hua Li

Improving mass detection by adaptive and multiscale processing in digitized mammograms

A new CAD mass detection system was developed using adaptive and multi-scale processing methods for improving detection sensitivity/specificity, and its robustness to the variation in mammograms. The major techniques developed in system design include: (1) image standardization by applying a series of preprocessing to remove extrinsic signal, extract breast area, and normalize the image intensity; (2) multi- mode processing by decomposing image features using directional wavelet transform and non-linear multi-scale representation using anisotropic diffusion; (3) adaptive processing in image segmentation using localized adaptive thresholding and adaptive clustering; and (4) combined `hard-`soft classification by using a modified fuzzy decision tree and committee decision-making method. Evaluations and comparisons were taken with a training dataset containing 30 normal and 47 abnormal mammograms with totally 70 masses, and an independent testing dataset consisting of 100 normal images, 39 images with 48 minimal cancers and 25 images with 25 benign masses. A high detection performance of sensitivity TP equals 93% with false positive rate FP equals 3.1 per image and a good generalizability with TP equals 80% and FP equals 2.0 per image are obtained.

3D Finite Element Modeling of Nonrigid Breast Deformation for Feature Registration in -ray and MR Images

Registering features in multiple mammographic views is an important technique to improve breast cancer detection rate. However, nonrigid breast deformation during X-ray imaging poses a severe challenge to the conventional 2D registration methods. We present a method that utilizes a 3D model to facilitate two-view registration by predicting breast deformation. At first, a finite element model of a breast is constructed using its MRIs. The model is capable of simulating both compression and decompression. Feature registration is then accomplished through a series of projections and compression-decompression operations. Experiments using real patient data demonstrate that a mammographic feature can be successfully registered from one view to another

Correspondence recovery in 2-view mammography

Predicting breast tissue deformation is of great significance in various medical practice including diagnosis and surgery. In breast surgery, surgeons are often concerned with a specific portion of the breast, e.g., tumor, which must be located accurately. Clinically it is important to combine information provided by images from several modalities or at different times. However, images taken by various techniques are often obtained under entirely different tissue configurations, compression, orientation or body position. Hence, some form of spatial nonrigid transformation of image data is required so that the tissues are represented in an equivalent configuration. In this paper we propose finite element method (FEM) based strategy for correspondence identification between image features identified in two view mammography. The algorithm models breast compression during mammography and allows for correspondence recovery of 2D features found in two views and reconstruction of their 3D locations. The algorithm was tested on MRI and mammography images of triple modality phantom and limited set of mammographic images.

Three dimensional finite element model for lesion correspondence in breast imaging

Predicting breast tissue deformation is of great significance in several medical applications such as biopsy, diagnosis, and surgery. In breast surgery, surgeons are often concerned with a specific portion of the breast, e.g., tumor, which must be located accurately beforehand. Also clinically it is important for combining the information provided by images from several modalities or at different times, for the detection/diagnosis, treatment planning and guidance of interventions. Multi-modality imaging of the breast obtained by X-ray mammography, MRI is thought to be best achieved through some form of data fusion technique. However, images taken by these various techniques are often obtained under entirely different tissue configurations, compression, orientation or body position. In these cases some form of spatial transformation of image data from one geometry to another is required such that the tissues are represented in an equivalent configuration. We propose to use a 3D finite element model for lesion correspondence in breast imaging. The novelty of the approach lies in the following facts: (1) Finite element is the most accurate technique for modeling deformable objects such as breast. The physical soundness and mathematical rigor of finite element method ensure the accuracy and reliability of breast modeling that is essential for lesion correspondence. (2) When both MR and mammographic images are available, a subject-specific 3D breast model will be built from MRIs. If only mammography is available, a generic breast model will be used for two-view mammography reading. (3) Incremental contact simulation of breast compression allows accurate capture of breast deformation and ensures the quality of lesion correspondence. (4) Balance between efficiency and accuracy is achieved through adaptive meshing. We have done intensive research based on phantom and patient data.

Wavelet-based image processing for digital mammography

This paper is to evaluate the importance of image preprocessing using multiresolution and multiorientation wavelet transforms on the performance of a previously reported computer assisted diagnostic (CAD) method for breast cancer screening, using digital mammography. An analysis of the influence of WTs on image feature extraction for mass detection is achieved by comparing the discriminate ability of features extracted with and without wavelet based image preprocessing using computed ROC. Three indexes are proposed to assess the segmentation of the mass area with comparison to ground truth. Dat was analyzed on region-of- interest database that included mass and normal regions from digitized mammograms with ground truth. The metrics for measurement of segmentation of the mass clearly demonstrates the importance of image preprocessing methods. Similarly, the relative improvement in performance is observed in feature extraction, where the Az values are increased. The improvement depends on the feature characteristics. The use of methodology in this paper result sin a significant improvement in feature extraction for the previously proposed CAD detection method. We are therefore exploring additional improvement in wavelet based image preprocessing methods, including adaptive methods, to achieve a further improvement in performance on larger image databases.

Classification of masses on mammograms using support vector machine

Mammography is the most effective method for early detection of breast cancer. However, the positive predictive value for classification of malignant and benign lesion from mammographic images is not very high. Clinical studies have shown that most biopsies for cancer are very low, between 15% and 30%. It is important to increase the diagnostic accuracy by improving the positive predictive value to reduce the number of unnecessary biopsies. In this paper, a new classification method was proposed to distinguish malignant from benign masses in mammography by Support Vector Machine (SVM) method. Thirteen features were selected based on receiver operating characteristic (ROC) analysis of classification using individual feature. These features include four shape features, two gradient features and seven Laws features. With these features, SVM was used to classify the masses into two categories, benign and malignant, in which a Gaussian kernel and sequential minimal optimization learning technique are performed. The data set used in this study consists of 193 cases, in which there are 96 benign cases and 97 malignant cases. The leave-one-out evaluation of SVM classifier was taken. The results show that the positive predict value of the presented method is 81.6% with the sensitivity of 83.7% and the false-positive rate of 30.2%. It demonstrated that the SVM-based classifier is effective in mass classification.

Image feature extraction for mass detection in digital mammography: effects of wavelet analysis

Multiresolution and multiorientation wavelet transforms (WTs), as the key CAD modules, were used in our previous study of CAD mass detection. The objective of this paper is to evaluate the roles of these WTs modules in the proposed CAD approach. A statistical analysis of the effects of WTs on image feature extraction for mass detection is taken including the effects of WTs on mass segmentation and a comparative study of discrimination ability of features extracted with WTs based and non-WTs based segmentation method. Three indexes are proposed to asses the segmentation. The effects of WTs on feature extraction are evaluated using ROC analysis of the feature discrimination ability. The statistical analysis demonstrates that the use of WTs modules results in a significant improvement in feature extraction for the previously proposed CAD mass detection method. The improvement, however, depends on the feature characteristics, large for boundary-related features while small for intensity-related features.

A Computer Assisted Diagnostic System for Mass Detection

Computer assisted diagnostic (CAD) methods have been proposed as a “second opinion” strategy for breast cancer screening using digital mammography. The reported methods have included the detection of either microcalcification clusters or masses [1]. Mass detection poses a more difficult problem compared to microcalcification cluster detection because masses are often: (a) of varying size, shape, and density, (b) exhibit poor image contrast, (c) are highly connected to the surrounding parenchymal tissue density, particularly for spiculated lesions, and (d) are surrounded by non-uniform tissue background with similar characteristics [2]–[4]. The segmentation of masses and the computation of related pixel intensity, morphological, and directional texture features poses difficult problems in terms of improved feature extraction methods, as required for classification methods that distinguish masses from normal tissues. Improved and robust feature extraction has not been emphasizes in the literature. Examples include features such as mass shape or mass margin analysis, or spiculations for spiculated lesions [5]. Mass detection is proposed here as a good clinical model for the motivation for proposing a new class of adaptive CAD methods for image preprocessing, to improve feature extraction. The methods proposed are useful for other CAD applications such as the detection of microcalcifications and lung nodules.

3D Object Localization and Visualization in Breast Tomosynthesis

This paper presents a preliminary study of CAD in breast tomosynthesis. The problem addressed is how to discriminate the objects on the in-focus plane from its blurred image superimposed on other planes. Instead of developing a preprocessing module to suppress the tomosynthesis artifacts in the reconstruction, we locate the object by utilizing the inherent 3D information in the reconstructed planar images. This study was done on tomosynthesis data of a stereo breast biopsy phantom (RMI 164A) reconstructed with the simple backprojection (or shift-and-add) method. The experimental results demonstrate the ability of the proposed method to identify the focal plane of objects. It can potentially be applied to remove the blurred out-of-plane structure and improve the clinical application of breast tomosynthesis.

Graph-based region growing for mass-segmentation in digital mammography

Mass segmentation is a vital step in CAD mass detection and classification. A challenge for mass segmentation in mammograms is that masses may contact with some surrounding tissues, which have the similar intensity. In this paper, a novel graph-based algorithm has been proposed to segment masses in mammograms. In the proposed algorithm, the procedure of region growing is represented as a growing tree whose root is the selected seed. Active leaves, which have the ability to grow, in the connection area between adjacent regions are deleted to stop growing, then separating the adjacent regions while keeping the spiculation of masses, which is a primary sign of malignancy for masses. The new constrained segmentation was tested with 20 cases in USF moffitt mammography database against the conventional region growing algorithm. The segmented mass regions were evaluated in terms of the overlap area with annotations made by the radiologist. We found that the new graph-based segmentation more closely match radiologists outlines of these masses.