Recent technological advancements in breast ultrasound

Evaluating diagnostic significance: The utilization of elastography and contrast-enhanced ultrasound for differential diagnosis in breast lesions

OBJECTIVE

The primary aim of this study is to assess the diagnostic efficacy of elastography and contrast-enhanced ultrasound (CEUS) in the identification of breast lesions subsequent to the optimization and correction of the BI-RADS category 4 classification obtained through conventional ultrasound. The objective is to augment both the specificity and accuracy of breast lesion diagnosis, thereby establishing a reliable framework for reducing unnecessary biopsies in clinical settings.

METHODS

A cohort comprising 50 cases of breast lesions classified under BI-RADS category 4 was collected during the period from November 2022 and November 2023. These cases were examined utilizing strain elastography (SE), shear wave elastography (SWE), and CEUS. Novel scoring methodologies for ultrasonic elastography (UE) and CEUS were formulated for this investigation. Subsequently, the developed UE and CEUS scoring systems were used to refine and optimize the conventional BI-RADS classification, either in isolation or in conjunction. Based on the revised classification, the benign group was classified as category 3 and the suspected malignant group was classified as category 4a and above, with pathological results serving as the definitive reference standard. The diagnostic efficacy of the optimized UE and CEUS, both independently and in combination, was meticulously scrutinized and compared using receiver operating characteristic (ROC) curve analysis, with pathological findings as the reference standard.

RESULTS

Within the study group, malignancy manifested in 11 cases. Prior to the implementation of the optimization criteria, 78% (39 out of 50) of patients underwent biopsies deemed unnecessary. Following the application of optimization criteria, specifically a threshold of≥8.5 points for the UE scoring method and≥6.5 points for the CEUS scoring method, the incidence of unnecessary biopsies diminished significantly. Reduction rates were observed at 53.8% (21 out of 39) with the UE protocol, 56.4% (22 out of 39) with the CEUS protocol, and 89.7% (35 out of 39) with the combined UE and CEUS optimization protocols.

CONCLUSION

The diagnostic efficacy of conventional ultrasound BI-RADS category 4 classification for breast lesions is enhanced following optimized correction using UE and CEUS, either independently or in conjunction. The application of the combined protocol demonstrates a notable reduction in the incidence of unnecessary biopsies.

ABUS tumor segmentation via decouple contrastive knowledge distillation

Objective.In recent years, deep learning-based methods have become the mainstream for medical image segmentation. Accurate segmentation of automated breast ultrasound (ABUS) tumor plays an essential role in computer-aided diagnosis. Existing deep learning models typically require a large number of computations and parameters.Approach. Aiming at this problem, we propose a novel knowledge distillation method for ABUS tumor segmentation. The tumor or non-tumor regions from different cases tend to have similar representations in the feature space. Based on this, we propose to decouple features into positive (tumor) and negative (non-tumor) pairs and design a decoupled contrastive learning method. The contrastive loss is utilized to force the student network to mimic the tumor or non-tumor features of the teacher network. In addition, we designed a ranking loss function based on ranking the distance metric in the feature space to address the problem of hard-negative mining in medical image segmentation.Main results. The effectiveness of our knowledge distillation method is evaluated on the private ABUS dataset and a public hippocampus dataset. The experimental results demonstrate that our proposed method achieves state-of-the-art performance in ABUS tumor segmentation. Notably, after distilling knowledge from the teacher network (3D U-Net), the Dice similarity coefficient (DSC) of the student network (small 3D U-Net) is improved by 7%. Moreover, the DSC of the student network (3D HR-Net) reaches 0.780, which is very close to that of the teacher network, while their parameters are only 6.8% and 12.1% of 3D U-Net, respectively.Significance. This research introduces a novel knowledge distillation method for ABUS tumor segmentation, significantly reducing computational demands while achieving state-of-the-art performance. The method promises enhanced accuracy and feasibility for computer-aided diagnosis in diverse imaging scenarios.

Bubble pulsation characteristics in multi-bubble systems affected by bubble size polydispersity and spatial structure

This study seeks to explore the bubble pulsation characteristics in multi-bubble environment with a special focus on the influences of the size polydispersity and the two-dimensional structure of bubbles. Three representative configurations of three interacting bubbles are formed by setting the initial radii of cavitation bubbles and inter-bubble distances appropriately, then the pulsation characteristics of a small bubble are investigated and compared by the bifurcation analysis. The results illustrate that the bubble size polydispersity and two-dimensional structure would greatly affect the bubble pulsations (i.e., the amplitude and nonlinearity of pulsations). Furthermore, the effects of two-dimensional structure are strong at a small inter-bubble distance of the large and small bubbles while the bubble size polydispersity always significantly affects the bubble pulsations for all cases. Moreover, the influences of both bubble size polydispersity and two-dimensional structure can be enhanced as the acoustic pressure increases, which can also become stronger when the large bubble is located at the same side as the small bubble and the initial radius of large bubble increases. Additionally, the effects would also be increased when the tissue viscoelasticity varies within a certain range. The present findings shed new light on the dynamics of multiple polydisperse microbubbles in viscoelastic tissues, potentially contributing to an optimization of their applications with ultrasound excitation.

Best Practice Guideline - DEGUM Recommendations on Breast Ultrasound

For many years, breast ultrasound has been used in addition to mammography as an important method for clarifying breast findings. However, differences in the interpretation of findings continue to be problematic 1 2. These differences decrease the diagnostic accuracy of ultrasound after detection of a finding and complicate interdisciplinary communication and the comparison of scientific studies 3. In 1999, the American College of Radiology (ACR) created a working group (International Expert Working Group) that developed a classification system for ultrasound examinations based on the established BI-RADS classification of mammographic findings under consideration of literature data 4. Due to differences in content, the German Society for Ultrasound in Medicine (DEGUM) published its own BI-RADS-analogue criteria catalog in 2006 3. In addition to the persistence of differences in content, there is also an issue with formal licensing with the current 5th edition of the ACR BI-RADS catalog, even though the content is recognized by the DEGUM as another system for describing and documenting findings. The goal of the Best Practice Guideline of the Breast Ultrasound Working Group of the DEGUM is to provide colleagues specialized in senology with a current catalog of ultrasound criteria and assessment categories as well as best practice recommendations for the various ultrasound modalities.

Deep weakly-supervised breast tumor segmentation in ultrasound images with explicit anatomical constraints

Breast tumor segmentation is an important step in the diagnostic procedure of physicians and computer-aided diagnosis systems. We propose a two-step deep learning framework for breast tumor segmentation in breast ultrasound (BUS) images which requires only a few manual labels. The first step is breast anatomy decomposition handled by a semi-supervised semantic segmentation technique. The input BUS image is decomposed into four breast anatomical structures, namely fat, mammary gland, muscle and thorax layers. Fat and mammary gland layers are used as constrained region to reduce the search space for breast tumor segmentation. The second step is breast tumor segmentation performed in a weakly-supervised learning scenario where only image-level labels are available. Breast tumors are first recognized by a classification network and then segmented by the proposed class activation mapping and deep level set (CAM-DLS) method. For breast anatomy decomposition, the proposed framework achieves Dice similarity coefficient (DSC) of 83.0 ± 11.8%, 84.3 ± 10.0%, 80.7 ± 15.4% and 91.0 ± 11.4% for fat, mammary gland, muscle and thorax layers, respectively. For breast tumor recognition, the proposed framework achieves sensitivity of 95.8%, precision of 92.4%, specificity of 93.9%, accuracy of 94.8% and F1-score of 0.941. For breast tumor segmentation, the proposed framework achieves DSC of 77.3% and intersection-over-union (IoU) of 66.0%. In conclusion, the proposed framework could efficiently perform breast tumor recognition and segmentation simultaneously in a weakly-supervised setting with anatomical constraints.

A multicenter, hospital-based and non-inferiority study for diagnostic efficacy of automated whole breast ultrasound for breast cancer in China

This study is the first multi-center non-inferiority study that aims to critically evaluate the effectiveness of HHUS/ABUS in China breast cancer detection. This was a multicenter hospital-based study. Five hospitals participated in this study. Women (30–69 years old) with defined criteria were invited for breast examination by HHUS, ABUS or/and mammography. For BI-RADS category 3, an additional magnetic resonance imaging (MRI) test was provided to distinguish the true negative results from false negative results. For women classified as BI-RADS category 4 or 5, either core aspiration biopsy or surgical biopsy was done to confirm the diagnosis. Between February 2016 and March 2017, 2844 women signed the informed consent form, and 1947 of them involved in final analysis (680 were 30 to 39 years old, 1267 were 40 to 69 years old).For all participants, ABUS sensitivity (91.81%) compared with HHUS sensitivity (94.70%) with non-inferior Z tests, P = 0.015. In the 40–69 age group, non-inferior Z tests showed that ABUS sensitivity (93.01%) was non-inferior to MG sensitivity (86.02%) with P < 0.001 and HHUS sensitivity (95.44%) was non-inferior to MG sensitivity (86.02%) with P < 0.001. Sensitivity of ABUS and HHUS are all superior to that of MG with P < 0.001 by superior test.For all participants, ABUS specificity (92.89%) was non-inferior to HHUS specificity (89.36%) with P < 0.001. Superiority test show that specificity of ABUS was superior to that of HHUS with P < 0.001. In the 40–69 age group, ABUS specificity (92.86%) was non-inferior to MG specificity (91.68%) with P < 0.001 and HHUS specificity (89.55%) was non-inferior to MG specificity (91.68%) with P < 0.001. ABUS is not superior to MG with P = 0.114 by superior test. The sensitivity of ABUS/HHUS is superior to that of MG. The specificity of ABUS/HHUS is non-inferior to that of MG. In China, for an experienced US radiologist, both HHUS and ABUS have better diagnostic efficacy than MG in symptomatic individuals.

Tumor segmentation in automated whole breast ultrasound using bidirectional LSTM neural network and attention mechanism

Accurate breast mass segmentation of automated breast ultrasound (ABUS) is a great help to breast cancer diagnosis and treatment. However, the lack of clear boundary and significant variation in mass shapes make the automatic segmentation very challenging. In this paper, a novel automatic tumor segmentation method SC-FCN-BLSTM is proposed by incorporating bi-directional long short-term memory (BLSTM) and spatial-channel attention (SC-attention) module into fully convolutional network (FCN). In order to decrease performance degradation caused by ambiguous boundaries and varying tumor sizes, an SC-attention module is designed to integrate both finer-grained spatial information and rich semantic information. Since ABUS is three-dimensional data, utilizing inter-slice context can improve segmentation performance. A BLSTM module with SC-attention is constructed to model the correlation between slices, which employs inter-slice context to assist segmentation for false positive elimination. The proposed method is verified on our private ABUS dataset of 124 patients with 170 volumes, including 3636 2D labeled slices. The Dice similarity coefficient (DSC), Recall, Precision and Hausdorff distance (HD) of the proposed method are 0.8178, 0.8067, 0.8292 and 11.1367. Experimental results demonstrate that the proposed method offered improved segmentation results compared with existing deep learning-based methods.

A 2-Phase Merge Filter Approach to Computer-Aided Detection of Breast Tumors on 3-Dimensional Ultrasound Imaging

OBJECTIVES

The role of image analysis in 3-dimensional (3D) automated breast ultrasound (ABUS) images is increasingly important because of its widespread use as a screening tool in whole-breast examinations. However, reviewing a large number of images acquired from ABUS is time-consuming and sometimes error prone. The aim of this study, therefore, was to develop an efficient computer-aided detection (CADe) algorithm to assist the review process.

METHODS

The proposed CADe algorithm consisted of 4 major steps. First, initial tumor candidates were formed by extracting and merging hypoechoic square cells on 2-dimensional (2D) transverse images. Second, a feature-based classifier was then constructed using 2D features to filter out nontumor candidates. Third, the remaining 2D candidates were merged longitudinally into 3D masses. Finally, a 3D feature-based classifier was used to further filter out nontumor masses to obtain the final detected masses. The proposed method was validated with 176 passes of breast images acquired by an Acuson S2000 automated breast volume scanner (Siemens Medical Solutions USA, Inc., Malvern, PA), including 44 normal passes and 132 abnormal passes containing 162 proven lesions (79 benign and 83 malignant).

RESULTS

The proposed CADe system could achieve overall sensitivity of 100% and 90% with 6.71 and 5.14 false-positives (FPs) per pass, respectively. Our results also showed that the average number of FPs per normal pass (7.16) was more than the number of FPs per abnormal pass (6.56) at 100% sensitivity.

CONCLUSIONS

The proposed CADe system has a great potential for becoming a good companion tool with ABUS imaging by ensuring high sensitivity with a relatively small number of FPs.

Production and clinical evaluation of breast lesion skin markers for automated three-dimensional ultrasonography of the breast: a pilot study

Objectives

Automated ultrasound of the breast has the advantage to have the whole breast scanned by technicians. Consequently, feedback to the radiologist about concurrent focal abnormalities (e.g., palpable lesions) is lost. To enable marking of patient- or physician-reported focal abnormalities, we aimed to develop skin markers that can be used without disturbing the interpretability of the image.

Methods

Disk-shaped markers were casted out of silicone. In this IRB-approved prospective study, 16 patients were included with a mean age of 57 (39–85). In all patients, the same volume was imaged twice using an automated breast ultrasound system, once with and once without a marker in place. Nine radiologists from two medical centers filled scoring forms regarding image quality, image interpretation, and confidence in providing a diagnosis based on the images.

Results

Marker adhesion was sufficient for automated scanning. Observer scores showed a significant shift in scores from excellent to good regarding diagnostic yield/image quality (χ², 15.99, p < 0.01), and image noise (χ², 21.20, p < 0.01) due to marker presence. In 93% of cases, the median score of observers “agree” with the statement that marker-induced noise did not influence image interpretability. Marker presence did not interfere with confidence in diagnosis (χ², 6.00, p = 0.20).

Conclusion

Inexpensive, easy producible skin markers can be used for accurate lesion marking in automated ultrasound examinations of the breast while image interpretability is preserved. Any marker-induced noise and decreased image quality did not affect confidence in providing a diagnosis.

Key Points

• The use of a skin marker enables the reporting radiologist to identify a location which a patient is concerned about.

• The developed skin marker can be used for accurate breast lesion marking in ultrasound examinations.

Assessment of 2D ultrasound fluid volume estimation accuracy in different shaped objects: an in vitro study

Background

Volume estimation of organs using two-dimensional (2D) ultrasonography is frequently warranted. Considering the influence of estimated volume on patient management, maintenance of its high accuracy is empirical. However, data are scarce regarding the accuracy of estimated volume of non-globular shaped objects of different volumes.

Purpose

To evaluate the volume estimation accuracy of different shaped and sized objects using high-end 2D ultrasound scanners.

Material and Methods

Globular (n=5); non-globular elongated (n=5), and non-globular near-spherical shaped (n=4) hollow plastic objects were scanned to estimate the volumes; actual volumes were compared with estimated volumes. T-test and one-way ANOVA were used to compare means; P<0.05 was considered significant.

Results

The actual volumes of the objects were in the range of 10–445 mL; estimated volumes ranged from 6.4–425 mL ( P=0.067). The estimated volume was lower than the actual volume; such volume underestimation was marked for non-globular elongated objects. Regardless of the scanner, the highest volume estimation error was for non-globular elongated objects (<40%) followed by non-globular near-spherical shaped objects (<23.88%); the lowest was for globular objects (<3.6%). Irrespective of the shape or the volume of the object, volume estimation difference among the scanners was not significant: globular (F=0.430, P=0.66); non-globular elongated (F=3.69, P=0.064); and non-globular near-spherical (F=4.00, P=0.06). A good inter-rater agreement (R=0.99, P<0.001) and a good correlation between actual versus estimated volumes (R=0.98, P<0.001) were noted.

Conclusion

The 2D ultrasonography can be recommended for volume estimation purposes of different shaped and different sized objects, regardless the type of the high-end scanner used.

Role of ultrasound, Color duplex Doppler and sono-elastography in the evaluation of renal allograft complications

Background

Renal transplantation could be considered for patients with end-stage renal disease. Ultrasonography is the imaging method chosen for renal allograft evaluation early in the postoperative period. Sono-elastography is used to estimate tissue stiffness. This study aimed to assess the correlation between sono-elastography and renal allograft histopathology in patients who had transplanted kidney and determine the efficacy of grayscale ultrasound, sono-elastography, and color duplex in evaluation renal allograft complications and correlation with renal function.

Results

Forty patients (26 males and 14 females) who underwent renal transplantation were included. Their ages ranging from 10:52 years; they all subjected to ultrasound, color Doppler, sono-elastography, and histopathology. The studded patients were divided into 3 groups according to histopathology: patient with normal results, patients with ATI, and patients with CAI. The difference between the mean elasticity values between the three groups was statistically highly significant (p value < 0.001) with high specificity, sensitivity, and accuracy in differentiating ATI and normal groups and also CAI and normal groups, while the lowest sensitivity noticed in differentiating between ATI and CAI groups that is mean elasticity was good to differentiating between ATI and normal groups and also between CAI and normal groups while it was less in differentiating between ATI and CAI groups. As regards the RI, the highest sensitivity of the RI was in differentiating ATI and normal with high sensitivity, specificity, and accuracy, and the lowest sensitivity, specificity, and accuracy of RI were in differentiating CAI and normal groups.

Conclusion

Transplanted renal allograft could be assessed by combined US, color duplex Doppler, and sono-elastography examination; also we can detect posttransplant complications early. Sono-elastography could be an efficient noninvasive method to diagnose and monitor kidney allograft rejection and follow-up of the renal allograft, which may give a further and possibly earlier prognostic index for chronic dysfunction in addition to serum creatinine.

Classification of Benign and Malignant Breast Tumors Using H-Scan Ultrasound Imaging

Breast cancer is one of the most common cancers among women worldwide. Ultrasound imaging has been widely used in the detection and diagnosis of breast tumors. However, due to factors such as limited spatial resolution and speckle noise, classification of benign and malignant breast tumors using conventional B-mode ultrasound still remains a challenging task. H-scan is a new ultrasound technique that images the relative size of acoustic scatterers. However, the feasibility of H-scan ultrasound imaging in the classification of benign and malignant breast tumors has not been investigated. In this paper, we proposed a new method based on H-scan ultrasound imaging to classify benign and malignant breast tumors. Backscattered ultrasound radiofrequency signals of 100 breast tumors were used (48 benign and 52 malignant cases). H-scan ultrasound images were constructed with the radiofrequency signals by matched filtering using Gaussian-weighted Hermite polynomials. Experimental results showed that benign breast tumors had more red components, while malignant breast tumors had more blue components in H-scan ultrasound images. There were significant differences between the RGB channels of H-scan ultrasound images of benign and malignant breast tumors. We conclude H-scan ultrasound imaging can be used as a new method for classifying benign and malignant breast tumors.

Initialization of active contours for segmentation of breast cancer via fusion of ultrasound, Doppler, and elasticity images

Active contours (snakes) are an efficient method for segmentation of ultrasound (US) images of breast cancer. However, the method produces inaccurate results if the seeds are initialized improperly (far from the true boundaries and close to the false boundaries). Therefore, we propose a novel initialization method based on the fusion of a conventional US image with elasticity and Doppler images. The proposed fusion method (FM) has been tested against four state-of-the-art initialization methods on 90 ultrasound images from a database collected by the Thammasat University Hospital of Thailand. The ground truth was hand-drawn by three leading radiologists of the hospital. The reference methods are: center of divergence (CoD), force field segmentation (FFS), Poisson Inverse Gradient Vector Flow (PIG), and quasi-automated initialization (QAI). A variety of numerical tests proves the advantages of the FM. For the raw US images, the percentage of correctly initialized contours is: FM-94.2%, CoD-0%, FFS-0%, PIG-26.7%, QAI-42.2%. The percentage of correctly segmented tumors is FM-84.4%, CoD-0%, FFS-0%, PIG-16.67%, QAI-22.44%. For reduced field of view US images, the percentage of correctly initialized contours is: FM-94.2%, CoD-0%, FFS-0%, PIG-65.6%, QAI-67.8%. The correctly segmented tumors are FM-88.9%, CoD-0%, FFS-0%, PIG-48.9%, QAI-44.5%. The accuracy, in terms of the average Hausdorff distance, is respectively 2.29 pixels, 33.81, 34.71, 7.7, and 8.4, whereas in terms of the Jaccard index, it is 0.9, 0.18, 0.19, 0.63, and 0.48.

Exploiting heat shock protein expression to develop a non-invasive diagnostic tool for breast cancer

Leveraging the unique surface expression of heat shock protein 90 (Hsp90) in breast cancer provides an exciting opportunity to develop rapid diagnostic tests at the point-of-care setting. Hsp90 has previously been shown to have elevated expression levels across all breast cancer receptor subtypes. We have developed a non-destructive strategy using HS-27, a fluorescently-tethered Hsp90 inhibitor, to assay surface Hsp90 expression on intact tissue specimens and validated our approach in clinical samples from breast cancer patients across estrogen receptor positive, Her2-overexpressing, and triple negative receptor subtypes. Utilizing a pre-clinical biopsy model, we optimized three imaging parameters that may affect the specificity of HS-27 based diagnostics – time between tissue excision and staining, agent incubation time, and agent dose, and translated our strategy to clinical breast cancer samples. Findings indicated that HS-27 florescence was highest in tumor tissue, followed by benign tissue, and finally followed by mammoplasty negative control samples. Interestingly, fluorescence in tumor samples was highest in Her2+ and triple negative subtypes, and inversely correlated with the presence of tumor infiltrating lymphocytes indicating that HS-27 fluorescence increases in aggressive breast cancer phenotypes. Development of a Gaussian support vector machine classifier based on HS-27 fluorescence features resulted in a sensitivity and specificity of 82% and 100% respectively when classifying tumor and benign conditions, setting the stage for rapid and automated tissue diagnosis at the point-of-care.

Review of quantitative multiscale imaging of breast cancer

Breast cancer is the most common cancer among women worldwide and ranks second in terms of overall cancer deaths. One of the difficulties associated with treating breast cancer is that it is a heterogeneous disease with variations in benign and pathologic tissue composition, which contributes to disease development, progression, and treatment response. Many of these phenotypes are uncharacterized and their presence is difficult to detect, in part due to the sparsity of methods to correlate information between the cellular microscale and the whole-breast macroscale. Quantitative multiscale imaging of the breast is an emerging field concerned with the development of imaging technology that can characterize anatomic, functional, and molecular information across different resolutions and fields of view. It involves a diverse collection of imaging modalities, which touch large sections of the breast imaging research community. Prospective studies have shown promising results, but there are several challenges, ranging from basic physics and engineering to data processing and quantification, that must be met to bring the field to maturity. This paper presents some of the challenges that investigators face, reviews currently used multiscale imaging methods for preclinical imaging, and discusses the potential of these methods for clinical breast imaging.

Elastography techniques in the evaluation of deep vein thrombosis

Deep venous thrombosis (DVT) is a significant medical problem with an incidence of 1 in 1,000 adults and greatly reduces quality of life through post-thrombotic syndrome. Treatment choice for DVT can be influenced by the age of the clot. While new endovascular catheter techniques treat venous clots to potentially prevent post-thrombotic syndrome, they require improved imaging techniques to accurately determine clot age. This review investigates experimental and clinical evidence of elastography techniques for aging DVT. Strain elastography and shear wave elastography are the most common techniques to age thrombus. These elastography techniques can distinguish between acute and chronic clots by characterizing tissue stiffness. When clot age cannot be determined with ultrasound duplex analysis, elastography may offer a helpful adjunct. However, further investigation is required to validate accuracy and reproducibility for clinical implementation of this novel technique.

Application of Ultrasound Elastography for Chronic Allograft Dysfunction in Kidney Transplantation

Interstitial fibrosis is the main characteristic of chronic allograft dysfunction, which remains the key factor affecting long-term allograft survival after kidney transplantation. Ultrasound elastography (UE), including real-time elastography, transient elastography, and acoustic radiation force impulse, has been applied widely in breast, thyroid, and liver diseases, especially in the assessment of liver fibrosis. Recently, numerous studies have reported the efficacy of UE methods in evaluating renal allograft fibrosis. This review aims to investigate the clinical applications, limitations, and future roles of UE in current clinical practice in light of changing management paradigms. In current clinical practice, UE methods, especially transient elastographic measurement, appear to be useful for ruling out fibrosis but do not have sufficient accuracy to distinguish between various stages of allograft fibrosis. Moreover, there remain considerable issues to be solved for the application of UE in kidney transplantation. Thus, UE methods cannot replace the crucial role of renal allograft biopsy in the diagnosis and evaluation of allograft fibrosis in kidney transplantation. Perhaps UE methods could be of more importance in the long-term observation and evaluation of allograft fibrosis during follow-up.