Radiomic Analysis in Contrast-Enhanced Spectral Mammography for Predicting Breast Cancer Histological Outcome

Recent Radiomics Advancements in Breast Cancer: Lessons and Pitfalls for the Next Future

Radiomics is an emerging translational field of medicine based on the extraction of high-dimensional data from radiological images, with the purpose to reach reliable models to be applied into clinical practice for the purposes of diagnosis, prognosis and evaluation of disease response to treatment. We aim to provide the basic information on radiomics to radiologists and clinicians who are focused on breast cancer care, encouraging cooperation with scientists to mine data for a better application in clinical practice. We investigate the workflow and clinical application of radiomics in breast cancer care, as well as the outlook and challenges based on recent studies. Currently, radiomics has the potential ability to distinguish between benign and malignant breast lesions, to predict breast cancer’s molecular subtypes, the response to neoadjuvant chemotherapy and the lymph node metastases. Even though radiomics has been used in tumor diagnosis and prognosis, it is still in the research phase and some challenges need to be faced to obtain a clinical translation. In this review, we discuss the current limitations and promises of radiomics for improvement in further research.

A Review of Breast Imaging for Timely Diagnosis of Disease

Breast cancer (BC) is the cancer with the highest incidence in women in the world. In this last period, the COVID-19 pandemic has caused in many cases a drastic reduction of routine breast imaging activity due to the combination of various factors. The survival of BC is directly proportional to the earliness of diagnosis, and especially during this period, it is at least fundamental to remember that a diagnostic delay of even just three months could affect BC outcomes. In this article we will review the state of the art of breast imaging, starting from morphological imaging, i.e., mammography, tomosynthesis, ultrasound and magnetic resonance imaging and contrast-enhanced mammography, and their most recent evolutions; and ending with functional images, i.e., magnetic resonance imaging and contrast enhanced mammography.

Early Prediction of Breast Cancer Recurrence for Patients Treated with Neoadjuvant Chemotherapy: A Transfer Learning Approach on DCE-MRIs

Cancer treatment planning benefits from an accurate early prediction of the treatment efficacy. The goal of this study is to give an early prediction of three-year Breast Cancer Recurrence (BCR) for patients who underwent neoadjuvant chemotherapy. We addressed the task from a new perspective based on transfer learning applied to pre-treatment and early-treatment DCE-MRI scans. Firstly, low-level features were automatically extracted from MR images using a pre-trained Convolutional Neural Network (CNN) architecture without human intervention. Subsequently, the prediction model was built with an optimal subset of CNN features and evaluated on two sets of patients from I-SPY1 TRIAL and BREAST-MRI-NACT-Pilot public databases: a fine-tuning dataset (70 not recurrent and 26 recurrent cases), which was primarily used to find the optimal subset of CNN features, and an independent test (45 not recurrent and 17 recurrent cases), whose patients had not been involved in the feature selection process. The best results were achieved when the optimal CNN features were augmented by four clinical variables (age, ER, PgR, HER2+), reaching an accuracy of 91.7% and 85.2%, a sensitivity of 80.8% and 84.6%, a specificity of 95.7% and 85.4%, and an AUC value of 0.93 and 0.83 on the fine-tuning dataset and the independent test, respectively. Finally, the CNN features extracted from pre-treatment and early-treatment exams were revealed to be strong predictors of BCR.

Radiomics and Artificial Intelligence Analysis with Textural Metrics Extracted by Contrast-Enhanced Mammography in the Breast Lesions Classification

The aim of the study was to estimate the diagnostic accuracy of textural features extracted by dual-energy contrast-enhanced mammography (CEM) images, by carrying out univariate and multivariate statistical analyses including artificial intelligence approaches. In total, 80 patients with known breast lesion were enrolled in this prospective study according to regulations issued by the local Institutional Review Board. All patients underwent dual-energy CEM examination in both craniocaudally (CC) and double acquisition of mediolateral oblique (MLO) projections (early and late). The reference standard was pathology from a surgical specimen for malignant lesions and pathology from a surgical specimen or fine needle aspiration cytology, core or Tru-Cut needle biopsy, and vacuum assisted breast biopsy for benign lesions. In total, 104 samples of 80 patients were analyzed. Furthermore, 48 textural parameters were extracted by manually segmenting regions of interest. Univariate and multivariate approaches were performed: non-parametric Wilcoxon–Mann–Whitney test; receiver operating characteristic (ROC), linear classifier (LDA), decision tree (DT), k-nearest neighbors (KNN), artificial neural network (NNET), and support vector machine (SVM) were utilized. A balancing approach and feature selection methods were used. The univariate analysis showed low accuracy and area under the curve (AUC) for all considered features. Instead, in the multivariate textural analysis, the best performance considering the CC view (accuracy (ACC) = 0.75; AUC = 0.82) was reached with a DT trained with leave-one-out cross-variation (LOOCV) and balanced data (with adaptive synthetic (ADASYN) function) and a subset of three robust textural features (MAD, VARIANCE, and LRLGE). The best performance (ACC = 0.77; AUC = 0.83) considering the early-MLO view was reached with a NNET trained with LOOCV and balanced data (with ADASYN function) and a subset of ten robust features (MEAN, MAD, RANGE, IQR, VARIANCE, CORRELATION, RLV, COARSNESS, BUSYNESS, and STRENGTH). The best performance (ACC = 0.73; AUC = 0.82) considering the late-MLO view was reached with a NNET trained with LOOCV and balanced data (with ADASYN function) and a subset of eleven robust features (MODE, MEDIAN, RANGE, RLN, LRLGE, RLV, LZLGE, GLV_GLSZM, ZSV, COARSNESS, and BUSYNESS). Multivariate analyses using pattern recognition approaches, considering 144 textural features extracted from all three mammographic projections (CC, early MLO, and late MLO), optimized by adaptive synthetic sampling and feature selection operations obtained the best results (ACC = 0.87; AUC = 0.90) and showed the best performance in the discrimination of benign and malignant lesions.

Radiomic Feature Reduction Approach to Predict Breast Cancer by Contrast-Enhanced Spectral Mammography Images

Contrast-enhanced spectral mammography (CESM) is an advanced instrument for breast care that is still operator dependent. The aim of this paper is the proposal of an automated system able to discriminate benign and malignant breast lesions based on radiomic analysis. We selected a set of 58 regions of interest (ROIs) extracted from 53 patients referred to Istituto Tumori "Giovanni Paolo II" of Bari (Italy) for the breast cancer screening phase between March 2017 and June 2018. We extracted 464 features of different kinds, such as points and corners of interest, textural and statistical features from both the original ROIs and the ones obtained by a Haar decomposition and a gradient image implementation. The features data had a large dimension that can affect the process and accuracy of cancer classification. Therefore, a classification scheme for dimension reduction was needed. Specifically, a principal component analysis (PCA) dimension reduction technique that includes the calculation of variance proportion for eigenvector selection was used. For the classification method, we trained three different classifiers, that is a random forest, a naïve Bayes and a logistic regression, on each sub-set of principal components (PC) selected by a sequential forward algorithm. Moreover, we focused on the starting features that contributed most to the calculation of the related PCs, which returned the best classification models. The method obtained with the aid of the random forest classifier resulted in the best prediction of benign/malignant ROIs with median values for sensitivity and specificity of 88.37% and 100%, respectively, by using only three PCs. The features that had shown the greatest contribution to the definition of the same were almost all extracted from the LE images. Our system could represent a valid support tool for radiologists for interpreting CESM images.

Response Predictivity to Neoadjuvant Therapies in Breast Cancer: A Qualitative Analysis of Background Parenchymal Enhancement in DCE-MRI

Background: For assessing the predictability of oncology neoadjuvant therapy results, the background parenchymal enhancement (BPE) parameter in breast magnetic resonance imaging (MRI) has acquired increased interest. This work aims to qualitatively evaluate the BPE parameter as a potential predictive marker for neoadjuvant therapy. Method: Three radiologists examined, in triple-blind modality, the MRIs of 80 patients performed before the start of chemotherapy, after three months from the start of treatment, and after surgery. They identified the portion of fibroglandular tissue (FGT) and BPE of the contralateral breast to the tumor in the basal control pre-treatment (baseline). Results: We observed a reduction of BPE classes in serial MRI checks performed during neoadjuvant therapy, as compared to baseline pre-treatment conditions, in 61.3% of patients in the intermediate step, and in 86.7% of patients in the final step. BPE reduction was significantly associated with sequential anthracyclines/taxane administration in the first cycle of neoadjuvant therapy compared to anti-HER2 containing therapies. The therapy response was also significantly related to tumor size. There were no associations with menopausal status, fibroglandular tissue (FGT) amount, age, BPE baseline, BPE in intermediate, and in the final MRI step. Conclusions: The measured variability of this parameter during therapy could predict therapy effectiveness in early stages, improving decision-making in the perspective of personalized medicine. Our preliminary results suggest that BPE may represent a predictive factor in response to neoadjuvant therapy in breast cancer, warranting future investigations in conjunction with radiomics.

Quantitative analysis of contrast enhanced spectral mammography grey value for early prediction of pathological response of breast cancer to neoadjuvant chemotherapy

A quantitative analysis of contrast-enhanced spectral mammography (CESM) enhancement was conducted for the early prediction of the pathological response after neoadjuvant chemotherapy (NAC). Retrospective analysis of the data of 111 patients was conducted, and all of them underwent NAC in our hospital and surgical resection after the end of all cycles from January 2018 to May 2019. They were divided into pathological complete response (PCR) and non-PCR groups. We determined whether a statistical difference in the percentage of CESM grey value reduction (ΔCGV) was present in the PCR and non-PCR groups and whether a statistical difference was observed in the diagnostic efficiency of craniocaudal (CC) and mediolateral oblique (MLO) view subtraction images. Independent sample t-test was used to compare different groups, the receiver operating characteristic (ROC) curve was used to compare the diagnostic efficacy of CC and MLO for pathological response after NAC, and the Delong test was used to compare the area under the ROC curve (AUC). Statistical significance was considered at P < 0.05. A statistical difference was observed in the ΔCGV in the PCR and non-PCR groups. No statistical difference was observed in the AUCs of CC and MLO view subtraction images. The ΔCGV can be used as a quantitative index to predict PCR early, and no statistical difference was observed in the diagnostic efficacy of CC and MLO view subtraction images.

A Proposal of Quantum-Inspired Machine Learning for Medical Purposes: An Application Case

Learning tasks are implemented via mappings of the sampled data set, including both the classical and the quantum framework. Biomedical data characterizing complex diseases such as cancer typically require an algorithmic support for clinical decisions, especially for early stage tumors that typify breast cancer patients, which are still controllable in a therapeutic and surgical way. Our case study consists of the prediction during the pre-operative stage of lymph node metastasis in breast cancer patients resulting in a negative diagnosis after clinical and radiological exams. The classifier adopted to establish a baseline is characterized by the result invariance for the order permutation of the input features, and it exploits stratifications in the training procedure. The quantum one mimics support vector machine mapping in a high-dimensional feature space, yielded by encoding into qubits, while being characterized by complexity. Feature selection is exploited to study the performances associated with a low number of features, thus implemented in a feasible time. Wide variations in sensitivity and specificity are observed in the selected optimal classifiers during cross-validations for both classification system types, with an easier detection of negative or positive cases depending on the choice between the two training schemes. Clinical practice is still far from being reached, even if the flexible structure of quantum-inspired classifier circuits guarantees further developments to rule interactions among features: this preliminary study is solely intended to provide an overview of the particular tree tensor network scheme in a simplified version adopting just product states, as well as to introduce typical machine learning procedures consisting of feature selection and classifier performance evaluation.

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.

Contrast-Enhanced Ultrasonography in the Diagnosis and Treatment Modulation of Breast Cancer

The aim of this paper is to highlight the role of contrast-enhanced ultrasound in breast cancer in terms of diagnosis, staging and follow-up of the post-treatment response. Contrast-enhanced ultrasound (CEUS) is successfully used to diagnose multiple pathologies and has also clinical relevance in breast cancer. CEUS has high accuracy in differentiating benign from malignant lesions by analyzing the enhancement characteristics and calculating the time-intensity curve’s quantitative parameters. It also has a significant role in axillary staging, especially when the lymph nodes are not suspicious on clinical examination and have a normal appearance on gray-scale ultrasound. The most significant clinical impact consists of predicting the response to neoadjuvant chemotherapy, which offers the possibility of adjusting the therapy by dynamically evaluating the patient. CEUS is a high-performance, feasible, non-irradiating, accessible, easy-to-implement imaging method and has proven to be a valuable addition to breast ultrasound.

Breast Cancer Type Classification Using Machine Learning

Background: Breast cancer is a heterogeneous disease defined by molecular types and subtypes. Advances in genomic research have enabled use of precision medicine in clinical management of breast cancer. A critical unmet medical need is distinguishing triple negative breast cancer, the most aggressive and lethal form of breast cancer, from non-triple negative breast cancer. Here we propose use of a machine learning (ML) approach for classification of triple negative breast cancer and non-triple negative breast cancer patients using gene expression data. Methods: We performed analysis of RNA-Sequence data from 110 triple negative and 992 non-triple negative breast cancer tumor samples from The Cancer Genome Atlas to select the features (genes) used in the development and validation of the classification models. We evaluated four different classification models including Support Vector Machines, K-nearest neighbor, Naïve Bayes and Decision tree using features selected at different threshold levels to train the models for classifying the two types of breast cancer. For performance evaluation and validation, the proposed methods were applied to independent gene expression datasets. Results: Among the four ML algorithms evaluated, the Support Vector Machine algorithm was able to classify breast cancer more accurately into triple negative and non-triple negative breast cancer and had less misclassification errors than the other three algorithms evaluated. Conclusions: The prediction results show that ML algorithms are efficient and can be used for classification of breast cancer into triple negative and non-triple negative breast cancer types.

Predicting of Sentinel Lymph Node Status in Breast Cancer Patients with Clinically Negative Nodes: A Validation Study

Simple Summary

Sentinel lymph node biopsy procedure is time consuming and expensive, but it is still the intra-operative exam capable of the best performance. However, sometimes, surgery is achieved without a clear diagnosis, so clinical decision support systems developed with artificial intelligence techniques are essential to assist current diagnostic procedures. In this work, we evaluated the usefulness of a CancerMath tool in the sentinel lymph nodes positivity prediction for clinically negative patients. We tested it on 993 patients referred to our institute characterized by sentinel lymph node status, tumor size, age, histologic type, grading, expression of estrogen receptor, progesterone receptor, HER2, and Ki-67. By training the CancerMath (CM) model on our dataset, we reached a sensitivity value of 72%, whereas the online one was 46%, despite a specificity reduction. It was found the addiction of the prognostic factors Her2 and Ki67 could help improve performances on the classification of particular types of patients.

Abstract

In the absence of lymph node abnormalities detectable on clinical examination or imaging, the guidelines provide for the dissection of the first axillary draining lymph nodes during surgery. It is not always possible to arrive at surgery without diagnostic doubts, and machine learning algorithms can support clinical decisions. The web calculator CancerMath (CM) allows you to estimate the probability of having positive lymph nodes valued on the basis of tumor size, age, histologic type, grading, expression of estrogen receptor, and progesterone receptor. We collected 993 patients referred to our institute with clinically negative results characterized by sentinel lymph node status, prognostic factors defined by CM, and also human epidermal growth factor receptor 2 (HER2) and Ki-67. Area Under the Curve (AUC) values obtained by the online CM application were comparable with those obtained after training its algorithm on our database. Nevertheless, by training the CM model on our dataset and using the same feature, we reached a sensitivity median value of 72%, whereas the online one was equal to 46%, despite a specificity reduction. We found that the addition of the prognostic factors Her2 and Ki67 could help improve performances on the classification of particular types of patients with the aim of reducing as much as possible the false positives that lead to axillary dissection. As showed by our experimental results, it is not particularly suitable for use as a support instrument for the prediction of metastatic lymph nodes on clinically negative patients.

Tumour Stroma Ratio Assessment Using Digital Image Analysis Predicts Survival in Triple Negative and Luminal Breast Cancer

We aimed to determine the clinical significance of tumour stroma ratio (TSR) in luminal and triple negative breast cancer (TNBC) using digital image analysis and machine learning algorithms. Automated image analysis using QuPath software was applied to a cohort of 647 breast cancer patients (403 luminal and 244 TNBC) using digital H&E images of tissue microarrays (TMAs). Kaplan-Meier and Cox proportional hazards were used to ascertain relationships with overall survival (OS) and breast cancer specific survival (BCSS). For TNBC, low TSR (high stroma) was associated with poor prognosis for both OS (HR 1.9, CI 1.1-3.3, p = 0.021) and BCSS (HR 2.6, HR 1.3-5.4, p = 0.007) in multivariate models, independent of age, size, grade, sTILs, lymph nodal status and chemotherapy. However, for luminal tumours, low TSR (high stroma) was associated with a favourable prognosis in MVA for OS (HR 0.6, CI 0.4-0.8, p = 0.001) but not for BCSS. TSR is a prognostic factor of most significance in TNBC, but also in luminal breast cancer, and can be reliably assessed using quantitative image analysis of TMAs. Further investigation into the contribution of tumour subtype stromal phenotype may further refine these findings.