Dorit D. Adler

Computerized analysis of mammographic microcalcifications in morphological and texture feature spaces

We are developing computerized feature extraction and classification methods to analyze malignant and benign microcalcifications on digitized mammograms. Morphological features that described the size, contrast, and shape of microcalcifications and their variations within a cluster were designed to characterize microcalcifications segmented from the mammographic background. Texture features were derived from the spatial gray-level dependence (SGLD) matrices constructed at multiple distances and directions from tissue regions containing microcalcifications. A genetic algorithm (GA) based feature selection technique was used to select the best feature subset from the multi-dimensional feature spaces. The GA-based method was compared to the commonly used feature selection method based on the stepwise linear discriminant analysis (LDA) procedure. Linear discriminant classifiers using the selected features as input predictor variables were formulated for the classification task. The discriminant scores output from the classifiers were analyzed by receiver operating characteristic (ROC) methodology and the classification accuracy was quantified by the area, Az, under the ROC curve. We analyzed a data set of 145 mammographic microcalcification clusters in this study. It was found that the feature subsets selected by the GA-based method are comparable to or slightly better than those selected by the stepwise LDA method. The texture features (Az = 0.84) were more effective than morphological features (Az = 0.79) in distinguishing malignant and benign microcalcifications. The highest classification accuracy (Az = 0.89) was obtained in the combined texture and morphological feature space. The improvement was statistically significant in comparison to classification in either the morphological (p = 0.002) or the texture (p = 0.04) feature space alone. The classifier using the best feature subset from the combined feature space and an appropriate decision threshold could correctly identify 35% of the benign clusters without missing a malignant cluster. When the average discriminant score from all views of the same cluster was used for classification, the Az value increased to 0.93 and the classifier could identify 50% of the benign clusters at 100% sensitivity for malignancy. Alternatively, if the minimum discriminant score from all views of the same cluster was used, the Az value would be 0.90 and a specificity of 32% would be obtained at 100% sensitivity. The results of this study indicate the potential of using combined morphological and texture features for computer-aided classification of microcalcifications.

Computer-aided classification of mammographic masses and normal tissue: linear discriminant analysis in texture feature space

We studied the effectiveness of using texture features derived from spatial grey level dependence (SGLD) matrices for classification of masses and normal breast tissue on mammograms. One hundred and sixty-eight regions of interest (ROIS) containing biopsy-proven masses and 504 ROIS containing normal breast tissue were extracted from digitized mammograms for this study. Eight features were calculated for each ROI. The importance of each feature in distinguishing masses from normal tissue was determined by stepwise linear discriminant analysis. Receiver operating characteristic (ROC) methodology was used to evaluate the classification accuracy. We investigated the dependence of classification accuracy on the input features, and on the pixel distance and bit depth in the construction of the SGLD matrices. It was found that five of the texture features were important for the classification. The dependence of classification accuracy on distance and bit depth was weak for distances greater than 12 pixels and bit depths greater than seven bits. By randomly and equally dividing the data set into two groups, the classifier was trained and tested on independent data sets. The classifier achieved an average area under the ROC curve, Az, of 0.84 during training and 0.82 during testing. The results demonstrate the feasibility of using linear discriminant analysis in the texture feature space for classification of true and false detections of masses on mammograms in a computer-aided diagnosis scheme.

Image feature selection by a genetic algorithm: application to classification of mass and normal breast tissue.

We investigated a new approach to feature selection, and demonstrated its application in the task of differentiating regions of interest (ROIs) on mammograms as either mass or normal tissue. The classifier included a genetic algorithm (GA) for image feature selection, and a linear discriminant classifier or a backpropagation neural network (BPN) for formulation of the classifier outputs. The GA-based feature selection was guided by higher probabilities of survival for fitter combinations of features, where the fitness measure was the area Az under the receiver operating characteristic (ROC) curve. We studied the effect of different GA parameters on classification accuracy, and compared the results to those obtained with stepwise feature selection. The data set used in this study consisted of 168 ROIs containing biopsy-proven masses and 504 ROIs containing normal tissue. From each ROI, a total of 587 features were extracted, of which 572 were texture features and 15 were morphological features. The GA was trained and tested with several different partitionings of the ROIs into training and testing sets. With the best combination of the GA parameters, the average test Az value using a linear discriminant classifier reached 0.90, as compared to 0.89 for stepwise feature selection. Test Az values with a BPN classifier and a more limited feature pool were 0.90 with GA-based feature selection, and 0.89 for stepwise feature selection. The use of a GA in tailoring classifiers with specific design characteristics was also discussed. This study indicates that a GA can provide versatility in the design of linear or nonlinear classifiers without a trade-off in the effectiveness of the selected features.

Doppler ultrasound color flow imaging in the study of breast cancer: Preliminary findings

A prospective study of the Doppler color flow features of 55 proved breast cancers was performed. On a three-level scale of low to marked vascularity, visual assessment of the color flow images classified 82% of the cancers as moderately or markedly vascular (minimal: 14%, moderate: 29%, marked: 53%). Four percent of the cancers had no detectable flow. In 29 women, a volume of tissue comparable to the cancer was scanned in the contralateral normal breast. Sixty-nine percent of the normal breasts had moderate or marked vascularity (minimal: 28%, moderate: 41%, marked: 28%), and 3% were avascular. There was poor distinction between normal tissues and cancer which suggests that more sensitive Doppler methods than were employed in this study may be needed in order to detect the small vessel flow reported to be rather specific for malignancy. The high, 82%, detection rate of tumor vessels in this study suggests the potential use of color flow Doppler for directing more specific but lengthy Doppler procedures.

AUTOMATED DETECTION OF BREAST MASSES ON MAMMOGRAMS USING ADAPTIVE CONTRAST ENHANCEMENT AND TEXTURE CLASSIFICATION

This paper presents segmentation and classification results of an automated algorithm for the detection of breast masses on digitized mammograms. Potential mass regions were first identified using density-weighted contrast enhancement (DWCE) segmentation applied to single-view mammograms. Once the potential mass regions had been identified, multiresolution texture features extracted from wavelet coefficients were calculated, and linear discriminant analysis (LDA) was used to classify the regions as breast masses or normal tissue. In this article the overall detection results for two independent sets of 84 mammograms used alternately for training and test were evaluated by free-response receiver operating characteristics (FROC) analysis. The test results indicate that this new algorithm produced approximately 4.4 false positive per image at a true positive detection rate of 90% and 2.3 false positives per image at a true positive rate of 80%.

Classification of mass and normal breast tissue on digital mammograms: Multiresolution texture analysis

We investigated the feasibility of using multiresolution texture analysis for differentiation of masses from normal breast tissue on mammograms. The wavelet transform was used to decompose regions of interest (ROIs) on digitized mammograms into several scales. Multiresolution texture features were calculated from the spatial gray level dependence matrices of (1) the original images at variable distances between the pixel pairs, (2) the wavelet coefficients at different scales, and (3) the wavelet coefficients up to certain scale and then at variable distances between the pixel pairs. In this study, 168 ROIs containing biopsy-proven masses and 504 ROIs containing normal parenchyma were used as the data set. The mass ROIs were randomly and equally divided into training and test groups along with corresponding normal ROIs from the same film. Stepwise linear discriminant analysis was used to select optimal features from the multiresolution texture feature space to maximize the separation of mass and normal tissue for all ROIs. We found that texture features at large pixel distances are important for the classification task. The wavelet transform can effectively condense the image information into its coefficients. With texture features based on the wavelet coefficients and variable distances, the area Az under the receiver operating characteristic curve reached 0.89 and 0.86 for the training and test groups, respectively. The results demonstrate that a linear discriminant classifier using the multiresolution texture features can effectively classify masses from normal tissue on mammograms.

Computerized classification of malignant and benign microcalcifications on mammograms: texture analysis using an artificial neural network.

We investigated the feasibility of using texture features extracted from mammograms to predict whether the presence of microcalcifications is associated with malignant or benign pathology. Eighty-six mammograms from 54 cases (26 benign and 28 malignant) were used as case samples. All lesions had been recommended for surgical biopsy by specialists in breast imaging. A region of interest (ROI) containing the microcalcifications was first corrected for the low-frequency background density variation. Spatial grey level dependence (SGLD) matrices at ten different pixel distances in both the axial and diagonal directions were constructed from the background-corrected ROI. Thirteen texture measures were extracted from each SGLD matrix. Using a stepwise feature selection technique, which maximized the separation of the two class distributions, subsets of texture features were selected from the multi-dimensional feature space. A backpropagation artificial neural network (ANN) classifier was trained and tested with a leave-one-case-out method to recognize the malignant or benign microcalcification clusters. The performance of the ANN was analysed with receiver operating characteristic (ROC) methodology. It was found that a subset of six texture features provided the highest classification accuracy among the feature sets studied. The ANN classifier achieved an area under the ROC curve of 0.88. By setting an appropriate decision threshold, 11 of the 28 benign cases were correctly identified (39% specificity) without missing any malignant cases (100% sensitivity) for patients who had undergone biopsy. This preliminary result indicates that computerized texture analysis can extract mammographic information that is not apparent by visual inspection. The computer-extracted texture information may be used to assist in mammographic interpretation, with the potential to reduce biopsies of benign cases and improve the positive predictive value of mammography.

Digitization requirements in mammography: Effects on computer‐aided detection of microcalcifications

We have developed a computerized method for detection of microcalcifications on digitized mammograms. The program has achieved an accuracy that can detect subtle microcalcifications which may potentially be missed by radiologists. In this study, we evaluated the dependence of the detection accuracy on the pixel size and pixel depth of the digitized mammograms. The mammograms were digitized with a laser film scanner at a pixel size of 0.035 mm x0.035 mm and 12-bit gray levels. Digitization with larger pixel sizes or fewer number of bits was simulated by averaging adjacent pixels or by eliminating the least significant bits, respectively. The SNR enhancement filter and the signal-extraction criteria in the computer program were adjusted to maximize the accuracy of signal detection for each pixel size. The overall detection accuracy was compared using the free response receiver operating characteristic curves. The results indicate that the detection accuracy decreases significantly as the pixel size increases from 0.035 mm x 0.035 mm to 0.07 mm x 0.07 mm (P < 0.007) and from 0.07 mm x 0.07 mm to 0.105 mm x 0.105 mm (P < 0.002). The detection accuracy is essentially independent of pixel depth from 12 to 9 bits and decreases significantly (P < 0.003) from 9 to 8 bits; a rapid decrease is observed as the pixel depth decreases further from 8 to 7 bits (P < 0.03) or from 7 to 6 bits (P < 0.02).(ABSTRACT TRUNCATED AT 250 WORDS)

Computer-aided diagnosis in mammography: classification of mass and normal tissue by texture analysis

Computer-aided diagnosis schemes are being developed to assist radiologists in mammographic interpretation. In this study, we investigated whether texture features could be used to distinguish between mass and non-mass regions in clinical mammograms. Forty-five regions of interest (ROIs) containing true masses with various degrees of visibility and 135 ROIs containing normal breast parenchyma were extracted manually from digitized mammograms as case samples. Spatial-grey-level-dependence (SGLD) matrices of each ROI were calculated and eight texture features were calculated from the SGLD matrices. The correlation and class-distance properties of extracted texture features were analysed. Selected texture features were input into a modified decision-tree classification scheme. The performance of the classifier was evaluated for different feature combinations and orders of features on the tree. A classification accuracy of about 89% sensitivity and 76% specificity was obtained for ordered features, sum average, correlation, and energy, during the training procedure. With a leave-one-out method, the test result was about 76% sensitivity and 64% specificity. The results of this preliminary study demonstrate the feasibility of using texture information for classification of mass and normal breast tissue, which will be likely to be useful for classifying true and false detections in computer-aided diagnosis programmes.

False‐positive reduction technique for detection of masses on digital mammograms: Global and local multiresolution texture analysis

We investigated the application of multiresolution global and local texture features to reduce false-positive detection in a computerized mass detection program. One hundred and sixty-eight digitized mammograms were randomly and equally divided into training and test groups. From these mammograms, two datasets were formed. The first dataset (manual) contained four regions of interest (ROIs) selected manually from each of the mammograms. One of the four ROIs contained a biopsy-proven mass and the other three contained normal parenchyma, including dense, mixed dense/fatty, and fatty tissues. The second dataset (hybrid) contained the manually extracted mass ROIs, along with normal tissue ROIs extracted by an automated Density-Weighted Contrast Enhancement (DWCE) algorithm as false-positive detections. A wavelet transform was used to decompose an ROI into several scales. Global texture features were derived from the low-pass coefficients in the wavelet transformed images. Local texture features were calculated from the suspicious object and the peripheral subregions. Linear discriminant models using effective features selected from the global, local, or combined feature spaces were established to maximize the separation between masses and normal tissue. Receiver Operating Characteristic (ROC) analysis was conducted to evaluate the classifier performance. The classification accuracy using global features were comparable to that using local features. With both global and local features, the average area, Az, under the test ROC curve, reached 0.92 for the manual dataset and 0.96 for the hybrid dataset, demonstrating statistically significant improvement over those obtained with global or local features alone. The results indicated the effectiveness of the combined global and local features in the classification of masses and normal tissue for false-positive reduction.