Computerized Analysis Of Mammographic Images For Detection And Characterization Of Breast Cancer

The identification and interpretation of the signs of breast cancer in mammographic images from screening programs can be very difficult due to the subtle and diversified appearance of breast disease. This book presents new image processing and pattern recognition techniques for computer-aided detection and diagnosis of breast cancer in its various forms. The main goals are: (1) the identification of bilateral asymmetry as an early sign of breast disease which is not detectable by other existing approaches; and (2) the detection and classification of masses and regions of architectural distortion, as benign lesions or malignant tumors, in a unified framework that does not require accurate extraction of the contours of the lesions. The innovative aspects of the work include the design and validation of landmarking algorithms, automatic Tabár masking procedures, and various feature descriptors for quantification of similarity and for contour independent classification of mammographic lesions. Characterization of breast tissue patterns is achieved by means of multidirectional Gabor filters. For the classification tasks, pattern recognition strategies, including Fisher linear discriminant analysis, Bayesian classifiers, support vector machines, and neural networks are applied using automatic selection of features and cross-validation techniques. Computer-aided detection of bilateral asymmetry resulted in accuracy up to 0.94, with sensitivity and specificity of 1 and 0.88, respectively. Computer-aided diagnosis of automatically detected lesions provided sensitivity of detection of malignant tumors in the range of [0.70, 0.81] at a range of falsely detected tumors of [0.82, 3.47] per image. The techniques presented in this work are effective in detecting and characterizing various mammographic signs of breast disease.

Architectural distortion is an important and early sign of breast cancer, but because of its subtlety, it is a common cause of false-negative findings on screening mammograms. Screening mammograms obtained prior to the detection of cancer could contain subtle signs of early stages of breast cancer, in particular, architectural distortion. This book presents image processing and pattern recognition techniques to detect architectural distortion in prior mammograms of interval-cancer cases. The methods are based upon Gabor filters, phase portrait analysis, procedures for the analysis of the angular spread of power, fractal analysis, Laws' texture energy measures derived from geometrically transformed regions of interest (ROIs), and Haralick's texture features. With Gabor filters and phase-portrait analysis, 4,224 ROIs were automatically obtained from 106 prior mammograms of 56 interval-cancer cases, including 301 true-positive ROIs related to architectural distortion, and from 52 mammograms of 13 normal cases. For each ROI, the fractal dimension, the entropy of the angular spread of power, 10 Laws' texture energy measures, and Haralick's 14 texture features were computed. The areas under the receiver operating characteristic (ROC) curves obtained using the features selected by stepwise logistic regression and the leave-one-image-out method are 0.77 with the Bayesian classifier, 0.76 with Fisher linear discriminant analysis, and 0.79 with a neural network classifier. Free-response ROC analysis indicated sensitivities of 0.80 and 0.90 at 5.7 and 8.8 false positives (FPs) per image, respectively, with the Bayesian classifier and the leave-one-image-out method. The present study has demonstrated the ability to detect early signs of breast cancer 15 months ahead of the time of clinical diagnosis, on the average, for interval-cancer cases, with a sensitivity of 0.8 at 5.7 FP/image. The presented computer-aided detection techniques, dedicated to accurate detection and localization of architectural distortion, could lead to efficient detection of early and subtle signs of breast cancer at pre-mass-formation stages. Table of Contents: Introduction / Detection of Early Signs of Breast Cancer / Detection and Analysis of Oriented Patterns / Detection of Potential Sites of Architectural Distortion / Experimental Set Up and Datasets / Feature Selection and Pattern Classification / Analysis of Oriented Patterns Related to Architectural Distortion / Detection of Architectural Distortion in Prior Mammograms / Concluding Remarks

X-ray mammography screening is the current mainstay for early breast cancer detection. It has been proven to detect breast cancer at an earlier stage and to reduce the number of women dying from the disease. However, it has a number of limitations. These current limitations in early breast cancer detection technology are driving a surge of new technological developments, from modifications of x-ray mammography such as computer programs that can indicate suspicious areas, to newer methods of detection such as magnetic resonance imaging (MRI) or biochemical tests on breast fluids. To explore the merits and drawbacks of these new breast cancer detection techniques, the Institute of Medicine of the National Academy of Sciences convened a committee of experts. During its year of operation, the committee examined the peer-reviewed literature, consulted with other experts in the field, and held two public workshops. In addition to identifying promising new technologies for early detection, the committee explored potential barriers that might prevent the development of new detection methods and their common usage. Such barriers could include lack of funding from agencies that support research and lack of investment in the commercial sector; complicated, inconsistent, or unpredictable federal regulations; inadequate insurance reimbursement; and limited access to or unacceptability of breast cancer detection technology for women and their doctors. Based on the findings of their study, the committee prepared a report entitled Mammography and Beyond: Developing Technology for Early Detection of Breast Cancer, which was published in the spring of 2001. This is a non-technical summary of that report.

This book provides a detailed assessment of the state of the art in automated techniques for the analysis of digital mammogram images. Topics covered include a variety of approaches for image processing and pattern recognition aimed at assisting the physician in the task of detecting tumors from evidence in mammogram images. The chapters are written by recognized experts in the field and are revised versions of papers selected from those presented at the “First International Workshop on Mammogram Image Analysis” held in San Jose as part of the 1993 Biomedical Image Processing conference. Contents:Automation in Mammography: Computer Vision and Human Perception (S Astley et al.)Restoration of Mammographic Images in the Presence of Signal-Dependent Noise (F Aghdasi et al.)Computer-Aided Detection and Diagnosis of Masses and Clustered Microcalcifications from Digital Mammograms (R M Nishikawa et al.)Feature Extraction for Computer-Aided Analysis of Mammograms (H Bårman et al.)Detection and Classification of Mammographic Calcifications (L Shen et al.)Comparative Evaluation of Pattern Recognition Techniques for Detection of Microcalcifications in Mammography (K S Woods et al.)Automated Detection of Breast Asymmetry Using Anatomical Features (P Miller & S Astley)Image Processing and Computer Aided Diagnosis in Digital Mammography “A Radiologist's Perspective” (E D Pisano & F Shtern)and other papers Readership: Computer scientists and biomedical engineers. keywords:Mammography;Digital;Analysis;Processing;Detection;Computer Vision;Image;Breast;Mammographic Image Analysis;Computer-Aided Diagnosis;Digital Mammography;Medical Image Analysis

The goal of this project is to demonstrate the clinical usefulness of computer-aided diagnosis (CAD) in mammographic detection of breast cancer. Our plan is to develop advanced CAD schemes for detection and characterization of clustered microcalcifications and masses by incorporating artificial neural networks and various image processing techniques. Clinical mammography workstations for automated detection of suspicious lesions in mammograms will be developed by integration of laser digitizer, high-speed computer and advanced CAD software. The prototype workstations will be used as a "second opinion" in interpreting mammograms by reducing observational errors. The outcomes of radiologists' image readings in the detection of breast cancer will be evaluated by examining radiologists' performance when reading films only and when reading film with the computer results. We believe that the outcomes of this demonstration project will lead to large-scale clinical trials and will result in commercial products for practical routine use in breast imaging.

Fractal analysis is useful in digital image processing for the characterization of shape roughness and gray-scale texture or complexity. Breast masses present shape and gray-scale characteristics in mammograms that vary between benign masses and malignant tumors. This book demonstrates the use of fractal analysis to classify breast masses as benign masses or malignant tumors based on the irregularity exhibited in their contours and the gray-scale variability exhibited in their mammographic images. A few different approaches are described to estimate the fractal dimension (FD) of the contour of a mass, including the ruler method, box-counting method, and the power spectral analysis (PSA) method. Procedures are also described for the estimation of the FD of the gray-scale image of a mass using the blanket method and the PSA method. To facilitate comparative analysis of FD as a feature for pattern classification of breast masses, several other shape features and texture measures are described in the book. The shape features described include compactness, spiculation index, fractional concavity, and Fourier factor. The texture measures described are statistical measures derived from the gray-level cooccurrence matrix of the given image. Texture measures reveal properties about the spatial distribution of the gray levels in the given image; therefore, the performance of texture measures may be dependent on the resolution of the image. For this reason, an analysis of the effect of spatial resolution or pixel size on texture measures in the classification of breast masses is presented in the book. The results demonstrated in the book indicate that fractal analysis is more suitable for characterization of the shape than the gray-level variations of breast masses, with area under the receiver operating characteristics of up to 0.93 with a dataset of 111 mammographic images of masses. The methods and results presented in the book are useful for computer-aided diagnosis of breast cancer. Table of Contents: Computer-Aided Diagnosis of Breast Cancer / Detection and Analysis of\newline Breast Masses / Datasets of Images of Breast Masses / Methods for Fractal Analysis / Pattern Classification / Results of Classification of Breast Masses / Concluding Remarks

This book provides a comprehensive description of the screening and clinical applications of digital breast tomosynthesis (DBT) and offers straightforward, clear guidance on use of the technique. Informative clinical cases are presented to illustrate how to take advantage of DBT in clinical practice. The importance of DBT as a diagnostic tool for both screening and diagnosis is increasing rapidly. DBT improves upon mammography by depicting breast tissue on a video clip made of cross‐sectional images reconstructed in correspondence with their mammographic planes of acquisition. DBT results in markedly reduced summation of overlapping breast tissue and offers the potential to improve mammographic breast cancer surveillance and diagnosis. This book will be an excellent practical teaching guide for beginners and a useful reference for more experienced radiologists.

In June 1998 the Fourth International Workshop on Digital Mammography was held in Nijmegen, The Netherlands, where it was hosted by the department of Radiology of the University Hospital Nijmegen. This series of meetings was initiated at the 1993 SPIE Biomedical Image Processing Conference in San Jose, USA, where a number of sessions were entirely devoted to mammographic image analysis. At very successful subsequent workshops held in York, UK (1994) and Chicago, USA (1996), the scope of the conference was broadened, establishing a platform for presentation and discussion of new developments in digital mammog raphy. Topics that are addressed at these meetings are computer-aided diagnosis, image processing, detector development, system design, observer performance and clinical evaluation. The goal is to bring researchers from universities, breast cancer experts, and engineers together, to exchange information and present new scientific developments in this rapidly evolving field. This book contains all the scientific papers and posters presented at the work shop in Nijmegen. Contributions came from as many as 20 different countries and 190 participants attended the meeting. At a technical exhibit companies demon strated new products and work in progress. Abstracts of all papers were reviewed by members of the scientific committee. Many of the accepted papers had excellent quality, but due to limited space not all of them could be included as full papers in these proceedings. Papers that were rated high by the reviewers are included as long or short papers, others appear as extended abstracts in the last chapter.