Kestrel is developing a commercial computer-based grading system for 2D and stereo visualization and analysis of digital fundus images with a user- selectable level of automation geared toward clinical application and use in epidemiological studies of ARMD.
The specific aims of this project are to develop and validate a commercial-quality Stereo ARMD grading system (ARMD GS) and to perform a longitudinal study that will demonstrate clinical and research capabilities not available through current light-box based methods. The ability to display digital images in a familiar 3D format allows the advantages of digital technology to greatly enhance traditional methods without sacrificing the benefits of manual grading. Kestrel will enhance the Phase I Stereo ARMD GS through an evolutionary path that provides improvements in stages. Initial enhancement will include automatic registration of time series images to facilitate longitudinal study and user-directed semi-automatic segmentation of lesions. The final stage will automatically segment the entire image at once without user direction. The various levels of system automation will allow the user a choice as to the most appropriate tool for a given application. The final system will be demonstrated and validated through a retrospective longitudinal study of 40 subjects.
A market exists for computer-aided grading system at the major epidemiological fundus grading centers throughout the world. As a screening and surveillance system, the 2,500 retail eyecare centers would benefit economically from the proposed system. Kestrel's system will streamline the grading centers' process, thus enabling the centers to keep pace with the growing research requirements in ophthalmology. Eyecare centers could improve their capability for screening for ARMD, diabetic retinopathy, and other diseases presenting in the reitna. At present, we initiated discussions for beta testing at sites such as the Joslin Eye Center, the University of Wisconsin, and Scheie Eye Center. The estimated market size for medical image management hardware and software is 2.52 billion dollars. Potential market share for computer-assisted stereo system is estimated at 50 million dollars over a ten year period. The application of the stereo retinal-image visualization and evaluation system extends to other retinal diseases, which require stereo viewing.
