CADx of mammographic masses and clustered microcalcifications: A review

CAPNet: Context attention pyramid network for computer-aided detection of microcalcification clusters in digital breast tomosynthesis

BACKGROUND AND OBJECTIVE

Computer-aided detection (CADe) of microcalcification clusters (MCs) in digital breast tomosynthesis (DBT) is crucial in the early diagnosis of breast cancer. Although convolutional neural network (CNN)-based detection models have achieved excellent performance in medical lesion detection, they are subject to some limitations in MC detection: 1) Most existing models employ the feature pyramid network (FPN) for multi-scale object detection; however, the rough feature sharing between adjacent layers in the FPN may limit the detection ability for small and low-contrast MCs; and 2) the MCs region only accounts for a small part of the annotation box, so the features extracted indiscriminately within the whole box may easily be affected by the background. In this paper, we develop a novel CNN-based CADe method to alleviate the impacts of the above limitations for the accurate and rapid detection of MCs in DBT.

METHODS

The proposed method has two parts: a novel context attention pyramid network (CAPNet) for intra-layer MC detection in two-dimensional (2D) slices and a three-dimensional (3D) aggregation procedure for aggregating 2D intra-layer MCs into a 3D result according to their connectivity in 3D space. The proposed CAPNet is based on an anchor-free and one-stage detection architecture and contains a context feature selection fusion (CFSF) module and a microcalcification response (MCR) branch. The CFSF module can efficiently enrich shallow layers' features by the complementary selection of local context features, aiming to reduce the missed detection of small and low-contrast MCs. The MCR branch is a one-layer branch parallel to the classification branch, which can alleviate the influence of the background region within the annotation box on feature extraction and enhance the ability of the model to distinguish MCs from normal breast tissue.

RESULTS

We performed a comparison experiment on an in-house clinical dataset with 648 DBT volumes, and the proposed method achieved impressive performance with a sensitivity of 91.56 % at 1 false positive per DBT volume (FPs/volume) and 93.51 % at 2 FPs/volume, outperforming other representative detection models.

CONCLUSIONS

The experimental results indicate that the proposed method is effective in the detection of MCs in DBT. This method can provide objective, accurate, and quick diagnostic suggestions for radiologists, presenting potential clinical value for early breast cancer screening.

FSE-Net: feature selection and enhancement network for mammogram classification

Objective. Early detection and diagnosis allow for intervention and treatment at an early stage of breast cancer. Despite recent advances in computer aided diagnosis systems based on convolutional neural networks for breast cancer diagnosis, improving the classification performance of mammograms remains a challenge due to the various sizes of breast lesions and difficult extraction of small lesion features. To obtain more accurate classification results, many studies choose to directly classify region of interest (ROI) annotations, but labeling ROIs is labor intensive. The purpose of this research is to design a novel network to automatically classify mammogram image as cancer and no cancer, aiming to mitigate or address the above challenges and help radiologists perform mammogram diagnosis more accurately.Approach. We propose a novel feature selection and enhancement network (FSE-Net) to fully exploit the features of mammogram images, which requires only mammogram images and image-level labels without any bounding boxes or masks. Specifically, to obtain more contextual information, an effective feature selection module is proposed to adaptively select the receptive fields and fuse features from receptive fields of different scales. Moreover, a feature enhancement module is designed to explore the correlation between feature maps of different resolutions and to enhance the representation capacity of low-resolution feature maps with high-resolution feature maps.Main results. The performance of the proposed network has been evaluated on the CBIS-DDSM dataset and INbreast dataset. It achieves an accuracy of 0.806 with an AUC of 0.866 on the CBIS-DDSM dataset and an accuracy of 0.956 with an AUC of 0.974 on the INbreast dataset.Significance. Through extensive experiments and saliency map visualization analysis, the proposed network achieves the satisfactory performance in the mammogram classification task, and can roughly locate suspicious regions to assist in the final prediction of the entire images.

A divide and conquer approach to maximise deep learning mammography classification accuracies

Breast cancer claims 11,400 lives on average every year in the UK, making it one of the deadliest diseases. Mammography is the gold standard for detecting early signs of breast cancer, which can help cure the disease during its early stages. However, incorrect mammography diagnoses are common and may harm patients through unnecessary treatments and operations (or a lack of treatment). Therefore, systems that can learn to detect breast cancer on their own could help reduce the number of incorrect interpretations and missed cases. Various deep learning techniques, which can be used to implement a system that learns how to detect instances of breast cancer in mammograms, are explored throughout this paper. Convolution Neural Networks (CNNs) are used as part of a pipeline based on deep learning techniques. A divide and conquer approach is followed to analyse the effects on performance and efficiency when utilising diverse deep learning techniques such as varying network architectures (VGG19, ResNet50, InceptionV3, DenseNet121, MobileNetV2), class weights, input sizes, image ratios, pre-processing techniques, transfer learning, dropout rates, and types of mammogram projections. This approach serves as a starting point for model development of mammography classification tasks. Practitioners can benefit from this work by using the divide and conquer results to select the most suitable deep learning techniques for their case out-of-the-box, thus reducing the need for extensive exploratory experimentation. Multiple techniques are found to provide accuracy gains relative to a general baseline (VGG19 model using uncropped 512 × 512 pixels input images with a dropout rate of 0.2 and a learning rate of 1 × 10-3) on the Curated Breast Imaging Subset of DDSM (CBIS-DDSM) dataset. These techniques involve transfer learning pre-trained ImagetNet weights to a MobileNetV2 architecture, with pre-trained weights from a binarised version of the mini Mammography Image Analysis Society (mini-MIAS) dataset applied to the fully connected layers of the model, coupled with using weights to alleviate class imbalance, and splitting CBIS-DDSM samples between images of masses and calcifications. Using these techniques, a 5.6% gain in accuracy over the baseline model was accomplished. Other deep learning techniques from the divide and conquer approach, such as larger image sizes, do not yield increased accuracies without the use of image pre-processing techniques such as Gaussian filtering, histogram equalisation and input cropping.

Breast cancer detection from thermal images using a Grunwald-Letnikov-aided Dragonfly algorithm-based deep feature selection method

Breast cancer is one of the deadliest diseases in women and its incidence is growing at an alarming rate. However, early detection of this disease can be life-saving. The rapid development of deep learning techniques has generated a great deal of interest in the medical imaging field. Researchers around the world are working on developing breast cancer detection methods using medical imaging. In the present work, we have proposed a two-stage model for breast cancer detection using thermographic images. Firstly, features are extracted from images using a deep learning model, called VGG16. To select the optimal subset of features, we use a meta-heuristic algorithm called the Dragonfly Algorithm (DA) in the second step. To improve the performance of the DA, a memory-based version of DA is proposed using the Grunwald-Letnikov (GL) method. The proposed two-stage framework has been evaluated on a publicly available standard dataset called DMR-IR. The proposed model efficiently filters out non-essential features and had 100% diagnostic accuracy on the standard dataset, with 82% fewer features compared to the VGG16 model.

Adoption of New Technologies: Artificial Intelligence

Over the past decade, artificial intelligence (AI) has been broadly applied to many aspects of human life, with recent groundbreaking successes in facial recognition, natural language processing, autonomous driving, and medical imaging. Gastroenterology has applied AI to a vast array of clinical problems, and some of the earliest prospective trials examining AI in medicine have been in computer vision applied to endoscopy. Evidence is mounting for 2 broad areas of AI as applied to gastroenterology: computer-aided detection and computer-aided diagnosis.

An Automated In-Depth Feature Learning Algorithm for Breast Abnormality Prognosis and Robust Characterization from Mammography Images Using Deep Transfer Learning

BACKGROUND

Diagnosing breast cancer masses and calcification clusters have paramount significance in mammography, which aids in mitigating the disease's complexities and curing it at early stages. However, a wrong mammogram interpretation may lead to an unnecessary biopsy of the false-positive findings, which reduces the patient's survival chances. Consequently, approaches that learn to discern breast masses can reduce the number of misconceptions and incorrect diagnoses. Conventionally used classification models focus on feature extraction techniques specific to a particular problem based on domain information. Deep learning strategies are becoming promising alternatives to solve the many challenges of feature-based approaches.

METHODS

This study introduces a convolutional neural network (ConvNet)-based deep learning method to extract features at varying densities and discern mammography's normal and suspected regions. Two different experiments were carried out to make an accurate diagnosis and classification. The first experiment consisted of five end-to-end pre-trained and fine-tuned deep convolution neural networks (DCNN). The in-depth features extracted from the ConvNet are also used to train the support vector machine algorithm to achieve excellent performance in the second experiment. Additionally, DCNN is the most frequently used image interpretation and classification method, including VGGNet, GoogLeNet, MobileNet, ResNet, and DenseNet. Moreover, this study pertains to data cleaning, preprocessing, and data augmentation, and improving mass recognition accuracy. The efficacy of all models is evaluated by training and testing three mammography datasets and has exhibited remarkable results.

RESULTS

Our deep learning ConvNet+SVM model obtained a discriminative training accuracy of 97.7% and validating accuracy of 97.8%, contrary to this, VGGNet16 method yielded 90.2%, 93.5% for VGGNet19, 63.4% for GoogLeNet, 82.9% for MobileNetV2, 75.1% for ResNet50, and 72.9% for DenseNet121.

CONCLUSIONS

The proposed model's improvement and validation are appropriated in conventional pathological practices that conceivably reduce the pathologist's strain in predicting clinical outcomes by analyzing patients' mammography images.

Deep Learning Based Capsule Neural Network Model for Breast Cancer Diagnosis Using Mammogram Images

Breast cancer is a commonly occurring disease in women all over the world. Mammogram is an efficient technique used for screening and identification of abnormalities over the breast region. Earlier identification of breast cancer enhances the prognosis of patients and is mainly based on the experience of the radiologist in interpretation of mammogram with quality of image. The advent of Deep Learning (DL) and Computer Vision techniques is widely used to perform breast cancer diagnosis. This paper presents a new Optimal Multi-Level Thresholding-based Segmentation with DL enabled Capsule Network (OMLTS-DLCN) breast cancer diagnosis model utilizing digital mammograms. The OMLTS-DLCN model involves an Adaptive Fuzzy based median filtering (AFF) technique as a pre-processing step to eradicate the noise that exists in the mammogram images. Besides, Optimal Kapur's based Multilevel Thresholding with Shell Game Optimization (SGO) algorithm (OKMT-SGO) is applied for breast cancer segmentation. In addition, the proposed model involves a CapsNet based feature extractor and Back-Propagation Neural Network (BPNN) classification model is employed to detect the existence of breast cancer. The diagnostic outcomes of the presented OMLTS-DLCN technique is examined by means of benchmark Mini-MIAS dataset and DDSM dataset. The experimental values obtained highlights the superior performance of the OMLTS-DLCN model with a higher accuracy of 98.50 and 97.55% on the Mini-MIAS dataset and DDSM dataset, respectively.

Artificial Intelligence and the Medical Physicist: Welcome to the Machine

Artificial intelligence (AI) is a branch of computer science dedicated to giving machines or computers the ability to perform human-like cognitive functions, such as learning, problem-solving, and decision making. Since it is showing superior performance than well-trained human beings in many areas, such as image classification, object detection, speech recognition, and decision-making, AI is expected to change profoundly every area of science, including healthcare and the clinical application of physics to healthcare, referred to as medical physics. As a result, the Italian Association of Medical Physics (AIFM) has created the “AI for Medical Physics” (AI4MP) group with the aims of coordinating the efforts, facilitating the communication, and sharing of the knowledge on AI of the medical physicists (MPs) in Italy. The purpose of this review is to summarize the main applications of AI in medical physics, describe the skills of the MPs in research and clinical applications of AI, and define the major challenges of AI in healthcare.

Study Processes and Applications of Ultrasomics in Precision Medicine

Ultrasomics is the science of transforming digitally encrypted medical ultrasound images that hold information related to tumor pathophysiology into mineable high-dimensional data. Ultrasomics data have the potential to uncover disease characteristics that are not found with the naked eye. The task of ultrasomics is to quantify the state of diseases using distinctive imaging algorithms and thereby provide valuable information for personalized medicine. Ultrasomics is a powerful tool in oncology but can also be applied to other medical problems for which a disease is imaged. To date there is no comprehensive review focusing on ultrasomics. Here, we describe how ultrasomics works and its capability in diagnosing disease in different organs, including breast, liver, and thyroid. Its pitfalls, challenges and opportunities are also discussed.

Radiomics and deep learning in lung cancer

Lung malignancies have been extensively characterized through radiomics and deep learning. By providing a three-dimensional characterization of the lesion, models based on radiomic features from computed tomography (CT) and positron-emission tomography (PET) have been developed to detect nodules, distinguish malignant from benign lesions, characterize their histology, stage, and genotype. Deep learning models have been applied to automatically segment organs at risk in lung cancer radiotherapy, stratify patients according to the risk for local and distant recurrence, and identify patients candidate for molecular targeted therapy and immunotherapy. Moreover, radiomics has also been applied successfully to predict side effects such as radiation- and immunotherapy-induced pneumonitis and differentiate lung injury from recurrence. Radiomics could also untap the potential for further use of the cone beam CT acquired for treatment image guidance, four-dimensional CT, and dose-volume data from radiotherapy treatment plans. Radiomics is expected to increasingly affect the clinical practice of treatment of lung tumors, optimizing the end-to-end diagnosis-treatment-follow-up chain. The main goal of this article is to provide an update on the current status of lung cancer radiomics.

A machine learning approach on multiscale texture analysis for breast microcalcification diagnosis

Background

Screening programs use mammography as primary diagnostic tool for detecting breast cancer at an early stage. The diagnosis of some lesions, such as microcalcifications, is still difficult today for radiologists. In this paper, we proposed an automatic binary model for discriminating tissue in digital mammograms, as support tool for the radiologists. In particular, we compared the contribution of different methods on the feature selection process in terms of the learning performances and selected features.

Results

For each ROI, we extracted textural features on Haar wavelet decompositions and also interest points and corners detected by using Speeded Up Robust Feature (SURF) and Minimum Eigenvalue Algorithm (MinEigenAlg). Then a Random Forest binary classifier is trained on a subset of a sub-set features selected by two different kinds of feature selection techniques, such as filter and embedded methods. We tested the proposed model on 260 ROIs extracted from digital mammograms of the BCDR public database. The best prediction performance for the normal/abnormal and benign/malignant problems reaches a median AUC value of 98.16% and 92.08%, and an accuracy of 97.31% and 88.46%, respectively. The experimental result was comparable with related work performance.

Conclusions

The best performing result obtained with embedded method is more parsimonious than the filter one. The SURF and MinEigen algorithms provide a strong informative content useful for the characterization of microcalcification clusters.

Artificial Intelligence for Mammography and Digital Breast Tomosynthesis: Current Concepts and Future Perspectives

Although computer-aided diagnosis (CAD) is widely used in mammography, conventional CAD programs that use prompts to indicate potential cancers on the mammograms have not led to an improvement in diagnostic accuracy. Because of the advances in machine learning, especially with use of deep (multilayered) convolutional neural networks, artificial intelligence has undergone a transformation that has improved the quality of the predictions of the models. Recently, such deep learning algorithms have been applied to mammography and digital breast tomosynthesis (DBT). In this review, the authors explain how deep learning works in the context of mammography and DBT and define the important technical challenges. Subsequently, they discuss the current status and future perspectives of artificial intelligence-based clinical applications for mammography, DBT, and radiomics. Available algorithms are advanced and approach the performance of radiologists-especially for cancer detection and risk prediction at mammography. However, clinical validation is largely lacking, and it is not clear how the power of deep learning should be used to optimize practice. Further development of deep learning models is necessary for DBT, and this requires collection of larger databases. It is expected that deep learning will eventually have an important role in DBT, including the generation of synthetic images.

Multiscale connected chain topological modelling for microcalcification classification

Computer-aided diagnosis (CAD) systems can be employed to help classify mammographic microcalcification clusters. In this paper, a novel method for the classification of the microcalcification clusters based on topology/connectivity has been introduced. The proposed method is distinct from existing techniques which concentrate on morphology and texture of microcalcifications and surrounding tissue. The proposed approach used multiscale morphological relationship of connectivity between microcalcifications where connected chains between nearest microcalcifications were generated at each scale. Subsequently, graph connectivity features at each scale were extracted to estimate the topological connectivity structure of microcalcification clusters for benign versus malignant classification. The proposed approach was evaluated using publicly available digitized datasets: MIAS and DDSM, in addition to the digital OPTIMAM dataset. The classification of features using KNN obtained a classification accuracy of 86.47±1.30%, 90.0±0.00%, 82.5±2.63%, 76.75±0.66% for the DDSM, MIAS-manual, MIAS-auto and OPTIMAM datasets respectively. The study showed that topological/connectivity modelling using a multiscale approach was appropriate for microcalcification cluster analysis and classification; topological connectivity and distribution can be linked to clinical understanding of microcalcification spatial distribution.

Deep Learning to Improve Breast Cancer Detection on Screening Mammography

The rapid development of deep learning, a family of machine learning techniques, has spurred much interest in its application to medical imaging problems. Here, we develop a deep learning algorithm that can accurately detect breast cancer on screening mammograms using an "end-to-end" training approach that efficiently leverages training datasets with either complete clinical annotation or only the cancer status (label) of the whole image. In this approach, lesion annotations are required only in the initial training stage, and subsequent stages require only image-level labels, eliminating the reliance on rarely available lesion annotations. Our all convolutional network method for classifying screening mammograms attained excellent performance in comparison with previous methods. On an independent test set of digitized film mammograms from the Digital Database for Screening Mammography (CBIS-DDSM), the best single model achieved a per-image AUC of 0.88, and four-model averaging improved the AUC to 0.91 (sensitivity: 86.1%, specificity: 80.1%). On an independent test set of full-field digital mammography (FFDM) images from the INbreast database, the best single model achieved a per-image AUC of 0.95, and four-model averaging improved the AUC to 0.98 (sensitivity: 86.7%, specificity: 96.1%). We also demonstrate that a whole image classifier trained using our end-to-end approach on the CBIS-DDSM digitized film mammograms can be transferred to INbreast FFDM images using only a subset of the INbreast data for fine-tuning and without further reliance on the availability of lesion annotations. These findings show that automatic deep learning methods can be readily trained to attain high accuracy on heterogeneous mammography platforms, and hold tremendous promise for improving clinical tools to reduce false positive and false negative screening mammography results. Code and model available at: https://github.com/lishen/end2end-all-conv .

Automatic segmentation and classification of breast lesions through identification of informative multiparametric PET/MRI features

Background

Multiparametric positron emission tomography/magnetic resonance imaging (mpPET/MRI) shows clinical potential for detection and classification of breast lesions. Yet, the contribution of features for computer-aided segmentation and diagnosis (CAD) need to be better understood. We proposed a data-driven machine learning approach for a CAD system combining dynamic contrast-enhanced (DCE)-MRI, diffusion-weighted imaging (DWI), and ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG)-PET.

Methods

The CAD incorporated a random forest (RF) classifier combined with mpPET/MRI intensity-based features for lesion segmentation and shape features, kinetic and spatio-temporal texture features, for lesion classification. The CAD pipeline detected and segmented suspicious regions and classified lesions as benign or malignant. The inherent feature selection method of RF and alternatively the minimum-redundancy-maximum-relevance feature ranking method were used.

Results

In 34 patients, we report a detection rate of 10/12 (83.3%) and 22/22 (100%) for benign and malignant lesions, respectively, a Dice similarity coefficient of 0.665 for segmentation, and a classification performance with an area under the curve at receiver operating characteristics analysis of 0.978, a sensitivity of 0.946, and a specificity of 0.936. Segmentation but not classification performance of DCE-MRI improved with information from DWI and FDG-PET. Feature ranking revealed that kinetic and spatio-temporal texture features had the highest contribution for lesion classification. ¹⁸F-FDG-PET and morphologic features were less predictive.

Conclusion

Our CAD enables the assessment of the relevance of mpPET/MRI features on segmentation and classification accuracy. It may aid as a novel computational tool for exploring different modalities/features and their contributions for the detection and classification of breast lesions.

Electronic supplementary material

The online version of this article (10.1186/s41747-019-0096-3) contains supplementary material, which is available to authorized users.

Detection of potential microcalcification clusters using multivendor for-presentation digital mammograms for short-term breast cancer risk estimation

PURPOSE

We explore using the number of potential microcalcification clusters detected in for-presentation mammographic images (the images which are typically accessible to large epidemiological studies) a marker of short-term breast cancer risk.

METHODS

We designed a three-step algorithm for detecting potential microcalcification clusters in for-presentation digital mammograms. We studied association with short-term breast cancer risk using a nested case control design, with a mammography screening cohort as a source population. In total, 373 incident breast cancer cases (diagnosed at least 3 months after a negative screen at study entry) and 1466 matched controls were included in our study. Conditional logistic regression Wald tests were used to test for association with the presence of microcalcifications at study entry. We compared results of these analyses to those obtained using a Computer-aided Diagnosis (CAD) software (VuComp) on corresponding for-processing images (images which are used clinically, but typically not saved).

RESULTS

We found a moderate agreement between our measure of potential microcalcification clusters on for-presentation images and a CAD measure on for-processing images. Similar evidence of association with short-term breast cancer risk was found (P =  1 × 10 - 10 and P =  9 × 10 - 09 , for our approach on for-presentation images and for the CAD measure on for-processing images, respectively) and interestingly both measures contributed independently to association with a short-term risk (P =  9 × 10 - 03 for the CAD measure, adjusted for our proposed method and P =  1 × 10 - 04 for our proposed method, adjusted for the CAD measure).

CONCLUSION

Meaningful measurement of potential microcalcifications, in the context of short-term breast cancer risk assessment, is feasible for for-presentation images across a range of vendors. Our algorithm for for-presentation images performs similarly to a CAD algorithm on for-processing images, hence our algorithm can be a useful tool for research on microcalcifications and their role on breast cancer risk, based on large-scale epidemiological studies with access to for-presentation images.

Breast Cancer Diagnosis in Digital Breast Tomosynthesis: Effects of Training Sample Size on Multi-Stage Transfer Learning Using Deep Neural Nets

In this paper, we developed a deep convolutional neural network (CNN) for the classification of malignant and benign masses in digital breast tomosynthesis (DBT) using a multi-stage transfer learning approach that utilized data from similar auxiliary domains for intermediate-stage fine-tuning. Breast imaging data from DBT, digitized screen-film mammography, and digital mammography totaling 4039 unique regions of interest (1797 malignant and 2242 benign) were collected. Using cross validation, we selected the best transfer network from six transfer networks by varying the level up to which the convolutional layers were frozen. In a single-stage transfer learning approach, knowledge from CNN trained on the ImageNet data was fine-tuned directly with the DBT data. In a multi-stage transfer learning approach, knowledge learned from ImageNet was first fine-tuned with the mammography data and then fine-tuned with the DBT data. Two transfer networks were compared for the second-stage transfer learning by freezing most of the CNN structures versus freezing only the first convolutional layer. We studied the dependence of the classification performance on training sample size for various transfer learning and fine-tuning schemes by varying the training data from 1% to 100% of the available sets. The area under the receiver operating characteristic curve (AUC) was used as a performance measure. The view-based AUC on the test set for single-stage transfer learning was 0.85 ± 0.05 and improved significantly (p <; 0.05$ ) to 0.91 ± 0.03 for multi-stage learning. This paper demonstrated that, when the training sample size from the target domain is limited, an additional stage of transfer learning using data from a similar auxiliary domain is advantageous.

Microcalcification Segmentation from Mammograms: A Morphological Approach

This publication presents a computer method for segmenting microcalcifications in mammograms. It makes use of morphological transformations and is composed of two parts. The first part detects microcalcifications morphologically, thus allowing the approximate area of their occurrence to be determined, the contrast to be improved, and noise to be reduced in the mammograms. In the second part, a watershed segmentation of microcalcifications is carried out. This study was carried out on a test set containing 200 ROIs 512 × 512 pixels in size, taken from mammograms from the Digital Database for Screening Mammography (DDSM), including 100 cases showing malignant lesions and 100 cases showing benign ones. The experiments carried out yielded the following average values of the measured indices: 80.5% (similarity index), 75.7% (overlap fraction), 70.8% (overlap value), and 19.8% (extra fraction). The average time of executing all steps of the methods used for a single ROI amounted to 0.83 s.

Locally adaptive decision in detection of clustered microcalcifications in mammograms

In computer-aided detection or diagnosis of clustered microcalcifications (MCs) in mammograms, the performance often suffers from not only the presence of false positives (FPs) among the detected individual MCs but also large variability in detection accuracy among different cases. To address this issue, we investigate a locally adaptive decision scheme in MC detection by exploiting the noise characteristics in a lesion area. Instead of developing a new MC detector, we propose a decision scheme on how to best decide whether a detected object is an MC or not in the detector output. We formulate the individual MCs as statistical outliers compared to the many noisy detections in a lesion area so as to account for the local image characteristics. To identify the MCs, we first consider a parametric method for outlier detection, the Mahalanobis distance detector, which is based on a multi-dimensional Gaussian distribution on the noisy detections. We also consider a non-parametric method which is based on a stochastic neighbor graph model of the detected objects. We demonstrated the proposed decision approach with two existing MC detectors on a set of 188 full-field digital mammograms (95 cases). The results, evaluated using free response operating characteristic (FROC) analysis, showed a significant improvement in detection accuracy by the proposed outlier decision approach over traditional thresholding (the partial area under the FROC curve increased from 3.95 to 4.25, p-value  <10^-4). There was also a reduction in case-to-case variability in detected FPs at a given sensitivity level. The proposed adaptive decision approach could not only reduce the number of FPs in detected MCs but also improve case-to-case consistency in detection.

Fully automated determination of arch angle on weight-bearing foot radiograph

BACKGROUND AND OBJECTIVE

Flatfeet can be evaluated by measuring the calcaneal-fifth metatarsal angle on a weight-bearing lateral foot radiograph. This study aimed to develop an automated method for determining the calcaneal-fifth metatarsal angle on weight-bearing lateral foot radiograph.

METHOD

The proposed method comprises four processing steps: (1) identification of the regions including the calcaneus and fifth metatarsal bones in a foot image; (2) delineation of the contours of the calcaneus and the fifth metatarsal; (3) determination of the tangential lines of the two bones from the contours; and (4) determination of the calcaneal-fifth metatarsal angle between the two tangential lines as arch angle.

RESULTS

The proposed method was evaluated using 300 weight-bearing lateral foot radiographs. The arch angles determined by the proposed method were compared with those measured by a radiologist, and the errors between the automatically and manually determined angles were used to evaluate the precision of the method. The average error in the proposed method was found to be 1.12° ± 1.57° In the study, in 73.33% of the cases, the arch angles could be determined automatically without redrawing any tangential lines; in 23.00% of the cases, the angles would be correctly determined by redrawing one of the tangential lines; further, in only 3.67% of the cases, both the calcaneal and fifth metatarsal tangential lines needed to be redrawn to determine the arch angles.

CONCLUSION

The results revealed that the proposed method has potential for assisting doctors in measuring the arch angles on weight-bearing lateral foot radiographs more efficiently.

Classification of micro-calcification in mammograms using scalable linear Fisher discriminant analysis

Breast cancer is one of the major causes of death in women. Computer Aided Diagnosis (CAD) systems are being developed to assist radiologists in early diagnosis. Micro-calcifications can be an early symptom of breast cancer. Besides detection, classification of micro-calcification as benign or malignant is essential in a complete CAD system. We have developed a novel method for the classification of benign and malignant micro-calcification using an improved Fisher Linear Discriminant Analysis (LDA) approach for the linear transformation of segmented micro-calcification data in combination with a Support Vector Machine (SVM) variant to classify between the two classes. The results indicate an average accuracy equal to 96% which is comparable to state-of-the art methods in the literature.

Graphical Abstract

Beyond imaging: The promise of radiomics

The domain of investigation of radiomics consists of large-scale radiological image analysis and association with biological or clinical endpoints. The purpose of the present study is to provide a recent update on the status of this rapidly emerging field by performing a systematic review of the literature on radiomics, with a primary focus on oncologic applications. The systematic literature search, performed in Pubmed using the keywords: "radiomics OR radiomic" provided 97 research papers. Based on the results of this search, we describe the methods used for building a model of prognostic value from quantitative analysis of patient images. Then, we provide an up-to-date overview of the results achieved in this field, and discuss the current challenges and future developments of radiomics for oncology.

Optimal reconstruction and quantitative image features for computer-aided diagnosis tools for breast CT

PURPOSE

The purpose of this study is to determine the optimal representative reconstruction and quantitative image feature set for a computer-aided diagnosis (CADx) scheme for dedicated breast computer tomography (bCT).

METHOD

We used 93 bCT scans that contain 102 breast lesions (62 malignant, 40 benign). Using an iterative image reconstruction (IIR) algorithm, we created 37 reconstructions with different image appearances for each case. In addition, we added a clinical reconstruction for comparison purposes. We used image sharpness, determined by the gradient of gray value in a parenchymal portion of the reconstructed breast, as a surrogate measure of the image qualities/appearances for the 38 reconstructions. After segmentation of the breast lesion, we extracted 23 quantitative image features. Using leave-one-out-cross-validation (LOOCV), we conducted the feature selection, classifier training, and testing. For this study, we used the linear discriminant analysis classifier. Then, we selected the representative reconstruction and feature set for the classifier with the best diagnostic performance among all reconstructions and feature sets. Then, we conducted an observer study with six radiologists using a subset of breast lesions (N = 50). Using 1000 bootstrap samples, we compared the diagnostic performance of the trained classifier to those of the radiologists.

RESULT

The diagnostic performance of the trained classifier increased as the image sharpness of a given reconstruction increased. Among combinations of reconstructions and quantitative image feature sets, we selected one of the sharp reconstructions and three quantitative image feature sets with the first three highest diagnostic performances under LOOCV as the representative reconstruction and feature set for the classifier. The classifier on the representative reconstruction and feature set achieved better diagnostic performance with an area under the ROC curve (AUC) of 0.94 (95% CI = [0.81, 0.98]) than those of the radiologists, where their maximum AUC was 0.78 (95% CI = [0.63, 0.90]). Moreover, the partial AUC, at 90% sensitivity or higher, of the classifier (pAUC = 0.085 with 95% CI = [0.063, 0.094]) was statistically better (P-value < 0.0001) than those of the radiologists (maximum pAUC = 0.009 with 95% CI = [0.003, 0.024]).

CONCLUSION

We found that image sharpness measure can be a good candidate to estimate the diagnostic performance of a given CADx algorithm. In addition, we found that there exists a reconstruction (i.e., sharp reconstruction) and a feature set that maximizes the diagnostic performance of a CADx algorithm. On this optimal representative reconstruction and feature set, the CADx algorithm outperformed radiologists.

Reducing false positives of microcalcification detection systems by removal of breast arterial calcifications

PURPOSE

In the past decades, computer-aided detection (CADe) systems have been developed to aid screening radiologists in the detection of malignant microcalcifications. These systems are useful to avoid perceptual oversights and can increase the radiologists' detection rate. However, due to the high number of false positives marked by these CADe systems, they are not yet suitable as an independent reader. Breast arterial calcifications (BACs) are one of the most frequent false positives marked by CADe systems. In this study, a method is proposed for the elimination of BACs as positive findings. Removal of these false positives will increase the performance of the CADe system in finding malignant microcalcifications.

METHODS

A multistage method is proposed for the removal of BAC findings. The first stage consists of a microcalcification candidate selection, segmentation and grouping of the microcalcifications, and classification to remove obvious false positives. In the second stage, a case-based selection is applied where cases are selected which contain BACs. In the final stage, BACs are removed from the selected cases. The BACs removal stage consists of a GentleBoost classifier trained on microcalcification features describing their shape, topology, and texture. Additionally, novel features are introduced to discriminate BACs from other positive findings.

RESULTS

The CADe system was evaluated with and without BACs removal. Here, both systems were applied on a validation set containing 1088 cases of which 95 cases contained malignant microcalcifications. After bootstrapping, free-response receiver operating characteristics and receiver operating characteristics analyses were carried out. Performance between the two systems was compared at 0.98 and 0.95 specificity. At a specificity of 0.98, the sensitivity increased from 37% to 52% and the sensitivity increased from 62% up to 76% at a specificity of 0.95. Partial areas under the curve in the specificity range of 0.8-1.0 were significantly different between the system without BACs removal and the system with BACs removal, 0.129 ± 0.009 versus 0.144 ± 0.008 (p<0.05), respectively. Additionally, the sensitivity at one false positive per 50 cases and one false positive per 25 cases increased as well, 37% versus 51% (p<0.05) and 58% versus 67% (p<0.05) sensitivity, respectively. Additionally, the CADe system with BACs removal reduces the number of false positives per case by 29% on average. The same sensitivity at one false positive per 50 cases in the CADe system without BACs removal can be achieved at one false positive per 80 cases in the CADe system with BACs removal.

CONCLUSIONS

By using dedicated algorithms to detect and remove breast arterial calcifications, the performance of CADe systems can be improved, in particular, at false positive rates representative for operating points used in screening.

Analysis of computer-aided detection techniques and signal characteristics for clustered microcalcifications on digital mammography and digital breast tomosynthesis

With IRB approval, digital breast tomosynthesis (DBT) images of human subjects were collected using a GE GEN2 DBT prototype system. Corresponding digital mammograms (DMs) of the same subjects were collected retrospectively from patient files. The data set contained a total of 237 views of DBT and equal number of DM views from 120 human subjects, each included 163 views with microcalcification clusters (MCs) and 74 views without MCs. The data set was separated into training and independent test sets. The pre-processing, object prescreening and segmentation, false positive reduction and clustering strategies for MC detection by three computer-aided detection (CADe) systems designed for DM, DBT, and a planar projection image generated from DBT were analyzed. Receiver operating characteristic (ROC) curves based on features extracted from microcalcifications and free-response ROC (FROC) curves based on scores from MCs were used to quantify the performance of the systems. Jackknife FROC (JAFROC) and non-parametric analysis methods were used to determine the statistical difference between the FROC curves. The difference between the CAD_DM and CAD_DBT systems when the false positive rate was estimated from cases without MCs did not reach statistical significance. The study indicates that the large search space in DBT may not be a limiting factor for CADe to achieve similar performance as that observed in DM.

Large scale deep learning for computer aided detection of mammographic lesions

Recent advances in machine learning yielded new techniques to train deep neural networks, which resulted in highly successful applications in many pattern recognition tasks such as object detection and speech recognition. In this paper we provide a head-to-head comparison between a state-of-the art in mammography CAD system, relying on a manually designed feature set and a Convolutional Neural Network (CNN), aiming for a system that can ultimately read mammograms independently. Both systems are trained on a large data set of around 45,000 images and results show the CNN outperforms the traditional CAD system at low sensitivity and performs comparable at high sensitivity. We subsequently investigate to what extent features such as location and patient information and commonly used manual features can still complement the network and see improvements at high specificity over the CNN especially with location and context features, which contain information not available to the CNN. Additionally, a reader study was performed, where the network was compared to certified screening radiologists on a patch level and we found no significant difference between the network and the readers.

Development of intelligent systems based on Bayesian regularization network and neuro-fuzzy models for mass detection in mammograms: A comparative analysis

Female breast cancer is the second most common cancer in the world. Several efforts in artificial intelligence have been made to help improving the diagnostic accuracy at earlier stages. However, the identification of breast abnormalities, like masses, on mammographic images is not a trivial task, especially for dense breasts. In this paper we describe our novel mass detection process that includes three successive steps of enhancement, characterization and classification. The proposed enhancement system is based mainly on the analysis of the breast texture. First of all, a filtering step with morphological operators and soft thresholding is achieved. Then, we remove from the filtered breast region, all the details that may interfere with the eventual masses, including pectoral muscle and galactophorous tree. The pixels belonging to this tree will be interpolated and replaced by the average of the neighborhood. In the characterization process, measurement of the Gaussian density in the wavelet domain allows the segmentation of the masses. Finally, a comparative classification mechanism based on the Bayesian regularization back-propagation networks and ANFIS techniques is proposed. The tests were conducted on the MIAS database. The results showed the robustness of the proposed enhancement method.

An Improved CAD System for Breast Cancer Diagnosis Based on Generalized Pseudo-Zernike Moment and Ada-DEWNN Classifier

In this paper, a novel framework of computer-aided diagnosis (CAD) system has been presented for the classification of benign/malignant breast tissues. The properties of the generalized pseudo-Zernike moments (GPZM) and pseudo-Zernike moments (PZM) are utilized as suitable texture descriptors of the suspicious region in the mammogram. An improved classifier- adaptive differential evolution wavelet neural network (Ada-DEWNN) is proposed to improve the classification accuracy of the CAD system. The efficiency of the proposed system is tested on mammograms from the Mammographic Image Analysis Society (mini-MIAS) database using the leave-one-out cross validation as well as on mammograms from the Digital Database for Screening Mammography (DDSM) database using 10-fold cross validation. The proposed method on MIAS-database attains a fair accuracy of 0.8938 and AUC of 0.935 (95 % CI = 0.8213-0.9831). The proposed method is also tested for in-plane rotation and found to be highly rotation invariant. In addition, the proposed classifier is tested and compared with some well-known existing methods using receiver operating characteristic (ROC) analysis using DDSM- database. It is concluded the proposed classifier has better area under the curve (AUC) (0.9289) and highly précised with 95 % CI, 0.8216 to 0.9834 and 0.0384 standard error.

Multi-View Probabilistic Classification of Breast Microcalcifications

Classification of clustered breast microcalcifications into benign and malignant categories is an extremely challenging task for computerized algorithms and expert radiologists alike. In this paper we apply a multi-view-classifier for the task. We describe a two-step classification method that is based on a view-level decision, implemented by a logistic regression classifier, followed by a stochastic combination of the two view-level indications into a single benign or malignant decision. The proposed method was evaluated on a large number of cases from a standardized digital database for screening mammography (DDSM). Experimental results demonstrate the advantage of the proposed multi-view classification algorithm that automatically learns the best way to combine the views.

Topological modeling and classification of mammographic microcalcification clusters

GOAL

The presence of microcalcification clusters is a primary sign of breast cancer; however, it is difficult and time consuming for radiologists to classify microcalcifications as malignant or benign. In this paper, a novel method for the classification of microcalcification clusters in mammograms is proposed.

METHODS

The topology/connectivity of individual microcalcifications is analyzed within a cluster using multiscale morphology. This is distinct from existing approaches that tend to concentrate on the morphology of individual microcalcifications and/or global (statistical) cluster features. A set of microcalcification graphs are generated to represent the topological structure of microcalcification clusters at different scales. Subsequently, graph theoretical features are extracted, which constitute the topological feature space for modeling and classifying microcalcification clusters. k-nearest-neighbors-based classifiers are employed for classifying microcalcification clusters.

RESULTS

The validity of the proposed method is evaluated using two well-known digitized datasets (MIAS and DDSM) and a full-field digital dataset. High classification accuracies (up to 96%) and good ROC results (area under the ROC curve up to 0.96) are achieved. A full comparison with related publications is provided, which includes a direct comparison.

CONCLUSION

The results indicate that the proposed approach is able to outperform the current state-of-the-art methods. Significance: This study shows that topology modeling is an important tool for microcalcification analysis not only because of the improved classification accuracy but also because the topological measures can be linked to clinical understanding.

Evaluating geodesic active contours in microcalcifications segmentation on mammograms

Breast cancer is the most commonly occurring type of cancer among women, and it is the major cause of female cancer-related deaths worldwide. Its incidence is increasing in developed as well as developing countries. Efficient strategies to reduce the high death rates due to breast cancer include early detection and tumor removal in the initial stages of the disease. Clinical and mammographic examinations are considered the best methods for detecting the early signs of breast cancer; however, these techniques are highly dependent on breast characteristics, equipment quality, and physician experience. Computer-aided diagnosis (CADx) systems have been developed to improve the accuracy of mammographic diagnosis; usually such systems may involve three steps: (i) segmentation; (ii) parameter extraction and selection of the segmented lesions and (iii) lesions classification. Literature considers the first step as the most important of them, as it has a direct impact on the lesions characteristics that will be used in the further steps. In this study, the original contribution is a microcalcification segmentation method based on the geodesic active contours (GAC) technique associated with anisotropic texture filtering as well as the radiologists' knowledge. Radiologists actively participate on the final step of the method, selecting the final segmentation that allows elaborating an adequate diagnosis hypothesis with the segmented microcalcifications presented in a region of interest (ROI). The proposed method was assessed by employing 1000 ROIs extracted from images of the Digital Database for Screening Mammography (DDSM). For the selected ROIs, the rate of adequately segmented microcalcifications to establish a diagnosis hypothesis was at least 86.9%, according to the radiologists. The quantitative test, based on the area overlap measure (AOM), yielded a mean of 0.52±0.20 for the segmented images, when all 2136 segmented microcalcifications were considered. Moreover, a statistical difference was observed between the AOM values for large and small microcalcifications. The proposed method had better or similar performance as compared to literature for microcalcifications with maximum diameters larger than 460μm. For smaller microcalcifications the performance was limited.

External validation of a publicly available computer assisted diagnostic tool for mammographic mass lesions with two high prevalence research datasets

PURPOSE

Lesions detected at mammography are described with a highly standardized terminology: the breast imaging-reporting and data system (BI-RADS) lexicon. Up to now, no validated semantic computer assisted classification algorithm exists to interactively link combinations of morphological descriptors from the lexicon to a probabilistic risk estimate of malignancy. The authors therefore aim at the external validation of the mammographic mass diagnosis (MMassDx) algorithm. A classification algorithm like MMassDx must perform well in a variety of clinical circumstances and in datasets that were not used to generate the algorithm in order to ultimately become accepted in clinical routine.

METHODS

The MMassDx algorithm uses a naïve Bayes network and calculates post-test probabilities of malignancy based on two distinct sets of variables, (a) BI-RADS descriptors and age ("descriptor model") and (b) BI-RADS descriptors, age, and BI-RADS assessment categories ("inclusive model"). The authors evaluate both the MMassDx (descriptor) and MMassDx (inclusive) models using two large publicly available datasets of mammographic mass lesions: the digital database for screening mammography (DDSM) dataset, which contains two subsets from the same examinations-a medio-lateral oblique (MLO) view and cranio-caudal (CC) view dataset-and the mammographic mass (MM) dataset. The DDSM contains 1220 mass lesions and the MM dataset contains 961 mass lesions. The authors evaluate discriminative performance using area under the receiver-operating-characteristic curve (AUC) and compare this to the BI-RADS assessment categories alone (i.e., the clinical performance) using the DeLong method. The authors also evaluate whether assigned probabilistic risk estimates reflect the lesions' true risk of malignancy using calibration curves.

RESULTS

The authors demonstrate that the MMassDx algorithms show good discriminatory performance. AUC for the MMassDx (descriptor) model in the DDSM data is 0.876/0.895 (MLO/CC view) and AUC for the MMassDx (inclusive) model in the DDSM data is 0.891/0.900 (MLO/CC view). AUC for the MMassDx (descriptor) model in the MM data is 0.862 and AUC for the MMassDx (inclusive) model in the MM data is 0.900. In all scenarios, MMassDx performs significantly better than clinical performance, P < 0.05 each. The authors furthermore demonstrate that the MMassDx algorithm systematically underestimates the risk of malignancy in the DDSM and MM datasets, especially when low probabilities of malignancy are assigned.

CONCLUSIONS

The authors' results reveal that the MMassDx algorithms have good discriminatory performance but less accurate calibration when tested on two independent validation datasets. Improvement in calibration and testing in a prospective clinical population will be important steps in the pursuit of translation of these algorithms to the clinic.

Development of an online, publicly accessible naive Bayesian decision support tool for mammographic mass lesions based on the American College of Radiology (ACR) BI-RADS lexicon

PURPOSE

To develop and validate a decision support tool for mammographic mass lesions based on a standardized descriptor terminology (BI-RADS lexicon) to reduce variability of practice.

MATERIALS AND METHODS

We used separate training data (1,276 lesions, 138 malignant) and validation data (1,177 lesions, 175 malignant). We created naïve Bayes (NB) classifiers from the training data with tenfold cross-validation. Our "inclusive model" comprised BI-RADS categories, BI-RADS descriptors, and age as predictive variables; our "descriptor model" comprised BI-RADS descriptors and age. The resulting NB classifiers were applied to the validation data. We evaluated and compared classifier performance with ROC-analysis.

RESULTS

In the training data, the inclusive model yields an AUC of 0.959; the descriptor model yields an AUC of 0.910 (P < 0.001). The inclusive model is superior to the clinical performance (BI-RADS categories alone, P < 0.001); the descriptor model performs similarly. When applied to the validation data, the inclusive model yields an AUC of 0.935; the descriptor model yields an AUC of 0.876 (P < 0.001). Again, the inclusive model is superior to the clinical performance (P < 0.001); the descriptor model performs similarly.

CONCLUSION

We consider our classifier a step towards a more uniform interpretation of combinations of BI-RADS descriptors. We provide our classifier at www.ebm-radiology.com/nbmm/index.html .

KEY POINTS

• We provide a decision support tool for mammographic masses at www.ebm-radiology.com/nbmm/index.html . • Our tool may reduce variability of practice in BI-RADS category assignment. • A formal analysis of BI-RADS descriptors may enhance radiologists' diagnostic performance.

A spatial shape constrained clustering method for mammographic mass segmentation

A novel clustering method is proposed for mammographic mass segmentation on extracted regions of interest (ROIs) by using deterministic annealing incorporating circular shape function (DACF). The objective function reported in this study uses both intensity and spatial shape information, and the dominant dissimilarity measure is controlled by two weighting parameters. As a result, pixels having similar intensity information but located in different regions can be differentiated. Experimental results shows that, by using DACF, the mass segmentation results in digitized mammograms are improved with optimal mass boundaries, less number of noisy patches, and computational efficiency. An average probability of segmentation error of 7.18% for well-defined masses (or 8.06% for ill-defined masses) was obtained by using DACF on MiniMIAS database, with 5.86% (or 5.55%) and 6.14% (or 5.27%) improvements as compared to the standard DA and fuzzy c-means methods.