Breast-lesions characterization using Quantitative Ultrasound features of peritumoral tissue

Evaluation of fragility fracture risk using deep learning based on ultrasound radio frequency signal

BACKGROUND

It was essential to identify individuals at high risk of fragility fracture and prevented them due to the significant morbidity, mortality, and economic burden associated with fragility fracture. The quantitative ultrasound (QUS) showed promise in assessing bone structure characteristics and determining the risk of fragility fracture.

AIMS

To evaluate the performance of a multi-channel residual network (MResNet) based on ultrasonic radiofrequency (RF) signal to discriminate fragility fractures retrospectively in postmenopausal women, and compared it with the traditional parameter of QUS, speed of sound (SOS), and bone mineral density (BMD) acquired with dual X-ray absorptiometry (DXA).

METHODS

Using QUS, RF signal and SOS were acquired for 246 postmenopausal women. An MResNet was utilized, based on the RF signal, to categorize individuals with an elevated risk of fragility fracture. DXA was employed to obtain BMD at the lumbar, hip, and femoral neck. The fracture history of all adult subjects was gathered. Analyzing the odds ratios (OR) and the area under the receiver operator characteristic curves (AUC) was done to evaluate the effectiveness of various methods in discriminating fragility fracture.

RESULTS

Among the 246 postmenopausal women, 170 belonged to the non-fracture group, 50 to the vertebral group, and 26 to the non-vertebral fracture group. MResNet was competent to discriminate any fragility fracture (OR = 2.64; AUC = 0.74), Vertebral fracture (OR = 3.02; AUC = 0.77), and non-vertebral fracture (OR = 2.01; AUC = 0.69). After being modified by clinical covariates, the efficiency of MResNet was further improved to OR = 3.31-4.08, AUC = 0.81-0.83 among all fracture groups, which significantly surpassed QUS-SOS (OR = 1.32-1.36; AUC = 0.60) and DXA-BMD (OR = 1.23-2.94; AUC = 0.63-0.76).

CONCLUSIONS

This pilot cross-sectional study demonstrates that the MResNet model based on the ultrasonic RF signal shows promising performance in discriminating fragility fractures in postmenopausal women. When incorporating clinical covariates, the efficiency of the modified MResNet is further enhanced, surpassing the performance of QUS-SOS and DXA-BMD in terms of OR and AUC. These findings highlight the potential of the MResNet as a promising approach for fracture risk assessment. Future research should focus on larger and more diverse populations to validate these results and explore its clinical applications.

Influences of Variability in Attenuation Compensation on the Estimation of Backscatter Coefficient of Median Nerves in Vivo

OBJECTIVE

Peripheral nerves remain a challenging target for medical imaging, given their size, anatomical complexity, and structural heterogeneity. Quantitative ultrasound (QUS) applies a set of techniques to estimate tissue acoustic parameters independent of the imaging platform. Many useful medical and laboratory applications for QUS have been reported, but challenges remain for deployment in vivo, especially for heterogeneous tissues. Several phenomena introduce variability in attenuation estimates, which may influence the estimation of other QUS parameters. For example, estimating the backscatter coefficient (BSC) requires compensation for the attenuation of overlying tissues between the transducer and the underlying tissue of interest. The purpose of this study is to extend prior studies by investigating the efficacy of several analytical methods of estimating attenuation compensation on QUS outcomes in the human median nerve.

METHODS

Median nerves were imaged at the volar wrist in vivo and beam-formed radiofrequency (RF) data were acquired. Six analytical approaches for attenuation compensation were compared: 1-2) attenuation estimated by applying spectral difference method (SDM) and spectral log difference method (SLDM) independently to regions of interest (ROIs) overlying the nerve and to the nerve ROI itself; 3-4) attenuation estimation by applying SDM and SLDM to ROIs overlying the nerve, and transferring these properties to the nerve ROI; and 5-6) methods that apply previously published values of tissue attenuation to the measured thickness of each overlying tissue. Mean between-subject estimates of BSC-related outcomes as well as within-subject variability of these outcomes were compared among the 6 methods.

RESULTS

Compensating for attenuation using SLDM and values from the literature reduced variability in BSC-based outcomes, compared to SDM. Variability in attenuation coefficients contributes substantially to variability in backscatter measurements.

CONCLUSION

This work has implications for the application of QUS to in vivo diagnostic assessments in peripheral nerves and possibly other heterogeneous tissues.

Predicting axillary lymph node metastasis in breast cancer based on ultrasound radiofrequency time-series analysis

BACKGROUND

The status of axillary lymph nodes (ALN) plays a critical role in the management of patients with breast cancer. It is an urgent demand to develop highly accurate, non-invasive methods for predicting ALN status.

PURPOSE

To evaluate the efficacy of ultrasound radiofrequency (URF) time-series parameters, in combination with clinical data, in predicting ALN metastasis in patients with breast cancer.

MATERIAL AND METHODS

We prospectively gathered clinicopathologic and ultrasonic data from patients diagnosed with breast cancer. Various machine-learning (ML) models were developed using all available features to determine the most efficient diagnostic model. Subsequently, distinct prediction models were created using the optimal ML model, and their diagnostic performances were evaluated and compared.

RESULTS

The study encompassed 240 patients, of whom 88 had lymph node metastases. A leave-one-out cross-validation (LOOCV) method was used to split the entire dataset into training and testing subsets. The random forest ML model outperformed the other algorithms, with an area under the curve (AUC) of 0.92. Prediction models based on clinical, ultrasonic, URF parameters, clinical + ultrasonic, clinical + URF, and ultrasonic + URF parameters had AUCs of 0.56, 0.79, 0.78, 0.90, 0.80, and 0.84, respectively, in the testing set. The comprehensive diagnostic model (clinical + ultrasonic + URF parameters) demonstrated strong diagnostic capability, with an AUC of 0.94 in the testing set, exceeding any single prediction model.

CONCLUSION

The combined model (clinical + ultrasonic + URF parameters) could be used preoperatively to predict lymph node status, offering valuable input for the design of individualized surgical approaches.

Machine Learning Model for Predicting Axillary Lymph Node Metastasis in Clinically Node Positive Breast Cancer Based on Peritumoral Ultrasound Radiomics and SHAP Feature Analysis

OBJECTIVE

This study seeks to construct a machine learning model that merges clinical characteristics with ultrasound radiomic analysis-encompassing both the intratumoral and peritumoral-to predict the status of axillary lymph nodes in patients with early-stage breast cancer.

METHODS

The study employed retrospective methods, collecting clinical information, ultrasound data, and postoperative pathological results from 321 breast cancer patients (including 224 in the training group and 97 in the validation group). Through correlation analysis, univariate analysis, and Lasso regression analysis, independent risk factors related to axillary lymph node metastasis in breast cancer were identified from conventional ultrasound and immunohistochemical indicators, and a clinical feature model was constructed. Additionally, features were extracted from ultrasound images of the intratumoral and its 1-5 mm peritumoral to establish a radiomics feature formula. Furthermore, by combining clinical features and ultrasound radiomics features, six machine learning models (Logistic Regression, Decision Tree, Support Vector Machine, Extreme Gradient Boosting, Random Forest, and K-Nearest Neighbors) were compared for diagnostic efficacy, and constructing a joint prediction model based on the optimal ML algorithm. The use of Shapley Additive Explanations (SHAP) enhanced the visualization and interpretability of the model during the diagnostic process.

RESULTS

Among the 321 breast cancer patients, 121 had axillary lymph node metastasis, and 200 did not. The clinical feature model had an AUC of 0.779 and 0.777 in the training and validation groups, respectively. Radiomics model analysis showed that the model including the Intratumor +3 mm peritumor area had the best diagnostic performance, with AUCs of 0.847 and 0.844 in the training and validation groups, respectively. The joint prediction model based on the XGBoost algorithm reached AUCs of 0.917 and 0.905 in the training and validation groups, respectively. SHAP analysis indicated that the Rad Score had the highest weight in the prediction model, playing a significant role in predicting axillary lymph node metastasis in breast cancer.

CONCLUSION

The predictive model, which integrates clinical features and radiomic characteristics using the XGBoost algorithm, demonstrates significant diagnostic value for axillary lymph node metastasis in breast cancer. This model can provide significant references for preoperative surgical strategy selection and prognosis evaluation for breast cancer patients, helping to reduce postoperative complications and improve long-term survival rates. Additionally, the utilization of SHAP enhancing the global and local interpretability of the model.

Development of a nomogram-based model combining intra- and peritumoral ultrasound radiomics with clinical features for differentiating benign from malignant in Breast Imaging Reporting and Data System category 3-5 nodules

Background

The differences in benign and malignant breast tumors are not only within the nodules but also involve changes in the surrounding tissues. Radiomics can reveal many details that are not discernible to the naked eye. This study aimed to distinguish between benign and malignant breast nodules using an ultrasound-based intra- and peritumoral radiomics model.

Methods

This study retrospectively collected the information from 379 patients with Breast Imaging Reporting and Data System (BI-RADS) category 3-5 nodules and clear pathological diagnosis of breast nodules screened by routine ultrasound examination in the Sixth People's Hospital Affiliated to Medical College of Shanghai Jiao Tong University from January 2017 to December 2022. The largest dimension of the lesion on the 2D ultrasound image was selected to outline the area of interest which was conformally and outwardly expanded automatically by 5 mm to extract intra- and peritumor radiomics features. The included cases were randomly divided into training sets and test sets in a ratio of 7:3. The optimal features of the included models were retained by statistical and machine learning methods of dimensionality reduction, and logistic regression was used as the classifier to build an intratumoral model and a combined intratumoral-peritumoral radiomics model, respectively; through single-factor and multifactor logistic regression, the optimal features that could predict benign and malignant breast tumors were screened. The clinical and imaging models were established by selecting independent risk factors as clinical and imaging features through univariate and multifactorial logistic regression.

Results

Among 379 BI-RADS category 3-5 breast nodules, there were 124 malignant nodules and 255 benign nodules; patients were aged 14 to 88 (46.22±15.51) years, and the age differences, radiomics score, and mass diameter between the training and test sets were not statistically significant (P>0.05). The intra- and peritumor radiomics model had an area under the curve (AUC) of 0.840 [95% confidence interval (CI): 0.766-0.914] in the test set. The model with intra- and peritumoral ultrasound radiomics features combined with clinical features had an AUC value of 0.960 (95% CI: 0.920-0.999).

Conclusions

The nomogram, developed using intratumoral and peritumoral radiomics features combined with clinical risk features, demonstrated superior performance in distinguishing between benign and malignant BI-RADS 3-5 lesions.

Multimodal radiomics and nomogram-based prediction of axillary lymph node metastasis in breast cancer: An analysis considering optimal peritumoral region

PURPOSE

To explore the optimal peri-tumoral regions on ultrasound (US) images and investigate the performance of multimodal radiomics for predicting axillary lymph node metastasis (ALNM).

METHODS

This retrospective study included 326 patients (training cohort: n = 162, internal validation cohort: n = 74, external validation cohort: n = 90). Intra-tumoral region of interests (ROIs) were delineated on US and digital mammography (DM) images. Peri-tumoral ROI (PTR) on US images were gained by dilating actual 0.5, 1.0, 1.5, 2.0, 2.5, 3.0 and 3.5 mm radius surrounding the tumor. Support vector machine (SVM) method was used to calculate the importance of radiomics features and to pick the 10 most important. Recursive feature elimination-SVM was used to evaluate the efficacy of models with different feature numbers used.

RESULTS

The PTR0.5mm yielded a maximum AUC of 0.802 (95% confidence interval (CI): 0.676-0.901) within the validation cohort using SVM classifier. The multimodal radiomics (intra-tumoral US and DM and US-based PTR0.5mm radiomics model) achieved the highest predictive ability (AUC = 0.888/0.844/0.835 and 95% CI = 0.829-0.936/0.741-0.929/0.752-0.896 for training/internal validation/external validation cohort, respectively).

CONCLUSION

The PTR0.5mm could be the optimal area for predicting ALNM. A favorable predictive accuracy for predicting ALNM was achieved using multimodal radiomics and its based nomogram.

Breast cancer patient-derived microtumors resemble tumor heterogeneity and enable protein-based stratification and functional validation of individualized drug treatment

Despite tremendous progress in deciphering breast cancer at the genomic level, the pronounced intra- and intertumoral heterogeneity remains a major obstacle to the advancement of novel and more effective treatment approaches. Frequent treatment failure and the development of treatment resistance highlight the need for patient-derived tumor models that reflect the individual tumors of breast cancer patients and allow a comprehensive analyses and parallel functional validation of individualized and therapeutically targetable vulnerabilities in protein signal transduction pathways. Here, we introduce the generation and application of breast cancer patient-derived 3D microtumors (BC-PDMs). Residual fresh tumor tissue specimens were collected from n = 102 patients diagnosed with breast cancer and subjected to BC-PDM isolation. BC-PDMs retained histopathological characteristics, and extracellular matrix (ECM) components together with key protein signaling pathway signatures of the corresponding primary tumor tissue. Accordingly, BC-PDMs reflect the inter- and intratumoral heterogeneity of breast cancer and its key signal transduction properties. DigiWest®-based protein expression profiling of identified treatment responder and non-responder BC-PDMs enabled the identification of potential resistance and sensitivity markers of individual drug treatments, including markers previously associated with treatment response and yet undescribed proteins. The combination of individualized drug testing with comprehensive protein profiling analyses of BC-PDMs may provide a valuable complement for personalized treatment stratification and response prediction for breast cancer.

Ultrasound Characterization of Cortical Bone Using Shannon Entropy

OBJECTIVE

Ultrasound backscattered signals encompass information on the microstructure of heterogeneous media such as cortical bone, in which pores act as scatterers and result in the scattering and multiple scattering of ultrasound waves. The objective of this study was to investigate whether Shannon entropy can be exploited to characterize cortical porosity.

METHODS

In the study described here, to demonstrate proof of concept, Shannon entropy was used as a quantitative ultrasound parameter to experimentally evaluate microstructural changes in samples with controlled scatterer concentrations made of a highly absorbing polydimethylsiloxane matrix (PDMS). Similar assessment was then performed using numerical simulations on cortical bone structures with varying average pore diameter (Ct.Po.Dm.), density (Ct.Po.Dn.) and porosity (Ct.Po.).

RESULTS

The results suggest that an increase in pore diameter and porosity lead to an increase in entropy, indicating increased levels of randomness in the signals as a result of increased scattering. The entropy-versus-scatterer volume fraction in PDMS samples indicates an initial increasing trend that slows down as the scatterer concentration increases. High levels of attenuation cause the signal amplitudes and corresponding entropy values to decrease drastically. The same trend is observed when porosity of the bone samples is increased above 15%.

CONCLUSION

Sensitivity of entropy to microstructural changes in highly scattering and absorbing media can potentially be exploited to diagnose and monitor osteoporosis.

Enhanced full-inversion-based ultrasound elastography for evaluating tumor response to neoadjuvant chemotherapy in patients with locally advanced breast cancer

PURPOSE

An enhanced ultrasound elastography technique is proposed for early assessment of locally advanced breast cancer (LABC) response to neoadjuvant chemotherapy (NAC).

METHODS

The proposed elastography technique inputs ultrasound radiofrequency data obtained through tissue quasi-static stimulation and adapts a strain refinement algorithm formulated based on fundamental principles of continuum mechanics, coupled with an iterative inverse finite element method to reconstruct the breast Young's modulus (E) images. The technique was explored for therapy response assessment using data acquired from 25 LABC patients before and at weeks 1, 2, and 4 after the NAC initiation (100 scans). The E ratio of tumor to the surrounding tissue was calculated at different scans and compared to the baseline for each patient. Patients' response to NAC was determined many months later using standard clinical and histopathological criteria.

RESULTS

Reconstructed E ratio changes obtained as early as one week after the NAC onset demonstrate very good separation between the two cohorts of responders and non-responders to NAC. Statistically significant differences were observed in the E ratio changes between the two patient cohorts at weeks 1 to 4 after treatment (p-value < 0.001; statistical power greater than 97%). A significant difference in axial strain ratio changes was observed only at week 4 (p-value = 0.01; statistical power = 76%). No significant difference was observed in tumor size changes at weeks 1, 2 or 4.

CONCLUSION

The proposed elastography technique demonstrates a high potential for chemotherapy response monitoring in LABC patients and superior performance compared to strain imaging.

Artificial Intelligence for Breast US

US is a widely available, commonly used, and indispensable imaging modality for breast evaluation. It is often the primary imaging modality for the detection and diagnosis of breast cancer in low-resource settings. In addition, it is frequently employed as a supplemental screening tool via either whole breast handheld US or automated breast US among women with dense breasts. In recent years, a variety of artificial intelligence systems have been developed to assist radiologists with the detection and diagnosis of breast lesions on US. This article reviews the background and evidence supporting the use of artificial intelligence tools for breast US, describes implementation strategies and impact on clinical workflow, and discusses potential emerging roles and future directions.

Subclavian brachial plexus metastasis from breast cancer: A case report

BACKGROUND

The common area of breast cancer metastases are bone, lung and liver. Brachial plexus metastasis from breast cancer is extremely rare. We report a case of subclavian brachial plexus metastasis from breast cancer 6 years postoperative, which were detected by ultrasound, magnetic resonance imaging (MRI) and ¹⁸F-fluorodeoxyglucose positron emission tomography and computed tomography (FDG-PET/CT).

CASE SUMMARY

Our study reports a 64-year-old woman who had right breast cancer and underwent radical mastectomy 6 years before. Ultrasound first revealed a soft lesion measuring 38 mm × 37 mm which located on the right side of the clavicle to the armpit subcutaneously. The right subclavian brachial plexus (beam level) was significantly thickened, wrapped around by a hypoechoic lesion, the surrounded axillary artery and vein were pressed. MRI brachial plexus scan showed that the right side of brachial plexus was enlarged compared with the left side and brachial plexus bundle in the distance showed a flake shadow. FDG-PET/CT revealed that the right side of brachial plexus nodular appearance with increased FDG metabolism. These results supported brachial plexus metastasis from breast cancer. Ultrasound exam also found many lesions between pectoralis major, deltoid muscle and inner upper arm. The lesion puncture was performed under ultrasound guidance and the tissue was sent for pathology. Pathology showed large areas of tumor cells in fibroblast tissue. Immunohistochemistry showed the following results: A2-1: GATA3 (+), ER (+, strong, 90%), PR (+, moderate, 10%), HER-2 (3+), Ki67 (+15%), P120 (membrane+), P63 (-), E-cadherin (+), CK5/6 (-). These results were consistent with the primary right breast cancer characteristics, thus supporting lesion metastasis from breast cancer.

CONCLUSION

The brachial plexus metastasis from breast cancer is uncommon. Ultrasound has great value in detecting brachial plexus metastasis of breast cancer. It is an easy, non-invasive and affordable method. Close attention should be paid to new grown out lesions in those patients who had a history of breast cancer when doing ultrasound review.

Ultrasound classification of breast masses using a comprehensive Nakagami imaging and machine learning framework

In this study we investigate the potential of parametric images formed from ultrasound B-mode scans using the Nakagami distribution for non-invasive classification of breast lesions and characterization of breast tissue. Through a sliding window technique, we generated seven types of Nakagami images for each patient scan in our dataset using basic and as well as derived parameters of the Nakagami distribution. To determine the suitable window size for image generation, we conducted an empirical analysis using 4 windows, which includes 3 column windows of lengths 0.1875 mm, 0.45 mm and 0.75 mm and widths of 0.002 mm, along with the standard square window with sides equal to three times the pulse length of incident ultrasound. From the parametric image sets generated using each window, we extracted a total of 72 features that consisted of morphometric, elemental and hybrid features. To our knowledge no other literature has conducted such a comprehensive analysis of Nakagami parametric images for the classification of breast lesions. Feature selection was performed to find the most useful subset of features from each of the parametric image sets for the classification of breast cancer. Analyzing the classification accuracy and Area under the Receiver Operating Characteristic (ROC) Curve (AUC) of the selected feature subsets, we determined that the selected features acquired from Nakagami parametric images generated using a column window of length 0.75 mm provides the best results for characterization of breast lesions. This optimal feature set provided a classification accuracy of 93.08%, an AUC of 0.9712, a False Negative Rate (FNR) of 0%, and a very low False Positive Rate (FPR) of 8.65%. Our results indicate that the high accuracy of such a procedure may assist in the diagnosis of breast cancer by helping to reduce false positive diagnoses.

Artificial Intelligence in Quantitative Ultrasound Imaging: A Survey

Quantitative ultrasound (QUS) imaging is a safe, reliable, inexpensive, and real-time technique to extract physically descriptive parameters for assessing pathologies. Compared with other major imaging modalities such as computed tomography and magnetic resonance imaging, QUS suffers from several major drawbacks: poor image quality and inter- and intra-observer variability. Therefore, there is a great need to develop automated methods to improve the image quality of QUS. In recent years, there has been increasing interest in artificial intelligence (AI) applications in medical imaging, and a large number of research studies in AI in QUS have been conducted. The purpose of this review is to describe and categorize recent research into AI applications in QUS. We first introduce the AI workflow and then discuss the various AI applications in QUS. Finally, challenges and future potential AI applications in QUS are discussed.

Breast Tumor Classification Using Intratumoral Quantitative Ultrasound Descriptors

Breast cancer is a global epidemic, responsible for one of the highest mortality rates among women. Ultrasound imaging is becoming a popular tool for breast cancer screening, and quantitative ultrasound (QUS) techniques are being increasingly applied by researchers in an attempt to characterize breast tissue. Several different quantitative descriptors for breast cancer have been explored by researchers. This study proposes a breast tumor classification system using the three major types of intratumoral QUS descriptors which can be extracted from ultrasound radiofrequency (RF) data: spectral features, envelope statistics features, and texture features. A total of 16 features were extracted from ultrasound RF data across two different datasets, of which one is balanced and the other is severely imbalanced. The balanced dataset contains RF data of 100 patients with breast tumors, of which 48 are benign and 52 are malignant. The imbalanced dataset contains RF data of 130 patients with breast tumors, of which 104 are benign and 26 are malignant. Holdout validation was used to split the balanced dataset into 60% training and 40% testing sets. Feature selection was applied on the training set to identify the most relevant subset for the classification of benign and malignant breast tumors, and the performance of the features was evaluated on the test set. A maximum classification accuracy of 95% and an area under the receiver operating characteristic curve (AUC) of 0.968 was obtained on the test set. The performance of the identified relevant features was further validated on the imbalanced dataset, where a hybrid resampling strategy was firstly utilized to create an optimal balance between benign and malignant samples. A maximum classification accuracy of 93.01%, sensitivity of 94.62%, specificity of 91.4%, and AUC of 0.966 were obtained. The results indicate that the identified features are able to distinguish between benign and malignant breast lesions very effectively, and the combination of the features identified in this research has the potential to be a significant tool in the noninvasive rapid and accurate diagnosis of breast cancer.

Accuracy of ultrasonic artificial intelligence in diagnosing benign and malignant breast diseases: A protocol for systematic review and meta-analysis

BACKGROUND

Artificial intelligence system is a deep learning system based on computer-assisted ultrasonic image diagnosis, which can extract morphological features of breast mass and conduct objective and efficient image analysis, thus automatically intelligent classification of breast mass, avoiding subjective error and improving the accuracy of diagnosis.[1-2] A large number of studies have confirmed that artificial intelligence (AI) has high effectiveness and reliability in the differential diagnosis of benign and malignant breast diseases.[3-4] However, the results of these studies have been contradictory. Therefore, this meta-analysis tested the hypothesis that artificial intelligence system is accurate in distinguishing benign and malignant breast diseases.

METHODS

We will search PubMed, Web of Science, Cochrane Library, and Chinese biomedical databases from their inceptions to the November 20, 2021, without language restrictions. Two authors will independently carry out searching literature records, scanning titles and abstracts, full texts, collecting data, and assessing risk of bias. Review Manager 5.2 and Stata14.0 software will be used for data analysis.

RESULTS

This systematic review will determine the accuracy of AI in the differential diagnosis of benign and malignant breast diseases.

CONCLUSION

Its findings will provide helpful evidence for the accuracy of AI in the differential diagnosis of benign and malignant breast diseases.

SYSTEMATIC REVIEW REGISTRATION

INPLASY2021110087.

FMRNet: A fused network of multiple tumoral regions for breast tumor classification with ultrasound images

PURPOSE

Recent studies have illustrated that the peritumoral regions of medical images have value for clinical diagnosis. However, the existing approaches using peritumoral regions mainly focus on the diagnostic capability of the single region and ignore the advantages of effectively fusing the intratumoral and peritumoral regions. In addition, these methods need accurate segmentation masks in the testing stage, which are tedious and inconvenient in clinical applications. To address these issues, we construct a deep convolutional neural network that can adaptively fuse the information of multiple tumoral-regions (FMRNet) for breast tumor classification using ultrasound (US) images without segmentation masks in the testing stage.

METHODS

To sufficiently excavate the potential relationship, we design a fused network and two independent modules to extract and fuse features of multiple regions simultaneously. First, we introduce two enhanced combined-tumoral (EC) region modules, aiming to enhance the combined-tumoral features gradually. Then, we further design a three-branch module for extracting and fusing the features of intratumoral, peritumoral, and combined-tumoral regions, denoted as the intratumoral, peritumoral, and combined-tumoral module. Especially, we design a novel fusion module by introducing a channel attention module to adaptively fuse the features of three regions. The model is evaluated on two public datasets including UDIAT and BUSI with breast tumor ultrasound images. Two independent groups of experiments are performed on two respective datasets using the fivefold stratified cross-validation strategy. Finally, we conduct ablation experiments on two datasets, in which BUSI is used as the training set and UDIAT is used as the testing set.

RESULTS

We conduct detailed ablation experiments about the proposed two modules and comparative experiments with other existing representative methods. The experimental results show that the proposed method yields state-of-the-art performance on both two datasets. Especially, in the UDIAT dataset, the proposed FMRNet achieves a high accuracy of 0.945 and a specificity of 0.945, respectively. Moreover, the precision (PRE = 0.909) even dramatically improves by 21.6% on the BUSI dataset compared with the existing method of the best result.

CONCLUSION

The proposed FMRNet shows good performance in breast tumor classification with US images, and proves its capability of exploiting and fusing the information of multiple tumoral-regions. Furthermore, the FMRNet has potential value in classifying other types of cancers using multiple tumoral-regions of other kinds of medical images.

Cystic (including atypical) and solid breast lesion classification using the different features of quantitative ultrasound parametric images

PURPOSE

The amount of ultrasound (US) breast examinations continues to grow rapidly because of the wider endorsement of breast cancer screening programs. Cysts are the most commonly diagnosed breast lesions. Atypical breast cysts can be a serious differentiation problem in the US. Our goal was to develop noninvasive automated US grayscale image analysis for the cystic and solid breast lesion differentiation based on mathematical image post-processing.

MATERIALS AND METHODS

We used a set of 217 ultrasound images of proven 107 cystic (including 53 atypical) and 110 solid lesions. Empirical statistical and morphological models of the lesions were used to obtain features. The AUC indicator and Student's t test were used to assess the quality of the individual features. The Pearson correlation matrix was used to calculate the correlation between various features. The LASSO and stepwise regression methods were used to determine the most significant features. Finally, the lesion classification was carried out by the various methods.

RESULTS

The use of LASSO regression for the feature selection made it possible to select the most significant features for classification. The sensitivity increased from 87.1% to 89.2% and the specificity-from 92.2 to 94.8%. After the correlation matrix construction, it was found that features with a high value of the correlation coefficient (0.72; 0.75) can also be used to improve the quality of the classification.

CONCLUSION

The construction of the empirical model of the lesion pixels brightness behavior can provide parameters that are important for the correct classification of ultrasound images. The optimal set of features with the maximum discriminant characteristics may not be consistent with the correlation of features and the value of the AUC index. Features with a low AUC index (in our case 0.72) can also be important for improving the quality of the classification.

Comparison of Ultrasound Features With Maximum Standardized Uptake Value Assessed by 18F-Fluorodeoxyglucose-Positron Emission Tomography/Computed Tomography for Prognosis of Estrogen Receptor+/Human Epithelial Growth Factor Receptor 2- Breast Cancer

ABSTRACT

estrogen receptor (ER)+/human epithelial growth factor receptor 2 (HER2)- breast cancers have less aggressive traits and a favorable prognosis when treated early. Prediction of prognosis for treatment outcome or survival in ER+/HER2- cancer is important. Ultrasound (US) is an effective and easy technique for breast cancer diagnosis and tumor characterization. Positron emission tomography/computed tomography (PET/CT) is widely used for diagnosis, staging, and therapeutic response in cancer evaluation, and a high maximum standardized uptake value (SUVmax) is associated with poor prognosis. The study aim was to compare the prognostic value of US features with that of the SUVmax assessed by PET/CT in ER+/HER- breast cancer patients. We retrospectively identified breast cancer patients in our institutional database who had undergone preoperative US and PET/CT, and 96 patients with invasive ductal carcinoma and ductal carcinoma in situ were included in this study. The US features of mass shape, margin, echo pattern, orientation, posterior features, boundary, and calcification in the mass were analyzed. We then analyzed the US features to look for correlations with SUVmax and associations with margins, boundaries, posterior features, histological grade, and ki-67 expression. High SUVmax was correlated with irregular shape, not-circumscribed margin, posterior acoustic enhancement, echogenic halo, and calcification in the mass (P < 0.05, all). Posterior acoustic enhancement was correlated with high ki-67 expression. Many US features of ER+/HER- breast cancer showed associations with SUVmax. Some US features of ER+/HER- breast cancer were useful for predicting prognosis.

Assessment and Prediction of Response to Neoadjuvant Chemotherapy in Breast Cancer: A Comparison of Imaging Modalities and Future Perspectives

Neoadjuvant chemotherapy (NAC) is becoming the standard of care for locally advanced breast cancer, aiming to reduce tumor size before surgery. Unfortunately, less than 30% of patients generally achieve a pathological complete response and approximately 5% of patients show disease progression while receiving NAC. Accurate assessment of the response to NAC is crucial for subsequent surgical planning. Furthermore, early prediction of tumor response could avoid patients being overtreated with useless chemotherapy sections, which are not free from side effects and psychological implications. In this review, we first analyze and compare the accuracy of conventional and advanced imaging techniques as well as discuss the application of artificial intelligence tools in the assessment of tumor response after NAC. Thereafter, the role of advanced imaging techniques, such as MRI, nuclear medicine, and new hybrid PET/MRI imaging in the prediction of the response to NAC is described in the second part of the review. Finally, future perspectives in NAC response prediction, represented by AI applications, are discussed.

Diagnostic performance of perilesional radiomics analysis of contrast-enhanced mammography for the differentiation of benign and malignant breast lesions

OBJECTIVE

To conduct perilesional region radiomics analysis of contrast-enhanced mammography (CEM) images to differentiate benign and malignant breast lesions.

METHODS AND MATERIALS

This retrospective study included patients who underwent CEM from November 2017 to February 2020. Lesion contours were manually delineated. Perilesional regions were automatically obtained. Seven regions of interest (ROIs) were obtained for each lesion, including the lesion ROI, annular perilesional ROIs (1 mm, 3 mm, 5 mm), and lesion + perilesional ROIs (1 mm, 3 mm, 5 mm). Overall, 4,098 radiomics features were extracted from each ROI. Datasets were divided into training and testing sets (1:1). Seven classification models using features from the seven ROIs were constructed using LASSO regression. Model performance was assessed by the AUC with 95% CI.

RESULTS

Overall, 190 women with 223 breast lesions (101 benign; 122 malignant) were enrolled. In the testing set, the annular perilesional ROI of 3-mm model showed the highest AUC of 0.930 (95% CI: 0.882-0.977), followed by the annular perilesional ROI of 1 mm model (AUC = 0.929; 95% CI: 0.881-0.978) and the lesion ROI model (AUC = 0.909; 95% CI: 0.857-0.961). A new model was generated by combining the predicted probabilities of the lesion ROI and annular perilesional ROI of 3-mm models, which achieved a higher AUC in the testing set (AUC = 0.940).

CONCLUSIONS

Annular perilesional radiomics analysis of CEM images is useful for diagnosing breast cancers. Adding annular perilesional information to the radiomics model built on the lesion information may improve the diagnostic performance.

KEY POINTS

• Radiomics analysis of the annular perilesional region of 3 mm in CEM images may provide valuable information for the differential diagnosis of benign and malignant breast lesions. • The radiomics information from the lesion region and the annular perilesional region may be complementary. Combining the predicted probabilities of the models constructed by the features from the two regions may improve the diagnostic performance of radiomics models.

Differential diagnosis between small breast phyllodes tumors and fibroadenomas using artificial intelligence and ultrasound data

Background

It is challenging to differentiate between phyllodes tumors (PTs) and fibroadenomas (FAs). Artificial intelligence (AI) can provide quantitative information regarding the morphology and textural features of lesions. This study attempted to use AI to evaluate the ultrasonic images of PTs and FAs and to explore the diagnostic performance of AI features in the differential diagnosis of PTs and FAs.

Methods

A total of 40 PTs and 290 FAs <5 cm in maximum diameter found in female patients were retrospectively analyzed. All tumors were segmented by doctors, and the features of the lesions were collated, including circularity, height-to-width ratio, margin spicules, margin coarseness (MC), margin indistinctness, margin lobulation (ML), internal calcification, angle between the long axis of the lesion and skin, energy, grey entropy, and grey mean. The differences between PTs and FAs were analyzed, and the diagnostic performance of AI features in the differential diagnosis of PTs and FAs was evaluated.

Results

Statistically significant differences (P<0.05) were found in the height-to-width ratio, ML, energy, and grey entropy between the PTs and FAs. Receiver operating characteristic (ROC) curve analysis of single features showed that the area under the curve [(AUC) 0.759] of grey entropy was the largest among the four features with statistically significant differences, and the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were 0.925, 0.459, 0.978, and 0.190, respectively. When considering the combinations of the features, the combination of height-to-width ratio, margin indistinctness, ML, energy, grey entropy, and internal calcification was the most optimal of the combinations of features with an AUC of 0.868, and a sensitivity, specificity, PPV, and NPV of 0.734, 0.900, 0.982, and 0.316, respectively.

Conclusions

Quantitative analysis of AI can identify subtle differences in the morphology and textural features between small PTs and FAs. Comprehensive consideration of multiple features is important for the differential diagnosis of PTs and FAs.

A new CNN architecture for efficient classification of ultrasound breast tumor images with activation map clustering based prediction validation

Effective ultrasound (US) analysis for preliminary breast tumor diagnosis is constrained due to the presence of complex echogenic patterns. Implementing pretrained models of convolutional neural networks (CNNs) which mostly focuses on natural images and using transfer learning seldom gives good results in medical domain. In this work, a CNN architecture, StepNet, with step-wise incremental convolution layers for each downsampled block was developed for classification of breast tumors as benign/malignant. To increase noise robustness and as an improvement over existing methodologies, neutrosophic preprocessing was performed, and the enhanced images were appended to the original image during training and data augmentation. The final layers' activation maps are clustered using fuzzy c-means clustering which qualify as a validation method for the prediction of StepNet. Using neutrosophic preprocessing alone had increased the validation accuracy from 0.84 to 0.93, while using neutrosophic preprocessing and augmentation had increased the accuracy to 0.98. StepNet has comparably less training and validation time than other state of the art architectures and methods and shows an increase in prediction accuracy even for challenging isoechoic and hypoechoic tumors.

Texture Analysis in the Diagnosis of Primary Breast Cancer: Comparison of High-Resolution Dedicated Breast Positron Emission Tomography (dbPET) and Whole-Body PET/CT

Objective: This retrospective study aimed to compare the ability to classify tumor characteristics of breast cancer (BC) of positron emission tomography (PET)-derived texture features between dedicated breast PET (dbPET) and whole-body PET/computed tomography (CT).

Methods: Forty-four BCs scanned by both high-resolution ring-shaped dbPET and whole-body PET/CT were analyzed. The primary BC was extracted with a standardized uptake value (SUV) threshold segmentation method. On both dbPET and PET/CT images, 38 texture features were computed; their ability to classify tumor characteristics such as tumor (T)-category, lymph node (N)-category, molecular subtype, and Ki67 levels was compared. The texture features were evaluated using univariate and multivariate analyses following principal component analysis (PCA). AUC values were used to evaluate the diagnostic power of the computed texture features to classify BC characteristics.

Results: Some texture features of dbPET and PET/CT were different between Tis-1 and T2-4 and between Luminal A and other groups, respectively. No association with texture features was found in the N-category or Ki67 level. In contrast, receiver-operating characteristic analysis using texture features' principal components showed that the AUC for classification of any BC characteristics were equally good for both dbPET and whole-body PET/CT.

Conclusions: PET-based texture analysis of dbPET and whole-body PET/CT may have equally good classification power for BC.

A novel quantitative and reference-free ultrasound analysis to discriminate different concentrations of bone mineral content

Bone fracture is a continuous process, during which bone mineral matrix evolves leading to an increase in hydroxyapatite and calcium carbonate content. Currently, no gold standard methods are available for a quantitative assessment of bone fracture healing. Moreover, the available tools do not provide information on bone composition. Whereby, there is a need for objective and non-invasive methods to monitor the evolution of bone mineral content. In general, ultrasound can guarantee a quantitative characterization of tissues. However, previous studies required measurements on reference samples. In this paper we propose a novel and reference-free parameter, based on the entropy of the phase signal calculated from the backscattered data in combination with amplitude information, to also consider absorption and scattering phenomena. The proposed metric was effective in discriminating different hydroxyapatite (from 10 to 50% w/v) and calcium carbonate (from 2 to 6% w/v) concentrations in bone-mimicking phantoms without the need for reference measurements, paving the way to their translational use for the diagnosis of tissue healing. To the best of our knowledge this is the first time that the phase entropy of the backscattered ultrasound signals is exploited for monitoring changes in the mineral content of bone-like materials.

Comparison of methods for texture analysis of QUS parametric images in the characterization of breast lesions

PURPOSE

Accurate and timely diagnosis of breast carcinoma is very crucial because of its high incidence and high morbidity. Screening can improve overall prognosis by detecting the disease early. Biopsy remains as the gold standard for pathological confirmation of malignancy and tumour grading. The development of diagnostic imaging techniques as an alternative for the rapid and accurate characterization of breast masses is necessitated. Quantitative ultrasound (QUS) spectroscopy is a modality well suited for this purpose. This study was carried out to evaluate different texture analysis methods applied on QUS spectral parametric images for the characterization of breast lesions.

METHODS

Parametric images of mid-band-fit (MBF), spectral-slope (SS), spectral-intercept (SI), average scatterer diameter (ASD), and average acoustic concentration (AAC) were determined using QUS spectroscopy from 193 patients with breast lesions. Texture methods were used to quantify heterogeneities of the parametric images. Three statistical-based approaches for texture analysis that include Gray Level Co-occurrence Matrix (GLCM), Gray Level Run-length Matrix (GRLM), and Gray Level Size Zone Matrix (GLSZM) methods were evaluated. QUS and texture-parameters were determined from both tumour core and a 5-mm tumour margin and were used in comparison to histopathological analysis in order to classify breast lesions as either benign or malignant. We developed a diagnostic model using different classification algorithms including linear discriminant analysis (LDA), k-nearest neighbours (KNN), support vector machine with radial basis function kernel (SVM-RBF), and an artificial neural network (ANN). Model performance was evaluated using leave-one-out cross-validation (LOOCV) and hold-out validation.

RESULTS

Classifier performances ranged from 73% to 91% in terms of accuracy dependent on tumour margin inclusion and classifier methodology. Utilizing information from tumour core alone, the ANN achieved the best classification performance of 93% sensitivity, 88% specificity, 91% accuracy, 0.95 AUC using QUS parameters and their GLSZM texture features.

CONCLUSIONS

A QUS-based framework and texture analysis methods enabled classification of breast lesions with >90% accuracy. The results suggest that optimizing method for extracting discriminative textural features from QUS spectral parametric images can improve classification performance. Evaluation of the proposed technique on a larger cohort of patients with proper validation technique demonstrated the robustness and generalization of the approach.

Application of ultrasound artificial intelligence in the differential diagnosis between benign and malignant breast lesions of BI-RADS 4A

Background

The classification of Breast Imaging Reporting and Data System 4A (BI-RADS 4A) lesions is mostly based on the personal experience of doctors and lacks specific and clear classification standards. The development of artificial intelligence (AI) provides a new method for BI-RADS categorisation. We analysed the ultrasonic morphological and texture characteristics of BI-RADS 4A benign and malignant lesions using AI, and these ultrasonic characteristics of BI-RADS 4A benign and malignant lesions were compared to examine the value of AI in the differential diagnosis of BI-RADS 4A benign and malignant lesions.

Methods

A total of 206 lesions of BI-RADS 4A examined using ultrasonography were analysed retrospectively, including 174 benign lesions and 32 malignant lesions. All of the lesions were contoured manually, and the ultrasonic morphological and texture features of the lesions, such as circularity, height-to-width ratio, margin spicules, margin coarseness, margin indistinctness, margin lobulation, energy, entropy, grey mean, internal calcification and angle between the long axis of the lesion and skin, were calculated using grey level gradient co-occurrence matrix analysis. Differences between benign and malignant lesions of BI-RADS 4A were analysed.

Results

Significant differences in margin lobulation, entropy, internal calcification and ALS were noted between the benign group and malignant group (P = 0.013, 0.045, 0.045, and 0.002, respectively). The malignant group had more margin lobulations and lower entropy compared with the benign group, and the benign group had more internal calcifications and a greater angle between the long axis of the lesion and skin compared with the malignant group. No significant differences in circularity, height-to-width ratio, margin spicules, margin coarseness, margin indistinctness, energy, and grey mean were noted between benign and malignant lesions.

Conclusions

Compared with the naked eye, AI can reveal more subtle differences between benign and malignant BI-RADS 4A lesions. These results remind us carefully observation of the margin and the internal echo is of great significance. With the help of morphological and texture information provided by AI, doctors can make a more accurate judgment on such atypical benign and malignant lesions.

Quantitative ultrasound radiomics for therapy response monitoring in patients with locally advanced breast cancer: Multi-institutional study results

BACKGROUND

Neoadjuvant chemotherapy (NAC) is the standard of care for patients with locally advanced breast cancer (LABC). The study was conducted to investigate the utility of quantitative ultrasound (QUS) carried out during NAC to predict the final tumour response in a multi-institutional setting.

METHODS

Fifty-nine patients with LABC were enrolled from three institutions in North America (Sunnybrook Health Sciences Centre (Toronto, Canada), MD Anderson Cancer Centre (Texas, USA), and Princess Margaret Cancer Centre (Toronto, Canada)). QUS data were collected before starting NAC and subsequently at weeks 1 and 4 during chemotherapy. Spectral tumour parametric maps were generated, and textural features determined using grey-level co-occurrence matrices. Patients were divided into two groups based on their pathological outcomes following surgery: responders and non-responders. Machine learning algorithms using Fisher's linear discriminant (FLD), K-nearest neighbour (K-NN), and support vector machine (SVM-RBF) were used to generate response classification models.

RESULTS

Thirty-six patients were classified as responders and twenty-three as non-responders. Among all the models, SVM-RBF had the highest accuracy of 81% at both weeks 1 and week 4 with area under curve (AUC) values of 0.87 each. The inclusion of week 1 and 4 features led to an improvement of the classifier models, with the accuracy and AUC from baseline features only being 76% and 0.68, respectively.

CONCLUSION

QUS data obtained during NAC reflect the ongoing treatment-related changes during chemotherapy and can lead to better classifier performances in predicting the ultimate pathologic response to treatment compared to baseline features alone.

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.