Search

Using a large set of ultrasound features does not necessarily ensure improved quantitative classification of breast tumors; rather, it often degrades the performance of a classifier. In this paper, we propose an effective feature reduction approach in the transform domain for improved multi-class classification of breast tumors. Feature transformation methods, such as empirical mode decomposition (EMD) and discrete wavelet transform (DWT), followed by a filter- or wrapper-based subset selection scheme are used to extract a set of non-redundant and more potential transform domain features through decorrelation of an optimally ordered sequence of N ultrasonic bi-modal (i.e., quantitative ultrasound and elastography) features. The proposed transform domain bi-modal reduced feature set with different conventional classifiers will classify 201 breast tumors into benign-malignant as well as BI-RADS ⩽ 3, 4, and 5 categories. For the latter case, an inadmissible error probability is defined for the subset selection using a wrapper/filter. The classifiers use train truth from histopathology/cytology for binary (i.e., benign-malignant) separation of tumors and then bi-modal BI-RADS scores from the radiologists for separating malignant tumors into BI-RADS category 4 and 5. A comparative performance analysis of several widely used conventional classifiers is also presented to assess their efficacy for the proposed transform domain reduced feature set for classification of breast tumors. The results show that our transform domain bimodal reduced feature set achieves improvement of 5.35%, 3.45%, and 3.98%, respectively, in sensitivity, specificity, and accuracy as compared to that of the original domain optimal feature set for benign-malignant classification of breast tumors. In quantitative classification of breast tumors into BI-RADS categories ⩽ 3, 4, and 5, the proposed transform domain reduced feature set attains improvement of 3.49%, 9.07%, and 3.06%, respectively, in likelihood of malignancy and 4.48% in inadmissible error probability compared to that of the original domain optimal subset. In summary, the construction of a transform domain reduced feature set by extracting complementary information from a large set of available bi-modal features and use of qualitative bi-modal BI-RADS can contribute to improved quantitative classification of breast tumors and thereby help reduce the number of unnecessary biopsies, securing a nearly minimum chance of a life-endangering diagnosis. [ABSTRACT FROM AUTHOR]

Multi-modal complementary information integrated noninvasive digital biomarkers are emerging with enormous potential for early detection of female breast cancer and reduction of mortality rate among women around the globe. The objective of this study is to design new quantitative ultrasound (QUS) biomarkers integrated from pathological information of tissue microstructure, conventional B-mode imaging-based acoustic and morphological features, and strain imaging-based tissue elasticity features for detecting cancer more effectively. We also develop simulation models to evaluate the estimation accuracy of effective scatterer diameter (ESD), a characteristic parameter of tissue pathology. The proposed biomarkers are designed by optimally integrating 2 QUS indirect pathology (i PATH) parameters with 27 QUS radiology and tissue elasticity (RATE) parameters. To determine effective biomarkers, several techniques are employed: a wrapper-based feature reduction technique in the original feature domain as well as in the empirical mode decomposition (EMD)-discrete wavelet transforms (DWT) domain, and a genetic algorithm-based feature weighting technique. Simulation phantoms determining the accuracy of the ESD estimation algorithm and explaining the increasing trend of ESD with the carcinogenic transformation of normal tissue are created in MATLAB and analyzed using Field II. The resulting average mean absolute percentage error of the ESD estimation algorithm on the simulation phantoms is 4.06%, proving its efficacy. On 139 in-vivo patient data, the three techniques yield accuracy, sensitivity, and specificity within the range of 97.84–99.28%, 97.84–100%, and 95.45–98.95%, respectively. The high values of these metrics indicate that the RATE- i PATH feature set can be used as non-invasive integrated biomarkers for computer-aided diagnosis of breast cancer. [ABSTRACT FROM AUTHOR]

Quantitative ultrasound (QUS) parameters are gaining attention recently as non-invasive biomarkers for soft tissue characterization. In this work, we use two QUS parameters, namely effective scatterer diameter (ESD) and mean scatterer spacing (MSS), for binary classification of breast lesions. To produce improved classification results, we propose a modified frequency-domain technique for ESD estimation of breast tissues from the diffuse component of backscattered radio-frequency (RF) data. Ensemble empirical mode decomposition (EEMD) is performed to separate the diffuse component from the coherent component of the backscattered RF data. A non-parametric Kolmogorov–Smirnov (K-S) test is employed for automatic selection of the intrinsic mode functions (IMFs). A novel multi-step system effect minimization process is also used. The ESD is estimated using a novel nearest neighborhood average regression line fitting (NNARLF) algorithm. Furthermore, we use an ameliorated EEMD domain autoregressive (AR) spectral estimation technique for MSS estimation. On using the ESD for binary classification of 139 lesions, we obtain high sensitivity, specificity, accuracy values of 95.45%, 95.79%, and 95.68%, respectively. The area under the receiver operating characteristics (ROC) curve is 0.95. On combining ESD with MSS we obtain even more improved sensitivity, specificity, and accuracy values of 97.73%, 95.79%, and 96.40%, respectively. The area under the ROC also increases to 0.97. This high classification performance demonstrates the potential of the use of these QUS parameters as non-invasive biomarkers for breast cancer detection. [ABSTRACT FROM AUTHOR]