Susan G Komen  
I've Been Diagnosed With Breast Cancer Someone I Know Was Diagnosed Share Your Story Join Us And Stay Informed Donate To End Breast Cancer
    Home > Research & Grants > Grants Program > Research Grants > Research Grants Awarded > Abstract

    Research Grants Awarded

    Biplane Correlation Imaging (BCI) for early detection of breast cancer

    Study Section:
    Postdoctoral Fellowship

    Scientific Abstract:
    The purpose of this research is to investigate the feasibility of using the Computer Aided Detection (CAD) with pair projection images to improve the detection of cancer in mammography images. It uses an imaging procedure called Bi-plane Correlation Imaging (BCI) in which digital radiographic images of the breast are acquired within a short interval of time from slightly different angles. Angular information is used to minimize the fundamental limiting factor imposed by anatomical noise on detection of lesions by identifying and positively reinforcing the lesion signals between different projections. It is anticipated that this method will overcome some of the most fundamental problems limiting the sensitivity and specificity of CAD in breast imaging. An existing prototype Full Field Digital Mammography (FFDM) mammography device capable of angular movement of the tube will be used. The technique will be optimized and evaluated using human subject data. Digital radiographic images of the breast at low exposure are acquired from different angulations. Angular information is used to minimize the fundamental limiting factor imposed by anatomical noise on detection of lesions by identifying and positively reinforcing the lesion signals between different projections . The overall exposure to the subjects will be less than that of conventional mammographic image acquisitions. Images will be processed in two stages. In stage one, the CAD routine uses morphological and textural features of the images to detect possible breast masses. Stage two includes BCI processing in which the results of CAD on a single projection are examined in terms of their geometrical correlation with a second projection, based on the predetermined shift of possible mass locations in that projection. The suspect masses with a geometrical correlation that coincides with the known location of the lesions in the second projection are scored as true positive (TP); otherwise they are scored as false positive (FP). The CAD-BCI method is a novel method that is applicable of 3-D breast radiography, and extracts morphological and textural features of x-ray images for the detection of breast masses. CAD-BCI has the potential for translation from a state-of-the art technology to clinic and improves breast cancer detection.

    Lay Abstract:
    One of the major deficiencies of mammography is its limitation in rendering masses and microcalcifications hidden in dense fibroglandular tissue due to the effect of the overlapping structures in the projection x-ray images. The normal tissue structure, so called anatomical noise, can prevent radiologists from seeing the relevant structural changes of tissue. To partially overcome this limitation, screening mammography usually involves acquiring two views of each breast: Cranio-Caudal (CC), and Mediolateral-Oblique (MLO). Radiologists compare mammograms at these views to confirm true lesions and to reduce false positives (FP). However , since the two views require two separate compressions, the image data from two views cannot be directly compared due to varying amount of tissue distortion. Furthermore the procedure involves compressing the breast twice with associate pain and comfort. The purpose of this project is to investigate the feasibility of an imaging procedure, namely, Bi-plane Correlation Imaging ( BCI ), in which digital radiographic images of the breast at low exposure is acquired from different angulations during a single low compression. Angular information is used to minimize the fundamental limiting factor imposed by anatomical noise on detection of lesions by identifying and positively reinforcing the lesion signals between different projections . Images are processed in two stages. In stage one, a robust Computer Aided Detection (CAD) routine is used to detect all possible breast masses. Stage two includes BCI processing in which the results of CAD on a single projection are examined in terms of their geometrical correlation with results in a second projection, based on the predetermined shift of possible mass locations in that projection. The suspect masses with a geometrical correlation that coincides with the known location of the lesions will be scored as true positive (TP) nodules; otherwise they will be scored as FP nodules. The CAD-BCI method is a novel method that utilizes three dimensional breast data, extracts morphological, and textural features of x-ray images for improved detection of breast masses . The spatial correlation information available from the stereo views of the lesions will also facilitate screening at lower total dose levels than those used in conventional mammography screening. CAD-BCI has the potential for translation from a state-of-the art technology to clinic and improves breast cancer detection .