Breast Cancer Diagnosis With Mammography

Breast cancer, which is the second-most common and leading cause of cancer death among women, has witnessed growing interest in the two last decades. Fortunately, its early detection is the most effective way to detect and diagnose breast cancer. Although mammography is the gold standard for screening, its difficult interpretation leads to an increase in missed cancers and misinterpreted non-cancerous lesion rates. Therefore, computer-aided diagnosis (CAD) systems can be a great helpful tool for assisting radiologists in mammogram interpretation. Nonetheless, these systems are limited by their black-box outputs, which decreases the radiologists' confidence. To circumvent this limit, content-based mammogram retrieval (CBMR) is used as an alternative to traditional CAD systems. Herein, authors systematically review the state-of-the-art on mammography-based breast cancer CAD methods, while focusing on recent advances in CBMR methods. In order to have a complete review, mammography imaging principles and its correlation with breast anatomy are also discussed.

Using Extreme Learning Machines and the Backprojection Algorithm as an Alternative to Reconstruct Electrical Impedance Tomography Images

Electrical Impedance Tomography (EIT) is an imaging technique based on the excitation of electrode pairs applied to the surface of the imaged region. The electrical potentials generated from alternating current excitation are measured and then applied to boundary-based reconstruction methods. When compared to other imaging techniques, EIT is considered a low-cost technique without ionizing radiation emission, safer for patients. However, the resolution is still low, depending on efficient reconstruction methods and low computational cost. EIT has the potential to be used as an alternative test for early detection of breast lesions in general. The most accurate reconstruction methods tend to be very costly as they use optimization methods as a support. Backprojection tends to be rapid but more inaccurate. In this work, the authors propose a hybrid method, based on extreme learning machines and backprojection for EIT reconstruction. The results were applied to numerical phantoms and were considered adequate, with potential to be improved using post processing techniques.

Developing and Using Computational Frameworks to Conduct Numerical Analysis and Calculate Temperature Profiles and to Classify Breast Abnormalities

In this chapter, computational tools that have been designed to analyze and classify infrared (IR) images will be presented. The function of such tools is to interconnect in a user-friendly way the algorithms that are used to map temperatures and to classify some breast pathologies. One of these performs texture mapping using IR breast images to relate temperatures measured to the points over the substitute tridimensional geometry mesh. Another computer-aided diagnosis (CAD) tool was adapted so that it could be used to evaluate individual patients. This methodology will be used when the computational framework approach for classification is described. Finally, graphical interfaces and their functionalities will be presented and explained. Some case studies will be presented in order to verify whether or not the computational classification framework is effective.

The Evolution of New Trends in Breast Thermography

In this chapter, the theoretical foundations of infrared radiation theory and the principles of the infrared imaging technique are presented. The use of infrared (IR) images has increased recently, especially due to the refinement and portability of thermographic cameras. As a result, this type of camera can be used for various medical applications. In this context, the use of IR images is proposed as an auxiliary tool for detecting disease and monitoring, especially for the early detection of breast cancer.

Classification of Breast Lesions in Frontal Thermographic Images Using a Diagnosis Aid Intelligent System

This study aims to assess the breast lesions classification in thermographic images using different configuration of an Extreme Learning Machine network as classifier. In this approach, the authors changed the number of neurons in the hidden layer and the type of kernel function to further explore the network in order to find a better solution for the classification problem. Authors also used different tools to perform features extraction to assess both texture and geometry information from the breast lesions. During the study, the authors found that the results changed not only due to the network parameters but also due to the features chosen to represent the thermographic images. A maximum accuracy of 95% was found for the differentiation of breast lesions.

Applications of the Use of Infrared Breast Images

In this chapter, procedures for and applications of using infrared (IR) imaging that have been developed will be presented and proposed means by which a better detailed understanding of breast cancer can be reached. It will be shown how such applications can be used as a basis for enhancing the use of breast thermographic imaging as a user-friendly and inexpensive tool for the early detection of breast cancer. The authors intend to show that IR imaging can also be used to validate temperature profiles that have been calculated and to classify breast abnormalities as set out in previous chapters. IR images can also be used to estimate thermophysical properties of the breast, and discussion of how this is done is included. The IR images were acquired at the Outpatients Clinic of Mastology of the Hospital das Clínicas of the Federal University of Pernambuco (HC/UFPE). The research project was registered in the Brazilian Health Ministry (CEP/CCS/UFPE nº 279/05) after being approved by the Ethics Committee of UFPE.

Imaging Techniques for Breast Cancer Diagnosis

Breast cancer is the most typical variety of cancer in women worldwide. Mammography is the “gold standard” for the analysis of the breast from an imaging perspective. Altogether, the techniques used within the management of cancer in all stages are multiple biomedical imaging. Imaging as a very important part of cancer clinical protocols can offer a range of knowledge regarding morphology, structure, metabolism, and functions. Supported by relevant literature, this text provides an outline of the previous and new modalities employed in the sector of breast imaging. Any progress in technology can result in increased imaging speed to satisfy physiological processes necessities. One of the problems within the designation of breast cancer is sensitivity limitation. To overcome this limitation, complementary imaging examinations are used that historically include screening, ultrasound, MRI, etc.

Computer Techniques for Detection of Breast Cancer and Follow Up Neoadjuvant Treatment

This chapter explores several steps of the thermal breast exams analysis process in detecting breast abnormality and evaluating the response of pre-surgical treatment. Topics concerning the process of acquiring, storing, and preprocessing these exams, including a novel segmentation proposal that uses collective intelligence techniques, will be discussed. In addition, various approaches to calculating statistical and geometric descriptors from thermal breast examinations are also considered of this chapter. These descriptors can be used at different stages of the analysis process of these exams. In this sense, two experiments will be presented. The first one explores the use of genetic algorithms in the feature selection process. The second conducts a preliminary study that intends to analyze some descriptors, already used in other works, in the process of evaluating preoperative treatment response. This evaluation is of fundamental importance since the response is directly associated with the prognosis of the disease.

Prototype of a Low-Cost Impedance Tomography Based on the Open-Hardware Paradigm

Electrical Impedance Tomography (EIT) is a noninvasive, painless, and ionizing radiation-free technique for image acquisition of a region of interest. It is performed through electrical parameters. The method is based on the application of an alternating electric current pattern of low intensity through electrodes arranged around the surface region in order to obtain the image and also to measure the excitation electrical potentials. The aim of this study is to develop a device based in open hardware. Furthermore, the authors aim to build a prototype of a data acquisition system based on EIT. This device is the first step to obtain a complete and portable tomography equipment at a low cost.

The Efforts of Deep Learning Approaches for Breast Cancer Detection Based on X-Ray Images

In this chapter, deep learning-based approaches, namely deep feature extraction, fine-tuning of pre-trained convolutional neural networks (CNN), and end-to-end training of a developed CNN model, are used to classify the malignant and normal breast X-ray images. For deep feature extraction, pre-trained deep CNN models such as ResNet18, ResNet50, ResNet101, VGG16, and VGG19 are used. For classification of the deep features, the support vector machines (SVM) classifier is used with various kernel functions namely linear, quadratic, cubic, and Gaussian, respectively. The aforementioned pre-trained deep CNN models are also used in fine-tuning procedure. A new CNN model is also proposed in end-to-end training fashion. The classification accuracy is used as performance measurements. The experimental works show that the deep learning has potential in detection of the breast cancer from the X-ray images. The deep features that are extracted from the ResNet50 model and SVM classifier with linear kernel function produced 94.7% accuracy score which the highest among all obtained.