Books+ Search Results
Computer-aided detection of architectural distortion in prior mammograms of interval cancer
Author
Title
Computer-aided detection of architectural distortion in prior mammograms of interval cancer [electronic resource] / Shantanu Banik, Rangaraj M. Rangayyan, and J.E. Leo Desautels.
ISBN
9781627050838 (electronic bk.)
9781627050821 (pbk.)
Published
San Rafael, Calif. (1537 Fourth Street, San Rafael, CA 94901 USA) : Morgan & Claypool, c2013.
Physical Description
1 online resource (xxiv, 169 p.) : ill., digital file.
Local Notes
Access is available to the Yale community.
Notes
Part of: Synthesis digital library of engineering and computer science.
Series from website.
Title from PDF t.p. (viewed on February 17, 2013).
Access and use
Access restricted by licensing agreement.
Summary
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.
Other formats
Also available in print.
Print version:
Format
Books / Online
Language
English
Added to Catalog
April 24, 2013
Series
Synthesis lectures on biomedical engineering, # 47
System details note
Mode of access: World Wide Web.
System requirements: Adobe Acrobat Reader.
Bibliography
Includes bibliographical references (p. 145-165).
Contents
Preface
Acknowledgments
List of symbols and abbreviations
1. Introduction
1.1 Breast cancer and mammography
1.2 Mammographic signs of breast cancer
1.2.1 Masses
1.2.2 Calcifications
1.2.3 Bilateral asymmetry
1.2.4 Architectural distortion
1.3 Event detection in medical images
1.3.1 Sensitivity and specificity
1.3.2 Receiver operating characteristics
1.3.3 Free-response receiver operating characteristics
1.4 Computer-aided diagnosis of breast cancer
1.5 Analysis of prior mammograms
1.6 Organization of the book
1.6.1 Hypothesis and aim of the work
1.6.2 Structure of the book
2. Detection of early signs of breast cancer
2.1 Detection of architectural distortion
2.2 Detection of architectural distortion by CAD systems
2.3 Detection of signs of cancer in prior mammograms
2.4 Remarks
3. Detection and analysis of oriented patterns
3.1 Oriented texture in biomedical images
3.2 Gabor filters
3.2.1 The real Gabor filter
3.2.2 The complex Gabor filter
3.3 Analysis of oriented patterns using phase portraits
3.3.1 Phase portraits
3.4 Analysis of orientation fields using phase portraits
3.5 Illustrative example
3.6 Remarks
4. Detection of potential sites of architectural distortion
4.1 Detection of architectural distortion
4.1.1 Segmentation of the breast portion in a mammogram
4.1.2 Extraction of the orientation field
4.1.3 Selection of curvilinear structures
4.1.4 Filtering and downsampling the orientation field
4.1.5 Estimating the phase-portrait maps
4.1.6 Shape-constrained phase-portrait model
4.2 Potential sites of architectural distortion
4.3 Remarks
5. Experimental set up and datasets
5.1 Datasets of mammograms
5.1.1 Interval cancer
5.1.2 Screen-detected cancer
5.2 Selection of ROIs
5.2.1 Interval-cancer cases
5.2.2 Screen-detected cancer cases
5.3 Remarks
6. Feature selection and pattern classification
6.1 Analysis of features and pattern classification
6.2 Characterization of features
6.2.1 Student's t-test and the p-value
6.2.2 The receiver operating characteristic (ROC) curve
6.2.3 FROC analysis
6.3 Feature selection
6.3.1 Logistic regression
6.3.2 Stepwise regression
6.4 Pattern classification
6.4.1 Fisher linear discriminant analysis
6.4.2 The Bayesian classifier
6.4.3 Neural networks
6.4.4 Support vector machines
6.5 Training and test sets
6.5.1 Cross-validation
6.5.2 The leave-one-out method
6.5.3 Effects of sample size and bias
6.6 Remarks
7. Analysis of oriented patterns related to architectural distortion
7.1 Features and patterns related to architectural distortion
7.2 Node value
7.3 Frequency-domain methods
7.3.1 Design of geometric transformations for spectral analysis
7.3.2 Fractal analysis
7.3.3 Estimation of FD
7.3.4 Angular spread of power
7.4 Analysis of texture
7.4.1 Statistical analysis using Haralick's measures
7.4.2 Structural analysis of texture using Laws' energy measures
7.4.3 Design of geometric transformations for the analysis of oriented patterns
7.5 Characterization of angular dispersion
7.5.1 Coherence
7.5.2 Orientation strength
7.5.3 Characterization of angular spread
7.6 Measures of angular dispersion
7.6.1 Tsallis entropy
7.6.2 Renyi entropy
7.7 Analysis of performance of individual features
7.8 Remarks
8. Detection of architectural distortion in prior mammograms
8.1 Analysis of performance of selected features from various sets
8.1.1 Statistical significance of differences in performance
8.1.2 Effects of various types of cross-validation and training sets
8.1.3 Analysis of performance with various methods of cross-validation and feature selection
8.2 Analysis of performance with subsets of the interval-cancer dataset
8.3 Cross-validation of performance in detection using different datasets
8.4 Comparative analysis and discussion
8.5 Remarks
9. Concluding remarks
A. List of empirically selected parameters
References
Authors' biographies.
Subjects
Subjects (Medical)
Breast Neoplasms - diagnosis.
Also listed under
Bookmark As
Citation
