ABSTRACT
A measurement technique has been developed for noninvasive breast cancer detection. The process involves the use of close-range stereophotogrammetry as a data acquisition device for the determination of breast surface concavities. We report the methodology used to detect these surface depressions, the rationale for the study, and our preliminary findings.
Subject(s)
Breast Neoplasms/diagnosis , Image Processing, Computer-Assisted , Photogrammetry/methods , Algorithms , Female , Humans , Mathematics , Models, Anatomic , Software DesignABSTRACT
A measurement technique has been developed for application in the area of noninvasive breast cancer detection. The measurement process involves the use of close-range stereophotogrammetry as a data acquisition device necessary for determination of breast volume and volume distribution. This report details the methodology used to acquire and analyze stereopair photographs necessary to document the validity and reliability of this application. The volume of a test object was determined by both water displacement and stereophotogrammetric analysis to estimate the precision of the proposed methodology. Additionally, the reliability component of the study was documented by analyzing variability of coordinates representing a series of locations marked on the surface of an irregularly shaped object. Both tests confirm that this stereometric analysis is a reliable and valid method of measurement and may be well suited for further development in the field of breast cancer detection.
Subject(s)
Biometry , Breast/anatomy & histology , Photogrammetry/methods , Photography/methods , Biomedical Engineering , Breast Neoplasms/diagnosis , Female , Humans , Models, AnatomicABSTRACT
Twelve female patients participated in a study designed to employ computer-assisted biostereometric analysis for the detection of breast masses. All breast masses were previously documented by physical examination and followed by xeromammography, stereophotography, and histopathologic confirmation of tumor type. "Contour mammograms" were produced from the biostereometric photographs. These data were analyzed first for tumor detection and location by visual inspection. A second analysis employed a computed algorithm designed to locate and measure surface aberrations that suggest the possibility of underlying breast tumor. Visual analysis yielded the exact location of breast tumors in eight of the ten malignancies; computer analysis exactly located nine of the ten malignancies. In the computer analysis, one malignancy not exactly located by quadrant was, nevertheless, located in the correct breast. The results of the study suggest that the biostereometric process may have future use in screening or prescreening procedures for breast cancer detection. It is noninvasive, applicable to large numbers of women and with suitable refinements, and capable of being fully automated.
