Predicting the molecular subtype of breast cancer based on mammography and ultrasound findings

Correlations of Imaging and Therapy in Breast Cancer Based on Molecular Patterns: An Important Issue in the Diagnosis of Breast Cancer

Breast cancer is a global health issue affecting countries worldwide, imposing a significant economic burden due to expensive treatments and medical procedures, given the increasing incidence. In this review, our focus is on exploring the distinct imaging features of known molecular subtypes of breast cancer, underlining correlations observed in clinical practice and reported in recent studies. The imaging investigations used for assessment include screening modalities such as mammography and ultrasonography, as well as more complex investigations like MRI, which offers high sensitivity for loco-regional evaluation, and PET, which determines tumor metabolic activity using radioactive tracers. The purpose of this review is to provide a better understanding as well as a revision of the imaging differences exhibited by the molecular subtypes and histopathological types of breast cancer.

Assessing the Influence of B-US, CDFI, SE, and Patient Age on Predicting Molecular Subtypes in Breast Lesions Using Deep Learning Algorithms

OBJECTIVES

Our study aims to investigate the impact of B-mode ultrasound (B-US) imaging, color Doppler flow imaging (CDFI), strain elastography (SE), and patient age on the prediction of molecular subtypes in breast lesions.

METHODS

Totally 2272 multimodal ultrasound imaging was collected from 198 patients. The ResNet-18 network was employed to predict four molecular subtypes from B-US imaging, CDFI, and SE of patients with different ages. All the images were split into training and testing datasets by the ratio of 80%:20%. The predictive performance on testing dataset was evaluated through 5 metrics including mean accuracy, precision, recall, F1-scores, and confusion matrix.

RESULTS

Based on B-US imaging, the test mean accuracy is 74.50%, the precision is 74.84%, the recall is 72.48%, and the F1-scores is 0.73. By combining B-US imaging with CDFI, the results were increased to 85.41%, 85.03%, 85.05%, and 0.84, respectively. With the integration of B-US imaging and SE, the results were changed to 75.64%, 74.69%, 73.86%, and 0.74, respectively. Using images from patients under 40 years old, the results were 90.48%, 90.88%, 88.47%, and 0.89. When images from patients who are above 40 years old, they were changed to 81.96%, 83.12%, 80.5%, and 0.81, respectively.

CONCLUSION

Multimodal ultrasound imaging can be used to accurately predict the molecular subtypes of breast lesions. In addition to B-US imaging, CDFI rather than SE contribute further to improve predictive performance. The predictive performance is notably better for patients under 40 years old compared with those who are 40 years old and above.

Predicting hormone receptor status in invasive breast cancer through radiomics analysis of long-axis and short-axis ultrasound planes

The hormone receptor (HR) status plays a significant role in breast cancer, serving as the primary guide for treatment decisions and closely correlating with prognosis. This study aims to investigate the predictive value of radiomics analysis in long-axis and short-axis ultrasound planes for distinguishing between HR-positive and HR-negative breast cancers. A cohort of 505 patients from two hospitals was stratified into discovery (Institute 1, 416 patients) and validation (Institute 2, 89 patients) cohorts. A comprehensive set of 788 ultrasound radiomics features was extracted from both long-axis and short-axis ultrasound planes, respectively. Utilizing least absolute shrinkage and selection operator (LASSO) regression analysis, distinct models were constructed for the long-axis and short-axis data. Subsequently, radiomics scores (Rad-scores) were computed for each patient. Additionally, a combined model was formulated by integrating data from long-axis and short-axis Rad-scores along with clinical factors. The diagnostic efficacy of all models was evaluated by the area under the receiver operating characteristic (ROC) curve (AUC). The long-axis and short-axis models, consisting of 11 features and 15 features, respectively, were established, yielding AUCs of 0.743 and 0.751 in the discovery cohort, and 0.795 and 0.744 in the validation cohort. The calculated long-axis and short-axis Rad-scores exhibited significant differences between HR-positive and HR-negative groups across all cohorts (all p < 0.001). Univariate analysis identified ultrasound-reported tumor size as an independent predictor. The combined model, incorporating long-axis and short-axis Rad-scores along with tumor size, achieved superior AUCs of 0.788 and 0.822 in the discovery and validation cohorts, respectively. The combined model effectively distinguishes between HR-positive and HR-negative breast cancers based on ultrasound radiomics features and tumor size, which may offer a valuable tool to facilitate treatment decision making and prognostic assessment.

Breast cancer stage and molecular subtype distribution: real-world insights from a regional oncological center in Hungary

OBJECTIVE

Examining the distribution of breast cancer (BC) stage and molecular subtype among women aged below (< 45 years), within (45-65 years), and above (> 65 years) the recommended screening age range helps to understand the screening program's characteristics and contributes to enhancing the effectiveness of BC screening programs.

METHODS

In this retrospective study, female patients with newly diagnosed BC from 2010 to 2020 were identified. The distribution of cases in terms of TNM stages, severity classes, and subtypes was analysed according to age groups.

RESULTS

A total of 3282 women diagnosed with BC were included in the analysis. Among these cases 51.4% were detected outside the screening age group, and these were characterized by a higher TNM stage compared to those diagnosed within the screening age band. We observed significantly higher relative frequency of advanced BC in the older age group compared to both the screening age population and women younger than 45 years (14.9% vs. 8.7% and 7.7%, P < 0.001). HR-/HER2- and HER+ tumours were relatively more frequent among women under age 45 years (HR-/HER2-: 23.6%, HER2+: 20.5%) compared to those within the screening age range (HR-/HER2-: 13.4%, HER2+: 13.9%) and the older age group (HR-/HER2-: 10.4%, HER2+: 11.5%).

CONCLUSIONS

The findings of our study shed light on potential areas for the improvement of BC screening programs (e.g., extending screening age group, adjusting screening frequency based on molecular subtype risk status) in Hungary and internationally, as well.

Breast Cancer Molecular Subtype Prediction: A Mammography-Based AI Approach

Breast cancer remains a leading cause of mortality among women, with molecular subtypes significantly influencing prognosis and treatment strategies. Currently, identifying the molecular subtype of cancer requires a biopsy-a specialized, expensive, and time-consuming procedure, often yielding to results that must be supported with additional biopsies due to technique errors or tumor heterogeneity. This study introduces a novel approach for predicting breast cancer molecular subtypes using mammography images and advanced artificial intelligence (AI) methodologies. Using the OPTIMAM imaging database, 1397 images from 660 patients were selected. The pretrained deep learning model ResNet-101 was employed to classify tumors into five subtypes: Luminal A, Luminal B1, Luminal B2, HER2, and Triple Negative. Various classification strategies were studied: binary classifications (one vs. all others, specific combinations) and multi-class classification (evaluating all subtypes simultaneously). To address imbalanced data, strategies like oversampling, undersampling, and data augmentation were explored. Performance was evaluated using accuracy and area under the receiver operating characteristic curve (AUC). Binary classification results showed a maximum average accuracy and AUC of 79.02% and 64.69%, respectively, while multi-class classification achieved an average AUC of 60.62% with oversampling and data augmentation. The most notable binary classification was HER2 vs. non-HER2, with an accuracy of 89.79% and an AUC of 73.31%. Binary classification for specific combinations of subtypes revealed an accuracy of 76.42% for HER2 vs. Luminal A and an AUC of 73.04% for HER2 vs. Luminal B1. These findings highlight the potential of mammography-based AI for non-invasive breast cancer subtype prediction, offering a promising alternative to biopsies and paving the way for personalized treatment plans.

Differentiation Between Phyllodes Tumors and Fibroadenomas of Breast Using Mammography-based Machine Learning Methods: A Preliminary Study

OBJECTIVE

To investigate the diagnostic performance of a mammography-based radiomics model for distinguishing phyllodes tumors (PTs) from fibroadenomas (FAs) of the breast.

MATERIALS AND METHODS

A total of 156 patients were retrospectively included (75 with PTs, 81 with FAs) and divided into training and validation groups at a ratio of 7:3. Radiomics features were extracted from craniocaudal and mediolateral oblique images. The least absolute shrinkage and selection operator (LASSO) algorithm and principal component analysis (PCA) were performed to select features. Three machine learning classifiers, including logistic regression (LR), K-nearest neighbor classifier (KNN) and support vector machine (SVM), were implemented in the radiomics model, imaging model and combined model. Receiver operating characteristic curves, area under the curve (AUC), sensitivity and specificity were computed.

RESULTS

Among 1084 features, the LASSO algorithm selected 17 features, and PCA further selected 6 features. Three machine learning classifiers yielded the same AUC of 0.935 in the validation group for the radiomics model. In the imaging model, KNN yielded the highest accuracy rate of 89.4% and AUC of 0.947 in the validation set. For the combined model, the SVM classifier reached the highest AUC of 0.918 with an accuracy rate of 86.2%, sensitivity of 83.9%, and specificity of 89.4% in the training group. In the validation group, LR yielded the highest AUC of 0.973. The combined model had a relatively higher AUC than the radiomics model or imaging model, especially in the validation group.

CONCLUSIONS

Mammography-based radiomics features demonstrate good diagnostic performance for discriminating PTs from FAs.

Deep learning radiopathomics based on preoperative US images and biopsy whole slide images can distinguish between luminal and non-luminal tumors in early-stage breast cancers

BACKGROUND

For patients with early-stage breast cancers, neoadjuvant treatment is recommended for non-luminal tumors instead of luminal tumors. Preoperative distinguish between luminal and non-luminal cancers at early stages will facilitate treatment decisions making. However, the molecular immunohistochemical subtypes based on biopsy specimens are not always consistent with final results based on surgical specimens due to the high intra-tumoral heterogeneity. Given that, we aimed to develop and validate a deep learning radiopathomics (DLRP) model to preoperatively distinguish between luminal and non-luminal breast cancers at early stages based on preoperative ultrasound (US) images, and hematoxylin and eosin (H&E)-stained biopsy slides.

METHODS

This multicentre study included three cohorts from a prospective study conducted by our team and registered on the Chinese Clinical Trial Registry (ChiCTR1900027497). Between January 2019 and August 2021, 1809 US images and 603 H&E-stained whole slide images (WSIs) from 603 patients with early-stage breast cancers were obtained. A Resnet18 model pre-trained on ImageNet and a multi-instance learning based attention model were used to extract the features of US and WSIs, respectively. An US-guided Co-Attention module (UCA) was designed for feature fusion of US and WSIs. The DLRP model was constructed based on these three feature sets including deep learning US feature, deep learning WSIs feature and UCA-fused feature from a training cohort (1467 US images and 489 WSIs from 489 patients). The DLRP model's diagnostic performance was validated in an internal validation cohort (342 US images and 114 WSIs from 114 patients) and an external test cohort (270 US images and 90 WSIs from 90 patients). We also compared diagnostic efficacy of the DLRP model with that of deep learning radiomics model and deep learning pathomics model in the external test cohort.

FINDINGS

The DLRP yielded high performance with area under the curve (AUC) values of 0.929 (95% CI 0.865-0.968) in the internal validation cohort, and 0.900 (95% CI 0.819-0.953) in the external test cohort. The DLRP also outperformed deep learning radiomics model based on US images only (AUC 0.815 [0.719-0.889], p = 0.027) and deep learning pathomics model based on WSIs only (AUC 0.802 [0.704-0.878], p = 0.013) in the external test cohort.

INTERPRETATION

The DLRP can effectively distinguish between luminal and non-luminal breast cancers at early stages before surgery based on pretherapeutic US images and biopsy H&E-stained WSIs, providing a tool to facilitate treatment decision making in early-stage breast cancers.

FUNDING

Natural Science Foundation of Guangdong Province (No. 2023A1515011564), and National Natural Science Foundation of China (No. 91959127; No. 81971631).

Preoperative prediction of Ki-67 status in invasive breast carcinoma using dynamic contrast-enhanced MRI, diffusion-weighted imaging and diffusion tensor imaging

Background

The Ki-67 is a beneficial marker of tumor aggressiveness. It is proliferation index that has been used to distinguish luminal B from luminal A breast cancers. By fast progress in quantitative radiology modalities, tumor biology and genetics can be assessed in a more accurate, predictive, and cost-effective method. The aim of this study was to assess the role of dynamic contrast-enhanced magnetic resonance imaging, diffusion-weighted imaging and diffusion tensor imaging in prediction of Ki-67 status in patients with invasive breast carcinoma estimate cut off values between breast cancer with high Ki-67 status and those with low Ki-67 status.

Results

Cut off ADC (apparent diffusion co-efficient) value of 0.657 mm2/s had 96.4% sensitivity, 75% specificity and 93.8% accuracy in differentiating cases with high Ki67 from those with low Ki67. Cut off maximum enhancement value of 1715 had 96.4% sensitivity, 75% specificity and 93.8% accuracy in differentiating cases with high Ki67 from those with low Ki67. Cut off washout rate of 0.73 I/S had 60.7% sensitivity, 75% specificity and 62.5% accuracy in differentiating cases with high Ki67 from those with low Ki67. Cut off time to peak value of 304 had 71.4% sensitivity, 75% specificity and 71.9% accuracy in differentiating cases with high Ki67 from those with low Ki67.

Conclusions

ADC, time to peak and maximum enhancement values had high sensitivity, specificity and accuracy in differentiating breast cancer with high Ki-67 status from those with low Ki-67 status.

Clinicopathological and ultrasound characteristics of breast cancer in BRCA1 and BRCA2 mutation carriers

PURPOSE

BRCA1 and BRCA2 tumors exhibit different characteristics. This study aimed to assess and compare the ultrasound findings and pathologic features of BRCA1 and BRCA2 breast cancers. To our knowledge, this is the first study to examine the mass formation, vascularity, and elasticity in breast cancers of BRCA-positive Japanese women.

METHODS

We identified patients with breast cancer harboring BRCA1 or BRCA2 mutations. After excluding patients who underwent chemotherapy or surgery before the ultrasound, we evaluated 89 cancers in BRCA1-positive and 83 in BRCA2-positive patients. The ultrasound images were reviewed by three radiologists in consensus. Imaging features, including vascularity and elasticity, were assessed. Pathological data, including tumor subtypes, were reviewed.

RESULTS

Significant differences in tumor morphology, peripheral features, posterior echoes, echogenic foci, and vascularity were observed between BRCA1 and BRCA2 tumors. BRCA1 breast cancers tended to be posteriorly accentuating and hypervascular. In contrast, BRCA2 tumors were less likely to form masses. In cases where a tumor formed a mass, it tended to show posterior attenuation, indistinct margins, and echogenic foci. In pathological comparisons, BRCA1 cancers tended to be triple-negative subtypes. In contrast, BRCA2 cancers tended to be luminal or luminal-human epidermal growth factor receptor 2 subtypes.

CONCLUSION

In the surveillance of BRCA mutation carriers, radiologists should be aware that the morphological differences between tumors are quite different between BRCA1 and BRCA2 patients.

Mammography-based radiomics analysis and imaging features for predicting the malignant risk of phyllodes tumours of the breast

AIM

To determine whether the mammography (MG)-based radiomics analysis and MG/ultrasound (US) imaging features could predict the malignant risk of phyllodes tumours (PTs) of the breast.

MATERIALS AND METHODS

Seventy-five patients with PTs were included retrospectively (39 with benign PTs, 36 with borderline/malignant PTs) and divided into thetraining (n=52) and validation groups (n=23). The clinical information, MG and US imaging characteristics, and histogram features were extracted from craniocaudal (CC) and mediolateral oblique (MLO) images. The lesion region of interest (ROI) and perilesional ROI were delineated. Multivariate logistic regression analysis was performed to determine the malignant factors of PTs. Receiver operating characteristic (ROC) curves were generated, and the area under the curve (AUC), sensitivity, and specificity were calculated.

RESULTS

There was no significant difference found in the clinical or MG/US features between benign and borderline/malignant PTs. In the lesion ROI, variance in the CC view and mean and variance in the MLO view were independent predictors. The AUC was 0.942, sensitivity and specificity were 96.3% and 92%, respectively, in the training group. In the validation group, the AUC was 0.879, the sensitivity was 91.7%, and the specificity was 81.8%. In the perilesional ROI, the AUCs were 0.904 and 0.939, sensitivities were 88.9% and 91.7%, and the specificities were 92% and 90.9% in the training and validation groups, respectively.

CONCLUSIONS

MG-based radiomic features could predict the risk of malignancy of patients with PTs and may be used as a potential tool to differentiate benign and borderline/malignant PTs.

Classifying Breast Cancer Metastasis Based on Imaging of Tumor Primary and Tumor Biology

The molecular classification of breast cancer has allowed for a better understanding of both prognosis and treatment of breast cancer. Imaging of the different molecular subtypes has revealed that biologically different tumors often exhibit typical features in mammography, ultrasound, and MRI. Here, we introduce the molecular classification of breast cancer and review the typical imaging features of each subtype, examining the predictive value of imaging with respect to distant metastases.

Machine learning analysis of breast ultrasound to classify triple negative and HER2+ breast cancer subtypes

OBJECTIVES

Early diagnosis of triple-negative (TN) and human epidermal growth factor receptor 2 positive (HER2+) breast cancer is important due to its increased risk of micrometastatic spread necessitating early treatment and for guiding targeted therapies. This study aimed to evaluate the diagnostic performance of machine learning (ML) classification of newly diagnosed breast masses into TN versus non-TN (NTN) and HER2+ versus HER2 negative (HER2-) breast cancer, using radiomic features extracted from grayscale ultrasound (US) b-mode images.

MATERIALS AND METHODS

A retrospective chart review identified 88 female patients who underwent diagnostic breast US imaging, had confirmation of invasive malignancy on pathology and receptor status determined on immunohistochemistry available. The patients were classified as TN, NTN, HER2+ or HER2- for ground-truth labelling. For image analysis, breast masses were manually segmented by a breast radiologist. Radiomic features were extracted per image and used for predictive modelling. Supervised ML classifiers included: logistic regression, k-nearest neighbour, and Naïve Bayes. Classification performance measures were calculated on an independent (unseen) test set. The area under the receiver operating characteristic curve (AUC), sensitivity (%), and specificity (%) were reported for each classifier.

RESULTS

The logistic regression classifier demonstrated the highest AUC: 0.824 (sensitivity: 81.8%, specificity: 74.2%) for the TN sub-group and 0.778 (sensitivity: 71.4%, specificity: 71.6%) for the HER2 sub-group.

CONCLUSION

ML classifiers demonstrate high diagnostic accuracy in classifying TN versus NTN and HER2+ versus HER2- breast cancers using US images. Identification of more aggressive breast cancer subtypes early in the diagnostic process could help achieve better prognoses by prioritizing clinical referral and prompting adequate early treatment.

Preoperative predictive factors affecting sentinel lymph node positivity in breast cancer and comparison of their effectiveness with existing nomograms

This study aimed to establish a strong regression model by revealing the preoperative predictive factors for sentinel lymph node (SLN) positivity in patients with early stage breast cancer (ESBC). In total, 445 patients who underwent SLN dissection for ESBC were included. All data that may be potential predictors of SLN positivity were retrospectively analyzed. Tumor size >2 cm, human epidermal growth factor receptor 2 (HER2) + status, lymphovascular invasion (LVI), palpable tumor, microcalcifications, multifocality or multicentricity, and axillary ultrasonographic findings were defined as independent predictors of SLN involvement. The area under the receiver operating characteristic (ROC) curve (AUC) values were 0.797, 0.808, and 0.870 for the Memorial Sloan-Kettering Cancer Center (MSKCC) nomogram, MD Anderson Cancer Center (MDACC) nomogram, and our regression model, respectively (P < .001). The recent model for predicting SLN status in ESBC was found to be stronger than existing nomograms. Parameters not included in current nomograms, such as palpable tumors, microcalcifications, and axillary ultrasonographic findings, are likely to make this model more meaningful.

Imaging Features of Breast Cancer Subtypes on Mammography and Ultrasonography: an Analysis of 479 Patients

To compare features of clinically defined subtypes of breast cancer on mammography (MG) and ultrasonography (USG). After obtaining approval from the institute ethics committee, a retrospective observational study was performed on biopsy-proven breast cancer patients who underwent baseline MG from 2016 to 2020. MG and USG features were evaluated and the patients were classified based on immunohistochemistry profile into luminal like (LL)-oestrogen receptor (ER)/progesterone receptor (PR) + , Her2neu-; basal like (BL)-ER/PR-, Her2neu-; Her2 like (HL)-Her2neu + . A total of 479 patients (mean age, 51.4 ± 11.7 years; all females) were included: LL-198 (41.3%), BL-121 (25.2%) and HL-160 (33.3%). On MG, round shape (21/115, 18.3%, p < 0.001); circumscribed (16/115, 13.9%, p < 0.001) and microlobulated margins (28/115, 24.4%) were associated with BL tumours. Associated suspicious calcifications (96/160, 60%, p < 0.001) and skin thickening or retraction (75/149, 50.3%, p < 0.001) were more common in HL. On USG, round shape (12/95, 12.8%, p = 0.005); circumscribed (8/94, 8.5%) and microlobulated margins (44/94, 46.8%) and posterior acoustic enhancement (7/95, 7.5%, p = 0.012) were associated with BL. The logistic regression analysis revealed that spiculated margins on MG favoured LL (OR: 8.5, p = 0.001); round shape (OR: 6.8), circumscribed (OR: 10.8) or microlobulated margins (OR: 3.5) (p < 0.001 for each) favoured BL; whereas associated features of calcifications (OR: 3.3) (p = 0.019) and skin retraction or thickening (OR: 1.8) (p < 0.001) favoured HL. On USG, circumscribed (OR: 5.9, p = 0.005) or microlobulated margins (OR: 3, p < 0.001) and posterior acoustic enhancement (OR: 9.5, p = 0.006) favoured BL. Clinically defined subtypes of breast cancer show significant differences in the imaging appearances on mammography and USG. BL tumours may not show the typical imaging features of malignancy, necessitating clinicopathological correlation for accurate diagnosis.

Association Between Contrast-Enhanced Ultrasound Characteristics and Molecular Subtypes of Breast Cancer

OBJECTIVE

This study aimed to investigate the correlation between contrast-enhanced ultrasound (CEUS) features and molecular subtypes of breast cancer (BC).

METHODS

A total of 116 patients (116 lesions) with pathologically diagnosed BC who received conventional ultrasound and CEUS before surgery were enrolled in this study. BC molecular subtypes were identified by postoperative pathological and immunohistochemical analysis as Luminal A (LA), Luminal B (LB), HER2 (H2) over-expression, and triple-negative (TN). Qualitative and quantitative CEUS characteristics were analyzed by one-way analysis of variance (continuous variables) or Pearson's χ² test or Fisher's exact probability method (categorical variables).

RESULTS

There were significant differences in enhancement speed and enhancement degree among the four subtypes (P < .05). The area under the curve (AUC), time to peak (TTP), and peak intensity (PI) differed among the four subtypes (P < .05). The AUC of the LA subtype (305.1 ± 188.4) was significantly smaller compared with the H2 (535.7 ± 222.0, P = .007) and TN subtypes (496.6 ± 254.7, P = .019). In addition, TTP was shorter in the H2 subtype (19.8 ± 4.9) compared with the other subtypes, and was significantly shorter than in the LA subtype (26.3 ± 7.2, P = .008) and LB subtype (23.1 ± 6.7, P = .036). The PI of the LA subtype (4.7 ± 2.3) was significantly lower than that of the LB (6.6 ± 2.3, P = .027), H2 (7.4 ± 2.2, P = .005), and TN subtypes (6.9 ± 2.6, P = .014).

CONCLUSIONS

CEUS features differed significantly among different molecular subtypes of BC. The enhancement patterns and parameters may be important predictive features of different subtypes of BC.

Stiffness in breast masses with posterior acoustic shadowing: significance of ultrasound real time shear wave elastography

Background

To assess the stiffness of benign breast masses in ultrasound images with posterior acoustic shadowing (PAS) and malignant lesions, and explore the significance of differential diagnosis using ultrasound real time shear wave elastography.

Material and methods

All 117 mammary masses (98 patients) with PAS were assessed by using routine ultrasound examination, and elastic modulus values were obtained with the real time shear wave elastography mode. All breast lesions were confirmed by surgery or biopsy. The significance of differences in ultrasound elastography values between breast benign and malignant masses with posterior acoustic shadowing was assessed, and the ROC curves of elasticity modulus values were analyzed.

Results

Among the 117 masses, 72 were benign and 45 were malignant. The two types of breast masses showed significant differences in size, margin, internal echo, calcification, and blood flow characteristics (P < 0.05), although the difference in orientation was not significant (P > 0.05). Emean, Emax and Esd obtained with real time shear wave elastography showed statistically significant differences between benign masses with posterior acoustic shadowing and breast cancer (P < 0.05), while Emin showed no significant difference between them (P = 0.633). Ultrasound real time shear wave elastography showed higher sensitivity and specificity than conventional ultrasound.

Conclusions

Benign and malignant breast masses with PAS show different ultrasound manifestations. Real time shear wave elastography can facilitate the differential diagnosis and treatment planning for these breast masses.

Application of preoperative ultrasound features combined with clinical factors in predicting HER2-positive subtype (non-luminal) breast cancer

BACKGROUND

Human epidermal growth factor receptor2+ subtype breast cancer has a high degree of malignancy and a poor prognosis. The aim of this study is to develop a prediction model for the human epidermal growth factor receptor2+ subtype (non-luminal) of breast cancer based on the clinical and ultrasound features related with estrogen receptor, progesterone receptor, and human epidermal growth factor receptor2.

METHODS

We collected clinical data and reviewed preoperative ultrasound images of enrolled breast cancers from September 2017 to August 2020. We divided the data into in three groups as follows. Group I: estrogen receptor ± , Group II: progesterone receptor ± and Group III: human epidermal growth factor receptor2 ± . Univariate and multivariate logistic regression analyses were used to analyze the clinical and ultrasound features related with biomarkers among these groups. A model to predict human epidermal growth factor receptor2+ subtype was then developed based on the results of multivariate regression analyses, and the efficacy was evaluated using the area under receiver operating characteristic curve, accuracy, sensitivity, specificity.

RESULTS

The human epidermal growth factor receptor2+ subtype accounted for 138 cases (11.8%) in the training set and 51 cases (10.1%) in the test set. In the multivariate regression analysis, age ≤ 50 years was an independent predictor of progesterone receptor + (p = 0.007), and posterior enhancement was a negative predictor of progesterone receptor + (p = 0.013) in Group II; palpable axillary lymph node, round, irregular shape and calcifications were independent predictors of the positivity for human epidermal growth factor receptor-2 in Group III (p = 0.001, p = 0.007, p = 0.010, p < 0.001, respectively). In Group I, shape was the only factor related to estrogen receptor status in the univariate analysis (p < 0.05). The area under receiver operating characteristic curve, accuracy, sensitivity, specificity of the model to predict human epidermal growth factor receptor2+ subtype breast cancer was 0.697, 60.14%, 72.46%, 58.49% and 0.725, 72.06%, 64.71%, 72.89% in the training and test sets, respectively.

CONCLUSIONS

Our study established a model to predict the human epidermal growth factor receptor2-positive subtype with moderate performance. And the results demonstrated that clinical and ultrasound features were significantly associated with biomarkers.

Predicting the molecular subtype of breast cancer and identifying interpretable imaging features using machine learning algorithms

OBJECTIVES

To evaluate the performance of interpretable machine learning models in predicting breast cancer molecular subtypes.

METHODS

We retrospectively enrolled 600 patients with invasive breast carcinoma between 2012 and 2019. The patients were randomly divided into a training (n = 450) and a testing (n = 150) set. The five constructed models were trained based on clinical characteristics and imaging features (mammography and ultrasonography). The model classification performances were evaluated using the area under the receiver operating characteristic (ROC) curve (AUC), accuracy, sensitivity, and specificity. Shapley additive explanation (SHAP) technique was used to interpret the optimal model output. Then we choose the optimal model as the assisted model to evaluate the performance of another four radiologists in predicting the molecular subtype of breast cancer with or without model assistance, according to mammography and ultrasound images.

RESULTS

The decision tree (DT) model performed the best in distinguishing triple-negative breast cancer (TNBC) from other breast cancer subtypes, yielding an AUC of 0.971; accuracy, 0.947; sensitivity, 0.905; and specificity, 0.941. The accuracy, sensitivity, and specificity of all radiologists in distinguishing TNBC from other molecular subtypes and Luminal breast cancer from other molecular subtypes have significantly improved with the assistance of DT model. In the diagnosis of TNBC versus other subtypes, the average sensitivity, average specificity, and average accuracy of less experienced and more experienced radiologists increased by 0.090, 0.125, 0.114, and 0.060, 0.090, 0.083, respectively. In the diagnosis of Luminal versus other subtypes, the average sensitivity, average specificity, and average accuracy of less experienced and more experienced radiologists increased by 0.084, 0.152, 0.159, and 0.020, 0.100, 0.048.

CONCLUSIONS

This study established an interpretable machine learning model to differentiate between breast cancer molecular subtypes, providing additional values for radiologists.

KEY POINTS

• Interpretable machine learning model (MLM) could help clinicians and radiologists differentiate between breast cancer molecular subtypes. • The Shapley additive explanations (SHAP) technique can select important features for predicting the molecular subtypes of breast cancer from a large number of imaging signs. • Machine learning model can assist radiologists to evaluate the molecular subtype of breast cancer to some extent.

Analysis of breast cancer subtypes and their correlations with receptors and ultrasound

The study aim was to evaluate the ultrasound (US) signs of the mammary lesions classified in the Breast Imaging-Reporting and Data System (BI-RADS) score category 3, 4, and 5, corresponding to US BI-RADS. It also followed the correlation between US changes of lesions suggestive for malignancy with the histopathological results and evaluated the proper management of those lesions. There were correlations of breast cancer (BC) subtypes with the receptors [estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (HER2)], and Ki67 index, and the signs of conventional ultrasonography and US elastography. We selected 108 female patients examined with US, mammography and fine-needle biopsy who presented suspicions for malignancy lesions. Following the immunohistochemical analysis, they were classified in one of the BC subtypes. According to chi-squared analysis of molecular cancer subtypes correlation to receptors and Ki67 index, we found significant associations between both luminal A and luminal B HER2-negative subtypes and hormone receptors (ER, PR). These have an inverse relationship with Ki67 index elevated values; luminal B HER2-positive subtype has a direct association with HER2 presence; HER2-enriched subtype was statistically significant associated to HER2 presence and elevated Ki67 index values but had an inverse relationship to hormone receptors (ER, PR); triple-negative subtype was strongly associated to Ki67 index values and inversely correlated to ER and PR. We found luminal A subtype as being the most common and luminal B HER2-positive subtype as having the fewer cases.

Clinico-radio-pathological Features and Biological Behavior of Breast Cancer in Young Indian Women: A Prospective Study

Aims The aim of the study is to evaluate the characteristic imaging features of breast cancer on mammogram, ultrasound, and magnetic resonance imaging (MRI) in women less than 40 years of age and to assess the degree of correlation between clinico-radio-pathological features and biological behavior. Methods and Materials A prospective observational study on consecutive women under 40 years of age evaluated with ultrasound of breast, digital mammogram, or contrast-enhanced breast MRI, diagnosed with breast cancer on histopathology and molecular analysis done at our center between January and December 2019 were included. Patient demographics, clinical presentation, family history, BRCA mutation status, imaging, pathological findings, and molecular status were determined. Results Out of 2,470 women diagnosed with breast cancer, 354 (14.3%) were less than 40 years of age who were included in this study. Mammography showed positive findings in 85%, ultrasonography in 94.3%, and MRI in 96.4% of women. Majority of the women (69.6%) presented in the late stage (Stage III and IV) with high-grade carcinoma in 39.5% and triple-negative breast cancer (TNBC) in 45.7%. Tumors with human epidermal growth factor-2neu expression were associated with the presence of microcalcifications ( p -value = 0.006), and TNBC with circumscribed margins or BI-RADS 3/4a category on imaging ( p -value = 0.007) and high-grade invasive carcinoma compared with others ( p -value <0.0001). Conclusion The incidence of breast cancer in Indian women less than 40 years of age is relatively high as compared with the West. The detection of breast cancer in young women remains challenging due to dense breast tissue, lower incidence rate, and lack of regular breast screening. While ultrasound is the recommended imaging method for evaluation of breast under the age of 40 years, we found a better characterization of lesions and higher cancer detection rates when they were also evaluated with mammography and MRI.

Role of mammogram and ultrasound imaging in predicting breast cancer subtypes in screening and symptomatic patients

BACKGROUND

Breast cancer (BC) radiogenomics, or correlation analysis of imaging features and BC molecular subtypes, can complement genetic analysis with less resource-intensive diagnostic methods to provide an early and accurate triage of BC. This is pertinent because BC is the most prevalent cancer amongst adult women, resulting in rising demands on public health resources.

AIM

To find combinations of mammogram and ultrasound imaging features that predict BC molecular subtypes in a sample of screening and symptomatic patients.

METHODS

This retrospective study evaluated 328 consecutive patients in 2017-2018 with histologically confirmed BC, of which 237 (72%) presented with symptoms and 91 (28%) were detected via a screening program. All the patients underwent mammography and ultrasound imaging prior to biopsy. The images were retrospectively read by two breast-imaging radiologists with 5-10 years of experience with no knowledge of the histology results to ensure statistical independence. To test the hypothesis that imaging features are correlated with tumor subtypes, univariate binomial and multinomial logistic regression models were performed. Our study also used the multivariate logistic regression (with and without interaction terms) to identify combinations of mammogram and ultrasound (US) imaging characteristics predictive of molecular subtypes.

RESULTS

The presence of circumscribed margins, posterior enhancement, and large size is correlated with triple-negative BC (TNBC), while high-risk microcalcifications and microlobulated margins is predictive of HER2-enriched cancers. Ductal carcinoma in situ is characterized by small size on ultrasound, absence of posterior acoustic features, and architectural distortion on mammogram, while luminal subtypes tend to be small, with spiculated margins and posterior acoustic shadowing (Luminal A type). These results are broadly consistent with findings from prior studies. In addition, we also find that US size signals a higher odds ratio for TNBC if presented during screening. As TNBC tends to display sonographic features such as circumscribed margins and posterior enhancement, resulting in visual similarity with benign common lesions, at the screening stage, size may be a useful factor in deciding whether to recommend a biopsy.

CONCLUSION

Several imaging features were shown to be independent variables predicting molecular subtypes of BC. Knowledge of such correlations could help clinicians stratify BC patients, possibly enabling earlier treatment or aiding in therapeutic decisions in countries where receptor testing is not readily available.

Comparison of clinicopathological characteristics and prognosis in breast cancer patients with different Breast Imaging Reporting and Data System categories

Purpose

The Breast Imaging Reporting and Data System (BI-RADS) is a systematic and standardized scheme of the radiological findings of breast. However, there were different BI-RADS categories between breast cancers as the clinical characteristics in previous studies. We analyzed the association of BI-RADS categories with the clinicopathological characteristics and prognosis of breast cancer.

Methods

A total of 44,184 patients with invasive breast cancers assigned to BI-RADS category 3, 4, or 5 in preoperative mammography or ultrasonography were analyzed retrospectively using large-scale data from the Korean Breast Cancer Society registration system. The difference in the clinicopathological factors and prognoses according to the BI-RADS categories (BI-RADS 3–4 and BI-RADS 5) were compared between the mammography and ultrasonography groups. Comparisons of the clinicopathological factors in both groups were made using logistic regression analysis, while the prognoses were based on the breast cancer-specific survival using the Kaplan-Meier method and Cox proportional hazards model.

Results

The factors associated with BI-RADS were T stage, N stage, palpability, histology grade, and lymphovascular invasion in the mammography group; and N stage, palpability, histology grade, and lymphovascular invasion in the ultrasonography group. In the survival analysis, there were significant differences in the breast cancer-specific survival of the BI-RADS category groups in both of the mammography (hazard ratio [HR], 3.366; P < 0.001) and ultrasonography (HR, 2.877; P < 0.001) groups.

Conclusion

In this study, the BI-RADS categories of preoperative mammography and ultrasonography of patients with invasive breast cancer were associated with prognosis and could be an important factor in making treatment decisions.

Predicting Molecular Subtypes of Breast Cancer with Mammography and Ultrasound Findings: Introduction of Sono-Mammometry Score

We studied the correlation of sonographic and digital mammographic features with molecular classification of breast cancer. Imaging features from 313 patients with preliminary ultrasound and digital mammogram between November 2017 and May 2020 were compared with histopathology and immunohistochemical analysis for the prediction of molecular classification of breast cancer. We also devised a score called “sono-mammometry” score consisting of few simple imaging features which can easily be performed in outpatient settings. We studied that non-triple-negative breast cancers are predominantly hypoechoic and strongly correlate with the presence of irregular spiculated margins along with peripheral echogenic halo, posterior shadowing, and microcalcifications, while there is considerable variation in imaging features of TNBC as some of its imaging features overlap with those of typical benign tumors. Although imaging characteristics are helpful in the prediction of molecular classification, the prognostication value of these imaging features is still weak. There is considerable variation in imaging features which warrants vigilance towards improved diagnostic performance. To help better understand these features, our sono-mammometry score can serve as straightforward test which is assumed to be functional and productive in resource-limited settings.

Deep learning with convolutional neural network in the assessment of breast cancer molecular subtypes based on US images: a multicenter retrospective study

OBJECTIVES

To evaluate the prediction performance of deep convolutional neural network (DCNN) based on ultrasound (US) images for the assessment of breast cancer molecular subtypes.

METHODS

A dataset of 4828 US images from 1275 patients with primary breast cancer were used as the training samples. DCNN models were constructed primarily to predict the four St. Gallen molecular subtypes and secondarily to identify luminal disease from non-luminal disease based on the ground truth from immunohistochemical of whole tumor surgical specimen. US images from two other institutions were retained as independent test sets to validate the system. The models' performance was analyzed using per-class accuracy, positive predictive value (PPV), and Matthews correlation coefficient (MCC).

RESULTS

The model achieved good performance in identifying the four breast cancer molecular subtypes in the two test sets, with accuracy ranging from 80.07% (95% CI, 76.49-83.23%) to 97.02% (95% CI, 95.22-98.16%) and 87.94% (95% CI, 85.08-90.31%) to 98.83% (95% CI, 97.60-99.43) for the two test cohorts for each sub-category, respectively. In terms of 4-class weighted average MCC, the model achieved 0.59 for test cohort A and 0.79 for test cohort B. Specifically, the DCNN also yielded good diagnostic performance in discriminating luminal disease from non-luminal disease, with a PPV of 93.29% (95% CI, 90.63-95.23%) and 88.21% (95% CI, 85.12-90.73%) for the two test cohorts, respectively.

CONCLUSION

Using pretreatment US images of the breast cancer, deep learning model enables the assessment of molecular subtypes with high diagnostic accuracy.

TRIAL REGISTRATION

Clinical trial number: ChiCTR1900027676 KEY POINTS: • Deep convolutional neural network (DCNN) helps clinicians assess tumor features with accuracy. • Multicenter retrospective study shows that DCNN derived from pretreatment ultrasound imagine improves the prediction of breast cancer molecular subtypes. • Management of patients becomes more precise based on the DCNN model.

Prevalence of Molecular Subtypes of Breast Cancer in India: a Systematic Review and Meta-analysis

In the last two decades, India has witnessed a substantial increase in the incidence of breast cancer and associated mortality. Studies on the prevalence of molecular subtypes of breast cancer in India have reported inconsistent results. Therefore, we conducted a systematic review of observational studies to document the prevalence of molecular subtypes of breast cancer. A complete literature search for observational studies was conducted in MEDLINE and EMBASE databases using key MeSH terms ((molecular classification) OR (molecular subtypes)) AND (breast cancer)) OR (breast carcinoma)) AND (prevalence)) AND (India). Two reviewers independently reviewed the retrieved studies. The screened studies satisfying the eligibility were included. The quality of included studies was assessed using the selected STROBE criteria. The overall pooled prevalence of luminal A, luminal B, HER2-enriched, and triple-negative breast cancer (TNBC) subtypes of breast cancer were 0.33 (95% CI 0.23-0.44), 0.17 (95% CI 0.12-0.23), 0.15 (95% CI 0.12-0.19), and 0.30 (95% CI 0.27-0.33), respectively. Subgroup analyses were performed by mean age of patients, time period, region, and sample size of the study. Among molecular subtypes of breast cancer, luminal A was the most prevalent subtype followed by TNBC, luminal B, and HER2-enriched subtypes. The overall prevalence of TNBC in India is high compared to other regions of the world. Additional research is warranted to identify the determinants of high TNBC in India. Differentiating TNBC from other molecular subtypes is important to guide therapeutic management of breast cancer.

Current Landscape of Breast Cancer Imaging and Potential Quantitative Imaging Markers of Response in ER-Positive Breast Cancers Treated with Neoadjuvant Therapy

In recent years, neoadjuvant treatment trials have shown that breast cancer subtypes identified on the basis of genomic and/or molecular signatures exhibit different response rates and recurrence outcomes, with the implication that subtype-specific treatment approaches are needed. Estrogen receptor-positive (ER+) breast cancers present a unique set of challenges for determining optimal neoadjuvant treatment approaches. There is increased recognition that not all ER+ breast cancers benefit from chemotherapy, and that there may be a subset of ER+ breast cancers that can be treated effectively using endocrine therapies alone. With this uncertainty, there is a need to improve the assessment and to optimize the treatment of ER+ breast cancers. While pathology-based markers offer a snapshot of tumor response to neoadjuvant therapy, non-invasive imaging of the ER disease in response to treatment would provide broader insights into tumor heterogeneity, ER biology, and the timing of surrogate endpoint measurements. In this review, we provide an overview of the current landscape of breast imaging in neoadjuvant studies and highlight the technological advances in each imaging modality. We then further examine some potential imaging markers for neoadjuvant treatment response in ER+ breast cancers.

Breast cancer imaging features as a predictor of the hormonal receptor status, HER2neu expression and molecular subtype

Background

Determination of the hormonal receptor (HR) status, HER2neu expression, and the molecular subtype has valuable diagnostic, therapeutic, and prognostic implications for breast cancer as breast cancer stratification during the last two decades has become dependent upon the underlying biology. The aim of this study is to assess the correlation between imaging features of breast cancer and the HR status, HER2neu expression, and the molecular subtype. Sixty breast cancer patients underwent breast ultrasound, mammography, and MRI evaluation. Pathological evaluation using immunohistochemistry and FISH was used to detect the HR status, HER2/neu expression, and the molecular subtype. Those findings were then correlated with the radiologic data.

Results

HR-positive tumors were associated with posterior acoustic shadowing (34/44, 77.3%; p = 0.004). Hormonal-negative tumors presenting as masses were more likely circumscribed on US and MRI compared to hormonal positive mass tumors (6/14, 42.9% vs 3/36, 7.7%; p = 0.003 on US and 6/13, 46.3% vs 3/36, 8.3%; P = 0.007 on MRI) and had malignant DCE kinetics with washout curves compared to the hormonal positive group (10/16, 62.5% vs 4/44, 9.1%; P < 0.001). HER2neu-positive tumors were significantly associated with calcifications and multifocality on mammography compared to HER2neu-negative group (9/13, 69% vs 12/34, 25.5%; P = 0.007) and (7/13, 53% vs 3/47, 6%; P < 0.001). TNBC and HER2neu-enriched were associated with washout kinetic curve pattern (57.1% and 66.7%, respectively). TNBCs were associated with circumscribed margins on US and MRI (6/9, 66.7%; P < 0.001).

Conclusion

Microcalcifications, margins, posterior acoustic features, and malignant washout kinetics strongly correlate with the hormonal receptor status, HER2neu status, and molecular subtype of breast cancer. These findings may suggest the molecular subtype of breast cancer and further expand the role of imaging.

Diagnostic value of diffusion-weighted imaging-derived apparent diffusion coefficient and its association with histological prognostic factors in breast cancer

Diffusion-weighted imaging (DWI) has been proven to be effective in detecting breast malignancies and has been widely implemented for breast imaging. However, the exact association between certain DWI biomarkers and well-known prognostic factors remains to be fully elucidated. By studying the association between the apparent diffusion coefficient (ADC) and prognostic factors, the present study aimed to explore the diagnostic value and prognostic potential of the ADC in breast lesions. The study included 539 female subjects with histopathologically confirmed breast lesions who underwent DWI of the breast tissue. The diagnoses comprised 307 subjects with malignant breast tumors and 232 with benign breast tumors. The maximum ADC and mean ADC (ADC_mean) values of the breast lesions were calculated. For malignant tumors, the association between ADC and major prognostic factors, including histological grade, nuclear grade and lymph node status, as well as estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (HER-2) and proliferation marker protein Ki-67.(Ki-67) status, were evaluated. The ADC_mean demonstrated the best diagnostic performance in distinguishing between malignant and benign lesions. With the optimum cut-off value at 1.30×10⁻³ mm²/sec, ADC_mean had a sensitivity and specificity of 84.1 and 90.2%, respectively. In those patients with malignant breast lesions, a decreased ADC was associated with breast lesions with high nuclear and histological grades, and lymph node-positive, ER-negative, PR-negative and HER-2-negative status, and Ki-67 ≥14%. In conclusion, the ADC is a useful imaging biomarker for differentiating between benign and malignant breast tumors. The marked association between the ADC and prognostic factors also demonstrated its value in evaluating the malignancy of breast lesions.

A different methylation profile of circadian genes promoter in breast cancer patients according to clinicopathological features

One of the supposed mechanisms that may lead to breast cancer (BC) is an alteration of circadian gene expression and DNA methylation. We undertook an integrated approach to identify methylation pattern of core circadian promoter regions in BC patients with regard to clinical features. We performed a quantitative methylation-specific real-time PCR analysis of a promoter methylation profile in 107 breast tumor and matched non-tumor tissues. A panel of circadian genes CLOCK, BMAL1, PERIOD (PER1, 2, 3), CRYPTOCHROME (CRY1, 2) and TIMELESS as well as their association with clinicopathological characteristics were included in the analysis. Three out of the eight analyzed genes exhibited marked hypermethylation (PER1, 2, 3), whereas CLOCK, BMAL1, CRY2 showed significantly lower promoter CpG methylation in the BC tissues when compared to the non-tumor tissues. Among variously methylated genes we found an association between the elevated methylation level of PERs promoter region and molecular subtypes, histological subtypes and tumor grading of BC. Methylation status may be associated with a gene expression level of circadian genes in BC patients. An aberrant methylation pattern in circadian genes in BC may provide information that could be used as novel biomarkers in clinics and molecular epidemiology as well as play an important role in BC etiology.

In-cell determination of Lactate Dehydrogenase Activity in a Luminal Breast Cancer Model ⁻ ex vivo Investigation of Excised Xenograft Tumor Slices Using dDNP Hyperpolarized [1-13C]pyruvate

[1-¹³C]pyruvate, the most widely used compound in dissolution-dynamic nuclear polarization (dDNP) magnetic resonance (MR), enables the visualization of lactate dehydrogenase (LDH) activity. This activity had been demonstrated in a wide variety of cancer models, ranging from cultured cells, to xenograft models, to human tumors in situ. Here we quantified the LDH activity in precision cut tumor slices (PCTS) of breast cancer xenografts. The Michigan Cancer Foundation-7 (MCF7) cell-line was chosen as a model for the luminal breast cancer type which is hormone responsive and is highly prevalent. The LDH activity, which was manifested as [1-¹³C]lactate production in the tumor slices, ranged between 3.8 and 6.1 nmole/nmole adenosine tri-phosphate (ATP) in 1 min (average 4.6 ± 1.0) on three different experimental set-ups consisting of arrested vs. continuous perfusion and non-selective and selective RF pulsation schemes and combinations thereof. This rate was converted to an expected LDH activity in a mass ranging between 3.3 and 5.2 µmole/g in 1 min, using the ATP level of these tumors. This indicated the likely utility of this approach in clinical dDNP of the human breast and may be useful as guidance for treatment response assessment in a large number of tumor types and therapies ex vivo.