Institute of Fundamental Technological Research

A new detailed dataset of breast ultrasound scans (BrEaST) containing images of benign and malignant lesions as well as normal tissue examples, is presented. The dataset consists of 256 breast scans collected from 256 patients. Each scan was manually annotated and labeled by a radiologist experienced in breast ultrasound examination. In particular, each tumor was identified in the image using a freehand annotation and labeled according to BIRADS features and lexicon. The histopathological classification of the tumor was also provided for patients who underwent a biopsy.
The BrEaST dataset is the first breast ultrasound dataset containing patient-level labels, image-level annotations, and tumor-level labels with all cases confirmed by follow-up care or core needle biopsy result. To enable research into breast disease detection, tumor segmentation and classification, the BrEaST dataset is made publicly available with the CC-BY 4.0 license.

The HIFU ablation technique is limited by the long duration of the procedure, which results from the large difference between the size of the HIFU beam’s focus and the tumor size. Ablation of large tumors requires treating them with a sequence of single HIFU beams, with a specific time interval in-between. The aim of this study was to evaluate the biological effects induced in a malignant solid tumor of the rat mammary gland, implanted in adult Wistar rats, during HIFU treatment according to a new ablation plan which allowed researchers to significantly shorten the duration of the procedure. We used a custom, automated, ultrasound imaging-guided HIFU ablation device. Tumors with a 1 mm thickness margin of healthy tissue were subjected to HIFU. Three days later, the animals were sacrificed, and the HIFU-treated tissues were harvested. The biological effects were studied, employing morphological, histological, immunohistochemical, and ultrastructural techniques. Massive cell death, hemorrhages, tissue loss, influx of immune cells, and induction of pro-inflammatory cytokines were observed in the HIFU-treated tumors. No damage to healthy tissues was observed in the area surrounding the safety margin. These results confirmed the efficacy of the proposed shortened duration of the HIFU ablation procedure and its potential for the treatment of solid tumors.

HIFU thermal ablation, breast cancer model, treatment plan, morphology, histology, ultrastructure, immune response, cell death, apoptosis, necrosis

High Intensity Focused Ultrasound (HIFU) is used in clinical practice for thermal ablation of malignant and benign solid tumors located in various organs. One of the reason limiting the wider use of this technology is the long treatment time resulting from i.a. the large difference between the size of the focal volume of the heating beam and the size of the tumor. Therefore, the treatment of large tumors requires scanning their volume with a sequence of single heating beams, the focus of which is moved in the focal plane along a specific trajectory with specific time and distance interval between sonications. To avoid an undesirable increase in the temperature of healthy tissues surrounding the tumor during scanning, the acoustic power and exposure time of each HIFU beam as well as the time intervals between sonications should be selected in such a way as to cover the entire volume of the tumor with necrosis as quickly as possible. This would reduce the costs of treatment. The aim of this study was to quantitatively evaluate the hypothesis that selecting the average acoustic power and exposure time for each individual heating beam, as well as the temporal intervals between sonications, can significantly shorten treatment time. Using 3D numerical simulations, the dependence of the duration of treatment of a tumor with a diameter of 5 mm or 9 mm (requiring multiple exposure to the HIFU beam) on the sonication parameters (acoustic power, exposure time) of each single beam capable of delivering the threshold thermal dose (CEM43 = 240 min) to the treated tissue volume was examined. The treatment duration was determined as the sum of exposure times to individual beams and time intervals between sonications. The tumor was located inside the ex vivo tissue sample at a depth of 12.6 mm. The thickness of the water layer between the HIFU transducer and the tissue was 50 mm. The sonication and scanning parameters selected using the developed algorithm shortened the duration of the ablation procedure by almost 14 times for a 5-mm tumor and 20 times for a 9-mm tumor compared to the duration of the same ablation plan when a HIFU beam was used of a constant acoustic power, constant exposure time (3 s) and constant long time intervals (120 s) between sonications. Results of calculations of the location and size of the necrotic lesion formed were experimentally verified on ex vivo pork loin samples, showing good agreement between them. In this way, it was proven that the proper selection of sonication and scanning parameters for each HIFU beam allows to significantly shorten the time of HIFU therapy.

HIFU ablation planning,HIFU therapy duration shortening,Tissue ex vivo,k-wave model,Experimental verification of therapy accuracy,Numerical simulation

We have designed and built low-cost compact ultrasound imaging-guided robotic HIFU (High-Intensity Focused Ultrasound) ablation device for thermal damage of solid tumors in small animals. Before this device is used to treat animals, experimental studies on ex vivo tissues were necessary to assess the accuracy of its targeting, ensuring the safety of therapy. The objective of this study was to assess the targeting accuracy of our device in the focal and axial plane of the HIFU beam using ex vivo tissue embedded in a reference cylindrical chamber inside which a coaxial cylindrical volume with a smaller diameter was ablated. HIFU beams with selected acoustic parameters, generated by a singe-element bowl-shaped 64-mm HIFU transducer operating at 1.08 MHz or 3.21 MHz frequency, were propagated in two-layer media: water-tissue (50 mm-40 mm) and focused at 12.6-mm depth below the tissue surface. Cylindrical necrotic lesions of various size were created by moving the chamber using a computercontrolled precise positioning unit. Lesions formed were compared with those intended for treatment using various visualization methods and displacement between their centers were determined. The targeting accuracy in the focal and axial planes was found to be respectively about 98% and 86% when determined from photos and about 88% and 76% when determined from MR images. The displacement between the centers of the necrotic lesion formed and planned for treatment was about 1 mm in the focal plane and about 2 mm in the axial plane. Our ablation device can be used as an effective and safe tool to plan, monitor and treat solid tumors in small animals and to test new anti-cancer drugs in preclinical studies.

We have designed and built ultrasound imaging-guided HIFU ablative device for preclinical studies on small animals. Before this device is used to treat animals, ex vivo tissue studies were necessary to determine the location and extent of necrotic lesions created inside tissue samples by HIFU beams depending on their acoustic properties. This will allow to plan the beam movement trajectory and the distance and time intervals between exposures leading to necrosis covering the entire treated volume without damaging the surrounding tissues. This is crucial for therapy safety. The objective of this study was to assess the impact of sonication parameters on the size of necrotic lesions formed by HIFU beams generated by 64-mm bowl-shaped transducer used, operating at 1.08 MHz or 3.21 MHz. Multiple necrotic lesions were created in pork loin samples at 12.6-mm depth below tissue surface during 3-s exposure to HIFU beams with fixed duty-cycle and varied pulse-duration or fixed pulse-duration and varied duty-cycle, propagated in two-layer media: water-tissue. After exposures, the necrotic lesions were visualized using magnetic resonance imaging and optical imaging (photos) after sectioning the samples. Quantitative analysis of the obtained results allowed to select the optimal sonication and beam movement parameters to suport planning of effective therapy.

automated ultrasound imaging-guided HIFU ablation system, ex vivo tissue, ultrasonic exposure parameters, extent of necrotic lesions

Image-guided High Intensity Focused Ultrasound (HIFU) technique is dynamically developing technology for treating solid tumors due to its non-invasive nature. Before a HIFU ablation system is ready for use, the exposure parameters of the HIFU beam capable of destroying the treated tissue without damaging the surrounding tissues should be selected to ensure the safety of therapy. The purpose of this work was to select the threshold acoustic power as well as the step and rate of movement of the HIFU beam, generated by a transducer intended to be used in the HIFU ablation system being developed, by using an array of thermocouples and numerical simulations. For experiments a bowl-shaped 64-mm, 1.05 MHz HIFU transducer with a 62.6 mm focal length (f-number 0.98) generated pulsed waves propagating in two-layer media: water/ex vivo pork loin tissue (50 mm/40 mm) was used. To determine a threshold power of the HIFU beam capable of creating the necrotic lesion in a small volume within the tested tissue during less than 3 s each tissue sample was sonicated by multiple parallel HIFU beams of different acoustic power focused at a depth of 12.6 mm below the tissue surface. Location of the maximum heating as well as the relaxation time of the tested tissue were determined from temperature variations recorded during and after sonication by five thermo-couples placed along the acoustic axis of each HIFU beam as well as from numerical simulations. The obtained results enabled to assess the location of each necrotic lesion as well as to determine the step and rate of the HIFU beam movement. The location and extent of the necrotic lesions created was verified using ultrasound images of tissue after sonication and visual inspection after cutting the samples. The threshold acoustic power of the HIFU beam capable of creating the local necrotic lesion in the tested tissue within 3 s without damaging of surrounding tissues was found to be 24 W, and the pause between sonications was found to be more than 40 s.

automated HIFU ablation system, threshold acoustic power of HIFU beam, ex vivo tissue, necrotic lesion, thermocouple array

Manual image segmentations are naturally subject to inaccuracies related to systematic errors (due to the tools used, eye-hand coordination, etc.). This was noted earlier when a simplified accuracy scale was proposed [1]. This scale arbitrarily divides a given range of values of the Kappa measurement parameter into classes: almost perfect (>0.80), substantial (0.61 - 0.80), moderate (0.41 - 0.60), fair (0.21 - 0.40), slight (0.00 - 0.21) and poor (< 0.00). However, the determination of threshold values between classes is not entirely clear and seems to be application-dependent. This is particularly important for images in which the tumor-normal tissue boundary can be very indistinct, as is observed in ultrasound imaging of the most common cancer in women - breast cancer [2]. In machine learning, there is an ongoing contest over the values of performance indicators obtained from new neural network architecture without accounting for any ground truth bias. This raises the question of what relevance, from a segmentation quality point of view, a gain at the level of single percentages has [3] if the references have much greater uncertainty. So far, research on this topic has been limited. The relationship between the segmentations of breast tumors on ultrasound images provided by three radiologists and those obtained using deep learning model has been studied in [4]. Unfortunately, the indicated segmentation contour sometimes varied widely in all three cases. A cursory analysis by multiple physicians, which focused only on the Kappa coefficient in the context of physicians’ BI-RADS category assignments, was conducted in the [5]. In this article, we present a preliminary analysis of the accuracy of experts’ manually prepared binary breast cancer masks on ultrasound images and their impact on performance metrics commonly used in machine learning. In addition, we examined how tumor type or BI-RADS category [6] affects the accuracy of tumor contouring.

Standard classification methods based on hand-crafted morphological and texture features have achieved good performance in breast mass differentiation in ultrasound (US).
In comparison to deep neural networks, commonly perceived as ‘black-box’ models, classical techniques are based on features that have well-understood medical and physical interpretation. However, classifiers based on morphological features commonly
underperform in the presence of the shadowing artifact and ill-defined mass borders, while texture based classifiers may fail when the US image is too noisy. Therefore, in practice it would be beneficial to select the classification method based on the appearance of the particular US image. In this work, we develop a deep meta-network that can automatically process input breast mass US images and recommend whether to apply the shape or
texture based classifier for the breast mass differentiation. Our preliminary results demonstrate that meta-learning techniques can be used to improve the performance of the standard classifiers based on handcrafted features. With the proposed meta-learning based approach, we achieved the area under the receiver operating characteristic curve of 0.95 and accuracy of 0.91.

breast mass classification, deep learning, meta-learning, morphological features, texture features

Local ultrasonic tissue ablation is induced by a rapid (<3s) rise in temperature in a small ellipsoidal volume (about 13mm3) inside the tissue to a cytotoxic level when exposed to a high-intensity focused ultrasound (HIFU) beam. The aim of this study was to develop a numerical tool to predict the location and extent of a necrotic lesion formed locally inside the ex vivo tissue as a result of exposure to a single or multiple HIFU beam, ensuring the efficacy and safety of destroying solid tumors. The proposed tool was based on modelling the non-linear propagation of acoustic waves and heat transfer in heterogeneous media using the k-wave toolbox. The wave propagation equations were solved for two-layer (water/tissue) media. The source of the acoustic waves was a spherical bowl-shaped transducer with a resonance frequency of 1.08 MHz. The distribution of heat sources was determined from the calculated acoustic pressure distribution in the HIFU beam. The obtained temperature distributions during heating and cooling allowed calculation of the thermal dose and prediction of the extent of the necrotic lesion. The obtained results of numerical simulations were compared with the experimental data from previous studies. The mean difference between the calculated and measured length or diameter of a single exposure induced necrotic lesion was approximately 1 mm. In the case of a necrotic lesion induced by multiple exposures, the mean difference between the measured and calculated cross-sectional area of the planned necrotic lesion covered with necrosis was approximately 11.2 %.

HIFU, modelling, numerical simulations, treatment planning, ex vivo experiments, necrotic lesions, thermal ablation, experimental verification