Abstract:
Objective: To develop a system for the interactive exploration and
examination of histology-derived data associated with breast tumors,
which may be used to evaluate the histologic grade of the tumor.
Study Design: The system integrates pathologist-generated prognostic
data with 2D image analysis data, 2D digital tissue cross-sections and
annotations, 3D tumor reconstructions and
volumetric analysis, 3D spatial tumor display, and recorded prognostic
information from
available cases in the DUCOM tumor databank. The system consists of
three components: 1)
a user interface for applying two-dimensional (2D) image processing,
segmentation and
annotation to a digitized histology slide, 2) a distance field
interpolation method for contour-based three-dimensional (3D)
reconstruction of breast tumors, and volumetric model analysis
routines, 3) a web-based database management interface for
interactive data browsing and
searching, and multi-modality visualization.
Results: The system has been implemented and deployed with data from
36 breast cancer
cases, 7 of which have been reconstructed in 3D.
Conclusion: Interactive visual analytics technology may be used to
create an effective breast tumor evaluation system.