficient (IBC), obtained with a system working at a central frequency of 45 MHz. The characteristic histological structures for each studied tumor type were well identified in the images. The IBC values observed in the two carcinoma types inside the affected region, were different between them, next to 10(-4) [Sr(-1).mm(-1)] for superficial BCC tissues, and to 10(-5) [Sr(-1).mm(-1)1] for BD tissues; moreover, in the deeper dermis (slight affected region) the backscatter was next to 10(-3) [Sr(-1).mm(-1)] for both tissue groups, and agrees with the values obtained for healthy skin both, in this study and in previous works. The results here obtained encourage the continuation of the work, with a higher number of samples, attempting to obtain more significant results.

A cost simulation for mammography examinations taking into account equipment failures and resource utilization characteristics.

OBJECTIVE: This work develops a cost analysis estimation for a mammography clinic, taking into account resource utilization and equipment failure rates. MATERIALS AND METHODS: Two standard clinic models were simulated, the first with one mammography equipment, two technicians and one doctor, and the second (based on an actually functioning clinic) with two equipments, three technicians and one doctor. Cost data and model parameters were obtained by direct measurements, literature reviews and other hospital data. A discrete-event simulation model was developed, in order to estimate the unit cost (total costs/number of examinations in a defined period) of mammography examinations at those clinics. The cost analysis considered simulated changes in resource utilization rates and in examination failure probabilities (failures on the image acquisition system). In addition, a sensitivity analysis was performed, taking into account changes in the probabilities of equipment failure types. RESULTS: For the two clinic configurations, the estimated mammography unit costs were, respectively, US$ 41.31 and US$ 53.46 in the absence of examination failures. As the examination failures increased up to 10% of total examinations, unit costs approached US$ 54.53 and US$ 53.95, respectively. The sensitivity analysis showed that type 3 (the most serious) failure increases had a very large impact on the patient attendance, up to the point of actually making attendance unfeasible. CONCLUSIONS: Discrete-event simulation allowed for the definition of the more efficient clinic, contingent on the expected prevalence of resource utilization and equipment failures.

Influence of temperature variations on the entropy and correlation of the Grey-Level Co-occurrence Matrix from B-Mode images.

In this work, the feasibility of texture parameters extracted from B-Mode images were explored in quantifying medium temperature variation. The goal is to understand how parameters obtained from the gray-level content can be used to improve the actual state-of-the-art methods for non-invasive temperature estimation (NITE). B-Mode images were collected from a tissue mimic phantom heated in a water bath. The phantom is a mixture of water, glycerin, agar-agar and graphite powder. This mixture aims to have similar acoustical properties to in vivo muscle. Images from the phantom were collected using an ultrasound system that has a mechanical sector transducer working at 3.5 MHz. Three temperature curves were collected, and variations between 27 and 44 degrees C during 60 min were allowed. Two parameters (correlation and entropy) were determined from Grey-Level Co-occurrence Matrix (GLCM) extracted from image, and then assessed for non-invasive temperature estimation. Entropy values were capable of identifying variations of 2.0 degrees C. Besides, it was possible to quantify variations from normal human body temperature (37 degrees C) to critical values, as 41 degrees C. In contrast, despite correlation parameter values (obtained from GLCM) presented a correlation coefficient of 0.84 with temperature variation, the high dispersion of values limited the temperature assessment.

Optimizing patient flow in a large hospital surgical centre by means of discrete-event computer simulation models.

OBJECTIVE: This study used the discrete-events computer simulation methodology to model a large hospital surgical centre (SC), in order to analyse the impact of increases in the number of post-anaesthetic beds (PABs), of changes in surgical room scheduling strategies and of increases in surgery numbers. METHODS: The used inputs were: number of surgeries per day, type of surgical room scheduling, anaesthesia and surgery duration, surgical teams' specialty and number of PABs, and the main outputs were: number of surgeries per day, surgical rooms' use rate and blocking rate, surgical teams' use rate, patients' blocking rate, surgery delays (minutes) and the occurrence of postponed surgeries. Two basic strategies were implemented: in the first strategy, the number of PABs was increased under two assumptions: (a) following the scheduling plan actually used by the hospital (the 'rigid' scheduling - surgical rooms were previously assigned and assignments could not be changed) and (b) following a 'flexible' scheduling (surgical rooms, when available, could be freely used by any surgical team). In the second, the same analysis was performed, increasing the number of patients (up to the system 'feasible maximum') but fixing the number of PABs, in order to evaluate the impact of the number of patients over surgery delays. CONCLUSION: It was observed that the introduction of a flexible scheduling/increase in PABs would lead to a significant improvement in the SC productivity.

Neuro-genetic non-invasive temperature estimation: intensity and spatial prediction.

OBJECTIVES: The existence of proper non-invasive temperature estimators is an essential aspect when thermal therapy applications are envisaged. These estimators must be good predictors to enable temperature estimation at different operational situations, providing better control of the therapeutic instrumentation. In this work, radial basis functions artificial neural networks were constructed to access temperature evolution on an ultrasound insonated medium. METHODS: The employed models were radial basis functions neural networks with external dynamics induced by their inputs. Both the most suited set of model inputs and number of neurons in the network were found using the multi-objective genetic algorithm. The neural models were validated in two situations: the operating ones, as used in the construction of the network; and in 11 unseen situations. The new data addressed two new spatial locations and a new intensity level, assessing the intensity and space prediction capacity of the proposed model. RESULTS: Good performance was obtained during the validation process both in terms of the spatial points considered and whenever the new intensity level was within the range of applied intensities. A maximum absolute error of 0.5 degrees C+/-10% (0.5 degrees C is the gold-standard threshold in hyperthermia/diathermia) was attained with low computationally complex models. CONCLUSION: The results confirm that the proposed neuro-genetic approach enables foreseeing temperature propagation, in connection to intensity and space parameters, thus enabling the assessment of different operating situations with proper temperature resolution.

A soft-computing methodology for noninvasive time-spatial temperature estimation.

The safe and effective application of thermal therapies is restricted due to lack of reliable noninvasive temperature estimators. In this paper, the temporal echo-shifts of backscattered ultrasound signals, collected from a gel-based phantom, were tracked and assigned with the past temperature values as radial basis functions neural networks input information. The phantom was heated using a piston-like therapeutic ultrasound transducer. The neural models were assigned to estimate the temperature at different intensities and points arranged across the therapeutic transducer radial line (60 mm apart from the transducer face). Model inputs, as well as the number of neurons were selected using the multiobjective genetic algorithm (MOGA). The best attained models present, in average, a maximum absolute error less than 0.5 degrees C, which is pointed as the borderline between a reliable and an unreliable estimator in hyperthermia/diathermia. In order to test the spatial generalization capacity, the best models were tested using spatial points not yet assessed, and some of them presented a maximum absolute error inferior to 0.5 degrees C, being "elected" as the best models. It should be also stressed that these best models present implementational low-complexity, as desired for real-time applications.

Complexity curve and grey level co-occurrence matrix in the texture evaluation of breast tumor on ultrasound images.

This work aims at investigating texture parameters in distinguishing malign and benign breast tumors on ultrasound images. A rectangular region of interest (ROI) containing the tumor and its neighboring was defined for each image. Five parameters were extracted from the complexity curve (CC) of the ROI. Another five parameters were calculated from the grey-level co-occurrence matrix (GLCM) also for the ROI. The same was carried out for internal tumor region, hence, totaling 20 parameters. The linear discriminant analysis was applied to sets of up to five parameters and then the performances were assessed. The most relevant individual parameters were the contrast (con) (from the GLCM over the ROI) and the maximum value (mvi) from the CC just for the tumor internal region). When they were taken together, a correct classification slightly over 80% of the breast tumors was achieved. The highest performance (accuracy=84.2%, sensitivity=87.0%, and specificity=78.8%) was obtained with mvi, con, the standard deviation of the pixel pairs and the entropy, both for GLCM, and the internal region contrast also from GLCM. Parameters extracted from the internal region generally performed better and were more significant than those from the ROI. Moreover, parameters calculated only from CC or GLCM resulted in no statistically significant performance difference. These findings suggest that the texture parameters can be useful to help radiologist in distinguishing between benign or malign breast tumors on ultrasound images.

Singular spectrum analysis applied to backscattered ultrasound signals from in vitro human cancellous bone specimens.

Mean scatterer spacing (MSS) holds particular promise for the detection of changes in quasiperiodic tissue microstructures such as may occur during development of disease in the liver, spleen, or bones. Many techniques that may be applied for MSS estimation (temporal and spectral autocorrelation, power spectrum and cepstrum, higher order statistics, and quadratic transformation) characterize signals that contain a mixture of periodic and nonperiodic contributions. In contrast, singular spectrum analysis (SSA), a method usually applied in nonlinear dynamics, first identifies components of signals corresponding to periodic structures and, second, identifies dominant periodicity. Thus, SSA may better separate periodic structures from nonperiodic structures and noise. Using an ultrasound echo simulation model, we previously demonstrated SSA's potential to identify MSS of structures in quasiperiodic scattering media. The current work aims to observe the behavior of MSS estimation by SSA using ultrasound measurements in phantom materials (two parallel, nylon-line phantoms and four foam phantoms of different densities). The SSA was able to estimate not only the nylon-line distances but also nylon-line thickness. The method also was sensitive to the average pore-size differences of the four sponges. The algorithms then were applied to characterize human cancellous bone microarchitectures. Using 1-MHz center-frequency, radio-frequency ultrasound signals, MSS was measured in 24 in vitro bone samples and ranged from 1.0 to 1.7 mm. The SSA MSS estimates correlate significantly to MSS measured independently from synchrotron microtomography, r2 = 0.68. Thus, application of SSA to backscattered ultrasound signals seems to be useful for providing information linked to tissue microarchitecture that is not evident from clinical images.