Reduction of false-positive recalls using a computerized mammographic image feature analysis scheme

Prediction of Short-Term Breast Cancer Risk with Fusion of CC- and MLO-Based Risk Models in Four-View Mammograms

This study performed and assessed a novel program to improve the accuracy of short-term breast cancer risk prediction by using information from craniocaudal (CC) and mediolateral-oblique (MLO) views of two breasts. An age-matched dataset of 556 patients with at least two sequential full-field digital mammography examinations was applied. In the second examination, 278 cases were diagnosed and pathologically verified as cancer, and 278 were negative, while all cases in the first examination were negative (not recalled). Two generalized linear-model-based risk prediction models were established with global- and local-based bilateral asymmetry features for CC and MLO views first. Then, a new fusion risk model was developed by fusing prediction results of the CC- and MLO-based risk models with an adaptive alpha-integration-based fusion method. The AUC of the fusion risk model was 0.72 ± 0.02, which was significantly higher than the AUC of CC- or MLO-based risk model (P < 0.05). The maximum odds ratio for CC- and MLO-based risk models were 8.09 and 5.25, respectively, and increased to 11.99 for the fusion risk model. For subgroups of patients aged 37-49 years, 50-65 years, and 66-87 years, the AUCs of 0.73, 0.71, and 0.75 for the fusion risk model were higher than AUC for CC- and MLO-based risk models. For the BIRADS 2 and 3 subgroups, the AUC values were 0.72 and 0.71 respectively for the fusion risk model which were higher than the AUC for the CC- and MLO-based risk models. This study demonstrated that the fusion risk model we established could effectively derive and integrate supplementary and useful information extracted from both CC and MLO view images and adaptively fuse them to increase the predictive power of the short-term breast cancer risk assessment model.

Improving mammography lesion classification by optimal fusion of handcrafted and deep transfer learning features

Objective.Handcrafted radiomics features or deep learning model-generated automated features are commonly used to develop computer-aided diagnosis schemes of medical images. The objective of this study is to test the hypothesis that handcrafted and automated features contain complementary classification information and fusion of these two types of features can improve CAD performance.Approach.We retrospectively assembled a dataset involving 1535 lesions (740 malignant and 795 benign). Regions of interest (ROI) surrounding suspicious lesions are extracted and two types of features are computed from each ROI. The first one includes 40 radiomic features and the second one includes automated features computed from a VGG16 network using a transfer learning method. A single channel ROI image is converted to three channel pseudo-ROI images by stacking the original image, a bilateral filtered image, and a histogram equalized image. Two VGG16 models using pseudo-ROIs and 3 stacked original ROIs without pre-processing are used to extract automated features. Five linear support vector machines (SVM) are built using the optimally selected feature vectors from the handcrafted features, two sets of VGG16 model-generated automated features, and the fusion of handcrafted and each set of automated features, respectively.Main Results.Using a 10-fold cross-validation, the fusion SVM using pseudo-ROIs yields the highest lesion classification performance with area under ROC curve (AUC = 0.756 ± 0.042), which is significantly higher than those yielded by other SVMs trained using handcrafted or automated features only (p < 0.05).Significance.This study demonstrates that both handcrafted and automated futures contain useful information to classify breast lesions. Fusion of these two types of features can further increase CAD performance.

Improving Performance of Breast Cancer Risk Prediction by Incorporating Optical Density Image Feature Analysis: An Assessment

RATIONALE AND OBJECTIVES

The purpose of this study is to improve accuracy of near-term breast cancer risk prediction by applying a new mammographic image conversion method combined with a two-stage artificial neural network (ANN)-based classification scheme.

MATERIALS AND METHODS

The dataset included 168 negative mammography screening cases. In developing and testing our new risk model, we first converted the original grayscale value (GV)-based mammographic images into optical density (OD)-based images. For each case, our computer-aided scheme then computed two types of image features representing bilateral asymmetry and the maximum of the image features computed from GV and OD images, respectively. A two-stage classification scheme consisting of three ANNs was developed. The first stage included two ANNs trained using features computed separately from GV and OD images of 138 cases. The second stage included another ANN to fuse the prediction scores produced by two ANNs in the first stage. The risk prediction performance was tested using the rest 30 cases.

RESULTS

With the two-stage classification scheme, the computed area under the receiver operating characteristic curve (AUC) was  0.816 ± 0.071, which was significantly higher than the AUC values of 0.669 ± 0.099 and 0.646 ± 0.099 achieved using two ANNs trained using GV features and OD features, respectively (P < .05).

CONCLUSION

This study demonstrated that applying an OD image conversion method can acquire new complimentary information to those acquired from the original images. As a result, fusion image features computed from these two types of images yielded significantly higher performance in near-term breast cancer risk prediction.

Pattern classification for breast lesion on FFDM by integration of radiomics and deep features

The radiomics model can be used in breast cancer detection via calculating quantitative image features. However, these features are explicitly designed, or handcrafted in advance, and this would limit their ability to characterize the lesion properly. This paper aims to build an integrated-features-based classification framework which cooperate the radiomics features and the deep features to classify benign and malignant breast lesions on full-filed digital mammography (FFDM). We propose a classification framework consists of three steps: (1) handcrafted features (HCFs) extraction and selection, (2) deep features (DFs) extraction and (3) the integrated features-based classification. Specifically, HCFs comprise the gray-level gap-length matrix (GLGLM) texture features and shape features, and DFs contain the pooled features and high-level fully-connected features. Then, a multi-classifier method is applied to construct our classification framework using integrated features for breast lesion classification. A total of 106 retrospective FFDM data (51 are malignant and 55 are benign) in both craniocaudal (CC) view and mediolateral oblique (MLO) view were included in this study. The areas under a receiver operating characteristic curve (AUC) value, accuracy, sensitivity, specificity and Youden's index, are used to examine the performance of our proposed method in differentiating benign and malignant breast lesion. Proposed framework trained on the concatenation of fully-connected features and HCFs can significantly improve classification performance (AUC of 94.6 %, accuracy of 96.4 %, sensitivity of 93.6 %, specificity of 98.9 % and Yonden's index of 92.5 %) compared with other features sets. Experimental results demonstrate that performance of proposed framework is improved, indicating the potential of concatenation of the fully-connected features and HCFs set in breast cancer patients.

A novel method of determining breast cancer risk using parenchymal textural analysis of mammography images on an Asian cohort

Historically, breast cancer risk prediction models are based on mammographic density measures, which are dichotomous in nature and generally categorize each voxel or area of the breast parenchyma as 'dense' or 'not dense'. Using these conventional methods, the structural patterns or textural components of the breast tissue elements are not considered or ignored entirely. This study presents a novel method to predict breast cancer risk that combines new texture and mammographic density based image features. We performed a comprehensive study of the correlation of 944 new and conventional texture and mammographic density features with breast cancer risk on a cohort of Asian women. We studied 250 breast cancer cases and 250 controls matched at full-field digital mammography (FFDM) status for age, BMI and ethnicity. Stepwise regression analysis identified relevant features to be included in a linear discriminant analysis (LDA) classifier model, trained and tested using a leave-one-out based cross-validation method. The area under the receiver operating characteristic (AUC) and adjusted odds ratios (ORs) were used as the two performance assessment indices in our study. For the LDA trained classifier, the adjusted OR was 6.15 (95% confidence interval: 3.55-10.64) and for Volpara volumetric breast density, 1.10 (0.67-1.81). The AUC for the LDA trained classifier was 0.68 (0.64-0.73), compared to 0.52 (0.47-0.57) for Volpara volumetric breast density (p   <  0.001). The regression analysis of OR values for the LDA classifier also showed a significant increase in slope (p   <  0.02). Mammographic texture features derived from digital mammograms are important quantitative measures for breast cancer risk assessment based models. Parenchymal texture analysis has an important role for stratifying breast cancer risk in women, which can be implemented to routine breast cancer screening strategies.

[Establishment of a deep feature-based classification model for distinguishing benign and malignant breast tumors on full-filed digital mammography]

OBJECTIVE

To develop a deep features-based model to classify benign and malignant breast lesions on full- filed digital mammography.

METHODS

The data of full-filed digital mammography in both craniocaudal view and mediolateral oblique view from 106 patients with breast neoplasms were analyzed. Twenty-three handcrafted features (HCF) were extracted from the images of the breast tumors and a suitable feature set of HCF was selected using t-test. The deep features (DF) were extracted from the 3 pre-trained deep learning models, namely AlexNet, VGG16 and GoogLeNet. With abundant breast tumor information from the craniocaudal view and mediolateral oblique view, we combined the two extracted features (DF and HCF) as the two-view features. A multi-classifier model was finally constructed based on the combined HCF and DF sets. The classification ability of different deep learning networks was evaluated.

RESULTS

Quantitative evaluation results showed that the proposed HCF+DF model outperformed HCF model, and AlexNet produced the best performances among the 3 deep learning models.

CONCLUSIONS

The proposed model that combines DF and HCF sets of breast tumors can effectively distinguish benign and malignant breast lesions on full-filed digital mammography.

Applying a new maximum local asymmetry feature analysis method to improve near-term breast cancer risk prediction

Quantitative assessment of mammographic asymmetry has been investigated for breast cancer risk prediction. A new asymmetry feature extraction method was proposed in this study to enhance the risk prediction accuracy of near-term breast cancer. Breast areas in each pair of bilateral mammographic images were divided into several pairs of matched local annular regions and the maximum local asymmetry features (MLAF) were extracted from these regions. Radial basis function network (RBFN) was used to merge these features for breast cancer risk prediction. The dataset included 560 negative subjects. The risk prediction performance was tested using a leave-one-case-out (LOCO) cross-validation method. Area under the receiver operating characteristic curve (AUC) was used as the risk prediction performance evaluation index. AUC  =  0.898  ±  0.013 was obtained by using the MLAFs extracted from the annular regions, which was significantly higher than the AUC value of 0.505  ±  0.025 achieved by using global asymmetry features computed from the whole breast regions (p  <  0.05, DeLong's test) and much higher than the AUC values of 0.825  ±  0.017 and 0.717  ±  0.021 achieved by using MLAFs extracted from horizontal strip regions and vertical strip regions. The study demonstrated that near-term breast cancer risk prediction could be improved by using the proposed feature extraction method.

False-positive reduction in computer-aided mass detection using mammographic texture analysis and classification

BACKGROUND AND OBJECTIVE

The aim of computer-aided-detection (CAD) systems for mammograms is to assist radiologists by marking region of interest (ROIs) depicting abnormalities. However, the confusing appearance of some normal tissues that visually look like masses results in a large proportion of marked ROIs with normal tissues. This paper copes with this problem and proposes a framework to reduce false positive masses detected by CAD.

METHODS

To avoid the error induced by the segmentation step, we proposed a segmentation-free framework with particular attention to improve feature extraction and classification steps. We investigated for the first time in mammogram analysis, Hilbert's image representation, Kolmogorov-Smirnov distance and maximum subregion descriptors. Then, a feature selection step is performed to select the most discriminative features. Moreover, we considered several classifiers such as Random Forest, Support Vector Machine and Decision Tree to distinguish between normal tissues and masses. Our experiments were carried out on a large dataset of 10168 ROIs (8254 normal tissues and 1914 masses) constructed from the Digital Database for Screening Mammography (DDSM). To simulate practical scenario, our normal regions are false positives asserted by a CAD system from healthy cases.

RESULTS

The combination of all the descriptors yields better results than each feature set used alone, and the difference is statistically significant. Besides, the feature selection steps yields a statistically significant increase in the accuracy values for the three classifiers. Finally, the random forest achieves the highest accuracy (81.09%), outperforming the SVM classifier (80.01%)) and decision tree (79.12%), but the difference is not statistically significant.

CONCLUSIONS

The accuracy of discrimination between normal and abnormal ROIs in mammograms obtained with the proposed gray level texture features sets are encouraging and comparable to these obtained with multiresolution features. Combination of several features as well as feature selection steps improve the results. To improve false positives reduction in CAD systems for breast cancer diagnosis, these features could be combined with multiresolution features.

Assessment of global and local region-based bilateral mammographic feature asymmetry to predict short-term breast cancer risk

This study aims to develop and test a new imaging marker-based short-term breast cancer risk prediction model. An age-matched dataset of 566 screening mammography cases was used. All 'prior' images acquired in the two screening series were negative, while in the 'current' screening images, 283 cases were positive for cancer and 283 cases remained negative. For each case, two bilateral cranio-caudal view mammograms acquired from the 'prior' negative screenings were selected and processed by a computer-aided image processing scheme, which segmented the entire breast area into nine strip-based local regions, extracted the element regions using difference of Gaussian filters, and computed both global- and local-based bilateral asymmetrical image features. An initial feature pool included 190 features related to the spatial distribution and structural similarity of grayscale values, as well as of the magnitude and phase responses of multidirectional Gabor filters. Next, a short-term breast cancer risk prediction model based on a generalized linear model was built using an embedded stepwise regression analysis method to select features and a leave-one-case-out cross-validation method to predict the likelihood of each woman having image-detectable cancer in the next sequential mammography screening. The area under the receiver operating characteristic curve (AUC) values significantly increased from 0.5863  ±  0.0237 to 0.6870  ±  0.0220 when the model trained by the image features extracted from the global regions and by the features extracted from both the global and the matched local regions (p  =  0.0001). The odds ratio values monotonically increased from 1.00-8.11 with a significantly increasing trend in slope (p  =  0.0028) as the model-generated risk score increased. In addition, the AUC values were 0.6555  ±  0.0437, 0.6958  ±  0.0290, and 0.7054  ±  0.0529 for the three age groups of 37-49, 50-65, and 66-87 years old, respectively. AUC values of 0.6529  ±  0.1100, 0.6820  ±  0.0353, 0.6836  ±  0.0302 and 0.8043  ±  0.1067 were yielded for the four mammography density sub-groups (BIRADS from 1-4), respectively. This study demonstrated that bilateral asymmetry features extracted from local regions combined with the global region in bilateral negative mammograms could be used as a new imaging marker to assist in the prediction of short-term breast cancer risk.

Applying a new bilateral mammographic density segmentation method to improve accuracy of breast cancer risk prediction

PURPOSE

How to optimally detect bilateral mammographic asymmetry and improve risk prediction accuracy remains a difficult and unsolved issue. Our aim was to find an effective mammographic density segmentation method to improve accuracy of breast cancer risk prediction.

METHODS

A dataset including 168 negative mammography screening cases was used. We applied a mutual threshold to bilateral mammograms of left and right breasts to segment the dense breast regions. The mutual threshold was determined by the median grayscale value of all pixels in both left and right breast regions. For each case, we then computed three types of image features representing asymmetry, mean and the maximum of the image features, respectively. A two-stage classification scheme was developed to fuse the three types of features. The risk prediction performance was tested using a leave-one-case-out cross-validation method.

RESULTS

By using the new density segmentation method, the computed area under the receiver operating characteristic curve was 0.830 ± 0.033 and overall prediction accuracy was 81.0%, significantly higher than those of 0.633 ± 0.043 and 57.1% achieved by using the previous density segmentation method ([Formula: see text], t-test).

CONCLUSIONS

A new mammographic density segmentation method based on a bilateral mutual threshold can be used to more effectively detect bilateral mammographic density asymmetry and help significantly improve accuracy of near-term breast cancer risk prediction.

Computer-aided classification of mammographic masses using visually sensitive image features

PURPOSE

To develop a new computer-aided diagnosis (CAD) scheme that computes visually sensitive image features routinely used by radiologists to develop a machine learning classifier and distinguish between the malignant and benign breast masses detected from digital mammograms.

METHODS

An image dataset including 301 breast masses was retrospectively selected. From each segmented mass region, we computed image features that mimic five categories of visually sensitive features routinely used by radiologists in reading mammograms. We then selected five optimal features in the five feature categories and applied logistic regression models for classification. A new CAD interface was also designed to show lesion segmentation, computed feature values and classification score.

RESULTS

Areas under ROC curves (AUC) were 0.786±0.026 and 0.758±0.027 when to classify mass regions depicting on two view images, respectively. By fusing classification scores computed from two regions, AUC increased to 0.806±0.025.

CONCLUSION

This study demonstrated a new approach to develop CAD scheme based on 5 visually sensitive image features. Combining with a "visual aid" interface, CAD results may be much more easily explainable to the observers and increase their confidence to consider CAD generated classification results than using other conventional CAD approaches, which involve many complicated and visually insensitive texture features.

Developing a new case based computer-aided detection scheme and an adaptive cueing method to improve performance in detecting mammographic lesions

The purpose of this study is to evaluate a new method to improve performance of computer-aided detection (CAD) schemes of screening mammograms with two approaches. In the first approach, we developed a new case based CAD scheme using a set of optimally selected global mammographic density, texture, spiculation, and structural similarity features computed from all four full-field digital mammography images of the craniocaudal (CC) and mediolateral oblique (MLO) views by using a modified fast and accurate sequential floating forward selection feature selection algorithm. Selected features were then applied to a 'scoring fusion' artificial neural network classification scheme to produce a final case based risk score. In the second approach, we combined the case based risk score with the conventional lesion based scores of a conventional lesion based CAD scheme using a new adaptive cueing method that is integrated with the case based risk scores. We evaluated our methods using a ten-fold cross-validation scheme on 924 cases (476 cancer and 448 recalled or negative), whereby each case had all four images from the CC and MLO views. The area under the receiver operating characteristic curve was AUC  =  0.793  ±  0.015 and the odds ratio monotonically increased from 1 to 37.21 as CAD-generated case based detection scores increased. Using the new adaptive cueing method, the region based and case based sensitivities of the conventional CAD scheme at a false positive rate of 0.71 per image increased by 2.4% and 0.8%, respectively. The study demonstrated that supplementary information can be derived by computing global mammographic density image features to improve CAD-cueing performance on the suspicious mammographic lesions.

Association Between Changes in Mammographic Image Features and Risk for Near-Term Breast Cancer Development

The purpose of this study is to develop and test a new computerized model for predicting near-term breast cancer risk based on quantitative assessment of bilateral mammographic image feature variations in a series of negative full-field digital mammography (FFDM) images. The retrospective dataset included series of four sequential FFDM examinations of 335 women. The last examination in each series ("current") and the three most recent "prior" examinations were obtained. All "prior" examinations were interpreted as negative during the original clinical image reading, while in the "current" examinations 159 cancers were detected and pathologically verified and 176 cases remained cancer-free. From each image, we initially computed 158 mammographic density, structural similarity, and texture based image features. The absolute subtraction value between the left and right breasts was selected to represent each feature. We then built three support vector machine (SVM) based risk models, which were trained and tested using a leave-one-case-out based cross-validation method. The actual features used in each SVM model were selected using a nested stepwise regression analysis method. The computed areas under receiver operating characteristic curves monotonically increased from 0.666±0.029 to 0.730±0.027 as the time-lag between the "prior" (3 to 1) and "current" examinations decreases. The maximum adjusted odds ratios were 5.63, 7.43, and 11.1 for the three "prior" (3 to 1) sets of examinations, respectively. This study demonstrated a positive association between the risk scores generated by a bilateral mammographic feature difference based risk model and an increasing trend of the near-term risk for having mammography-detected breast cancer.

Assessment of performance and reproducibility of applying a content-based image retrieval scheme for classification of breast lesions

PURPOSE

To develop a new computer-aided diagnosis (CAD) scheme using a content-based image retrieval (CBIR) approach for classification between the malignant and benign breast lesions depicted on the digital mammograms and assess CAD performance and reproducibility.

METHODS

An image dataset including 820 regions of interest (ROIs) was used. Among them, 431 ROIs depict malignant lesions and 389 depict benign lesions. After applying an image preprocessing process to define the lesion center, two image features were computed from each ROI. The first feature is an average pixel value of a mapped region generated using a watershed algorithm. The second feature is an average pixel value difference between a ROI's center region and the rest of the image. A two-step CBIR approach uses these two features sequentially to search for ten most similar reference ROIs for each queried ROI. A similarity based classification score was then computed to predict the likelihood of the queried ROI depicting a malignant lesion. To assess the reproducibility of the CAD scheme, we selected another independent testing dataset of 100 ROIs. For each ROI in the testing dataset, we added four randomly queried lesion center pixels and examined the variation of the classification scores.

RESULTS

The area under the ROC curve (AUC) = 0.962 ± 0.006 was obtained when applying a leave-one-out validation method to 820 ROIs. Using the independent testing dataset, the initial AUC value was 0.832 ± 0.040, and using the median classification score of each ROI with five queried seeds, AUC value increased to 0.878 ± 0.035.

CONCLUSIONS

The authors demonstrated that (1) a simple and efficient CBIR scheme using two lesion density distribution related features achieved high performance in classifying breast lesions without actual lesion segmentation and (2) similar to the conventional CAD schemes using global optimization approaches, improving reproducibility is also one of the challenges in developing CAD schemes using a CBIR based regional optimization approach.

A new approach to develop computer-aided detection schemes of digital mammograms

The purpose of this study is to develop a new global mammographic image feature analysis based computer-aided detection (CAD) scheme and evaluate its performance in detecting positive screening mammography examinations. A dataset that includes images acquired from 1896 full-field digital mammography (FFDM) screening examinations was used in this study. Among them, 812 cases were positive for cancer and 1084 were negative or benign. After segmenting the breast area, a computerized scheme was applied to compute 92 global mammographic tissue density based features on each of four mammograms of the craniocaudal (CC) and mediolateral oblique (MLO) views. After adding three existing popular risk factors (woman's age, subjectively rated mammographic density, and family breast cancer history) into the initial feature pool, we applied a sequential forward floating selection feature selection algorithm to select relevant features from the bilateral CC and MLO view images separately. The selected CC and MLO view image features were used to train two artificial neural networks (ANNs). The results were then fused by a third ANN to build a two-stage classifier to predict the likelihood of the FFDM screening examination being positive. CAD performance was tested using a ten-fold cross-validation method. The computed area under the receiver operating characteristic curve was AUC = 0.779   ±   0.025 and the odds ratio monotonically increased from 1 to 31.55 as CAD-generated detection scores increased. The study demonstrated that this new global image feature based CAD scheme had a relatively higher discriminatory power to cue the FFDM examinations with high risk of being positive, which may provide a new CAD-cueing method to assist radiologists in reading and interpreting screening mammograms.

Assessment of a Four-View Mammographic Image Feature Based Fusion Model to Predict Near-Term Breast Cancer Risk

The purpose of this study was to develop and assess a new quantitative four-view mammographic image feature based fusion model to predict the near-term breast cancer risk of the individual women after a negative screening mammography examination of interest. The dataset included fully-anonymized mammograms acquired on 870 women with two sequential full-field digital mammography examinations. For each woman, the first "prior" examination in the series was interpreted as negative (not recalled) during the original image reading. In the second "current" examination, 430 women were diagnosed with pathology verified cancers and 440 remained negative ("cancer-free"). For each of four bilateral craniocaudal and mediolateral oblique view images of left and right breasts, we computed and analyzed eight groups of global mammographic texture and tissue density image features. A risk prediction model based on three artificial neural networks was developed to fuse image features computed from two bilateral views of four images. The risk model performance was tested using a ten-fold cross-validation method and a number of performance evaluation indices including the area under the receiver operating characteristic curve (AUC) and odds ratio (OR). The highest AUC = 0.725 ± 0.026 was obtained when the model was trained by gray-level run length statistics texture features computed on dense breast regions, which was significantly higher than the AUC values achieved using the model trained by only two bilateral one-view images (p < 0.02). The adjustable OR values monotonically increased from 1.0 to 11.8 as model-generated risk score increased. The regression analysis of OR values also showed a significant increase trend in slope (p < 0.01). As a result, this preliminary study demonstrated that a new four-view mammographic image feature based risk model could provide useful and supplementary image information to help predict the near-term breast cancer risk.

Point-of-care and point-of-procedure optical imaging technologies for primary care and global health

Leveraging advances in consumer electronics and wireless telecommunications, low-cost, portable optical imaging devices have the potential to improve screening and detection of disease at the point of care in primary health care settings in both low- and high-resource countries. Similarly, real-time optical imaging technologies can improve diagnosis and treatment at the point of procedure by circumventing the need for biopsy and analysis by expert pathologists, who are scarce in developing countries. Although many optical imaging technologies have been translated from bench to bedside, industry support is needed to commercialize and broadly disseminate these from the patient level to the population level to transform the standard of care. This review provides an overview of promising optical imaging technologies, the infrastructure needed to integrate them into widespread clinical use, and the challenges that must be addressed to harness the potential of these technologies to improve health care systems around the world.