Diagnostic Performance of a Computer-Aided Image Analysis System For Breast Ultrasound

Ultrasonic Diagnosis of Breast Nodules Using Modified Faster R-CNN

Breast cancer has become the biggest threat to female health. Ultrasonic diagnosis of breast cancer based on artificial intelligence is basically a classification of benign and malignant tumors, which does not meet clinical demand. Besides, the current target detection method performs poorly in detecting small lesions, while it is clinically required to detect nodules below 2 mm. The objective of this study is to (a) propose a diagnostic method based on Breast Imaging Reporting and Data System (BI-RADS) and (b) increase its detectability of small lesions. We modified the framework of Faster R-CNN (Faster Region-based Convolutional Neural Network) by introducing multi-scale feature extraction and multi-resolution candidate bound extraction into the network. Then, it was trained using 852 images of BI-RADS C2, 739 images of C3, and 1662 images of malignancy (BI-RADS 4a/4b/4c/5/6). We compared our model with unmodified Faster R-CNN and YOLO v3 (You Only Look Once v3). The mean average precision (mAP) is significantly increased to 0.913, while its average detection speed is slightly declined to 4.11 FPS (frames per second). Meanwhile, its detectivity of small lesions is effectively improved. Moreover, we also tentatively applied our model on video sequences and got satisfactory results. We modified Faster R-CNN and trained it partly based on BI-RADS. Its detectability of lesions, as well as small nodules, was significantly improved. In view of wide coverage of dataset and satisfactory test results, our method can basically meet clinical needs.

Computer-Aided Diagnosis for Breast Ultrasound Using Computerized BI-RADS Features and Machine Learning Methods

This work identifies effective computable features from the Breast Imaging Reporting and Data System (BI-RADS), to develop a computer-aided diagnosis (CAD) system for breast ultrasound. Computerized features corresponding to ultrasound BI-RADs categories were designed and tested using a database of 283 pathology-proven benign and malignant lesions. Features were selected based on classification performance using a "bottom-up" approach for different machine learning methods, including decision tree, artificial neural network, random forest and support vector machine. Using 10-fold cross-validation on the database of 283 cases, the highest area under the receiver operating characteristic (ROC) curve (AUC) was 0.84 from a support vector machine with 77.7% overall accuracy; the highest overall accuracy, 78.5%, was from a random forest with the AUC 0.83. Lesion margin and orientation were optimum features common to all of the different machine learning methods. These features can be used in CAD systems to help distinguish benign from worrisome lesions.

Ultrasonic multi-feature analysis procedure for computer-aided diagnosis of solid breast lesions

We have developed quantitative descriptors to provide an objective means of noninvasive identification of cancerous breast lesions. These descriptors include quantitative acoustic features assessed using spectrum analysis of ultrasonic radiofrequency (rf) echo signals and morphometric properties related to lesion shape. Acoustic features include measures of echogenicity, heterogeneity and shadowing, computed by generating spectral-parameter images of the lesion and surrounding tissue. Spectral-parameter values are derived from rf echo signals at each pixel using a sliding-window Fourier analysis. We derive quantitative acoustic features from spectral-parameter maps of the lesion and adjacent areas. We quantify morphometric features by geometric and fractal analysis of traced lesion boundaries. Initial results on biopsy-proven cases show that although a single parameter cannot reliably discriminate cancerous from noncancerous breast lesions, multi-feature analysis provides excellent discrimination for this data set. We have processed data for 130 biopsy-proven patients, acquired during routine ultrasonic examinations at three clinical sites and produced an area under the receiver-operating-characteristics (ROC) curve of 0.947 +/- 0.045. Among the quantitative descriptors, lesion-margin definition, spiculation and border irregularity are the most useful; some additional morphometric features (such as border irregularity) also are particularly effective in lesion classification. Our findings are consistent with many of the BI-RADS (Breast Imaging Reporting and Data System) breast-lesion-classification criteria in use today.