Reproducibility of image analysis for breast ultrasound computer-aided diagnosis

Acoustical Imaging 03/2008; 29(1):397-402. DOI: 10.1007/978-1-4020-8823-0_55


We employ a Case-Based Reasoning approach to analyze breast masses in ultrasound and to classify them for level of suspicion for cancer following the ACR BI-RADS® protocol. Our computer-aided imaging system (Breast Companion®, BC) measures numeric features of the mass, determines Relative Similarity (RS) between the mass of interest and images in a database of masses with known findings and outcomes, then retrieves and displays the images of the most similar known masses instantaneously for the radiologist to review during interpretation. This study tested BC for reproducibility of performance in comparison to that of three radiologists under a variety of operating conditions. The long-term goal is to standardize diagnosis, reduce radiologist variability and reduce false positives.

Download full-text


Available from: Michael P Andre, Sep 29, 2015
19 Reads
  • [Show abstract] [Hide abstract]
    ABSTRACT: The purpose was to investigate the repeatability and bias of the output of two classifiers commonly used in computeraided diagnosis for the task of distinguishing benign from malignant lesions. Classifier training and testing were performed within a bootstrap approach using a dataset of 125 sonographic breast lesions (54 malignant, 71 benign). The classifiers investigated were linear discriminant analysis (LDA) and a Bayesian Neural Net (BNN) with 5 hidden units. Both used the same 4 input lesion features. The bootstrap .632plus area under the ROC curve (AUC) was used as a summary performance metric. On an individual case basis, the variability of the classifier output was used in a detailed performance evaluation of repeatability and bias. The LDA obtained an AUC value of 0.87 with 95% confidence interval [0.81; 0.92]. For the BNN, those values were 0.86 and [.76; .93], respectively. The classifier outputs for individual cases displayed better repeatability (less variability) for the LDA than for the BNN and for the LDA the maximum repeatability (lowest variability) lied in the middle of the range of possible outputs, while the BNN was least repeatable (highest variability) in this region. There was a small but significant systematic bias in the LDA output, however, while for the BNN the bias appeared to be weak. In summary, while ROC analysis suggested similar classifier performance, there were substantial differences in classifier behavior on a by-case basis. Knowledge of this behavior is crucial for successful translation and implementation of computer-aided diagnosis in clinical decision making.
    Proceedings of SPIE - The International Society for Optical Engineering 03/2010; DOI:10.1117/12.844889 · 0.20 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Our computer-aided diagnostic (CADx) tool uses advanced image processing and artificial intelligence to analyze findings on breast sonography images. The goal is to standardize reporting of such findings using well-defined descriptors and to improve accuracy and reproducibility of interpretation of breast ultrasound by radiologists. This study examined several factors that may impact accuracy and reproducibility of the CADx software, which proved to be highly accurate and stabile over several operating conditions. Keywords Breast cancer - Sonography - Computer-aided diagnosis - Image processing - Relative similarity - ROC analysis - Segmentation - Case-based reasoning
    Acoustical Imaging, 30 edited by Michael P Andre, Joie P Jones, Hua Lee, 01/2011: pages 3-10; Springer., ISBN: 978-90-481-3254-6
  • [Show abstract] [Hide abstract]
    ABSTRACT: We demonstrate that the coda of station-to-station Green's functions extracted from the ambient seismic field in southern California reach stability in the microseism band (5-10 s) after correlating six months of noise data. The coda stability makes it possible to retrieve Green's functions between stations that operate asynchronously through scattered waves as recorded by a network of fiducial stations. The Green's functions extracted from asynchronous and synchronous data have comparable quality as long as stable virtual coda are used, and both show good convergence to the Green's functions extracted from 1 year of seismic noise with ˜50 fiducial stations. This approach suggests that Green's functions can be extracted across seismic stations regardless of whether or not they are occupied simultaneously, which raises the prospect of a new mode for seismic experiments that seek to constrain Earth structure.
    Geophysical Research Letters 03/2012; 39(6):6301-. DOI:10.1029/2011GL050755 · 4.20 Impact Factor