Early prediction of response to neoadjuvant chemotherapy (NAC) in breast cancer is crucial for guiding therapy decisions. In this work, we propose a deep learning based approach the early NAC ultrasound (US) imaging. We used transfer with convolutional neural networks (CNNs) develop models. The usefulness two techniques was examined. First, CNN pre-trained on ImageNet dataset utilized. Second, applied double learning, additionally fine-tuned mass US images differentiate malignant and benign...

Abstract The presented studies evaluate for the first time efficiency of tumour classification based on quantitative analysis ultrasound data originating from tissue surrounding tumour. 116 patients took part in study after qualifying biopsy due to suspicious breast changes. RF signals collected and tumour-surroundings were processed determine measures consisting Nakagami distribution shape parameter, entropy, texture parameters. utility parameters benign malignant lesions was assessed...

Monitoring Neoadjuvant chemotherapy (NAC) effects is necessary to capture resistant patients and stop or change treatment. The aim of this study was assess the tumor response at an early stage, after first doses NAC, based on variability backscattered ultrasound energy, backscatter statistics. statistics has not previously been used monitor NAC effects. B-mode images raw radio frequency data from breast tumors were obtained using scanner before 1 week each cycle. Twenty-four malignant...

To investigate the performance of multiparametric ultrasound for evaluation treatment response in breast cancer patients undergoing neoadjuvant chemotherapy (NAC). The IRB approved this prospective study. Breast who were scheduled to undergo NAC invited participate Changes tumour echogenicity, stiffness, maximum diameter, vascularity and integrated backscatter coefficient (IBC) assessed prior 7 days after four consecutive cycles. Residual malignant cell (RMC) measurement at surgery was...

Deep neural networks have achieved good performance in breast mass classification ultrasound imaging. However, their usage clinical practice is still limited due to the lack of explainability decisions conducted by networks. In this study, address problem, we generated saliency maps indicating image regions important for network's decisions.Ultrasound images were collected from 272 masses, including 123 malignant and 149 benign. Transfer learning was applied develop a deep network...

Attenuation of ultrasound in tissue can be estimated from the propagating pulse center frequency downshift. This method assumes that envelope emitted approximated by a Gaussian function and attenuation linearly depends on frequency. The resulting downshift mean not only but also bandwidth propagation distance. kind approach is valid for narrowband pulses shallow penetration depth. However, short deep penetration, rather large received spectra are modified limited receiving system. In this...

Ultrasound (US) imaging is widely used for the tissue characterization. However, US images commonly suffer from speckle noise, which degrades perceived image quality. Various deep learning approaches have been proposed denoising, but most of them lack interpretability how network processing (black box problem). In this work, we utilize a reinforcement (RL) approach, pixelRL, to denoising. The technique utilizes set easily interpretable and filtering operations applied in pixel-wise manner....

Neoadjuvant chemotherapy (NAC) is widely used in the treatment of breast cancer. However, to date, there are no fully reliable, non-invasive methods for monitoring NAC. In this article, we propose a new method classifying NAC-responsive and unresponsive tumors using quantitative ultrasound.

Echocardiographic assessment of systolic and diastolic function the heart is often limited by image quality. However, aortic root well visualized in most patients. We hypothesize that motion may correlate with left ventricle heart. Data obtained from 101 healthy volunteers (mean age 46.6 ± 12.4) was used study. The data contained sequences standard two-dimensional (2D) echocardiographic B-mode (brightness mode, classical ultrasound grayscale presentation) images corresponding to single...

Objective. Prediction of the response to neoadjuvant chemotherapy (NAC) in breast cancer is important for patient outcomes. In this work, we propose a deep learning based approach NAC prediction ultrasound (US) imaging.Approach.We develop recurrent neural networks that can process serial US imaging data predict We present models either raw radio-frequency (RF) or regular images. The proposed evaluated on 204 sequences from 51 cancers. Each sequence included collected before and after each...

Neo-adjuvant chemotherapy (NAC) is used in breast cancer before tumor surgery to reduce the size of and risk spreading. Monitoring effects NAC important because a number cases response therapy poor requires change treatment. A new method that uses quantitative ultrasound assess has been presented. The aim was detect unresponsive tumors at an early stage treatment.The assumes scattering different for responsive nonresponsive tumors. assessment based on differences between histograms echo...

The pathological states of biological tissue are often resulted in attenuation changes. Thus, information about attenuating properties is valuable for the physician and could be useful ultrasonic diagnosis. We currently developing a technique parametric imaging we intend to apply it vivo characterization tissue. estimation method based on echoes mean frequency changes due dispersion, presented. Doppler IQ was adopted estimate directly from raw RF data. Singular Spectrum Analysis used...

Standard classification methods based on hand-crafted morphological and texture features have achieved good performance in breast mass differentiation ultrasound (US). In comparison to deep neural networks, commonly perceived as 'black-box' models, classical techniques are that well-understood medical physical interpretation. However, classifiers underperform the presence of shadowing artifact ill-defined borders, while may fail when US image is too noisy. Therefore, practice it would be...