Automated Pap Screen with HPV and Multispectral Imaging

DESCRIPTION (provided by applicant): The standard of care for cervical cancer screening is the Papanicolaou (Pap) smear. Manual review of smears is extremely labor intensive, and present commercial automated systems have significant limitations. We propose to address these deficiencies with a novel approach to automated Pap screening that uses high-risk HPV immunofluorescence labeling together with Pap staining as well as multispectral imaging and texture analysis. The HPV fluorescence signal will be used for a low magnification survey of the slide. Regions that show fluorescence will then be examined at high magnification using multispectral imaging. In this project we will: a) develop protocols to detect an HPV-16 gene product using a monoclonal antibody attached to a fluorescent probe in a slide that also contains Pap stain; b) develop an acousto-optic tunable filter based automated multispectral imaging station; c) develop analysis and classification algorithms to use this multispectral data to distinguish between cells that require clinical follow-up and those that do not. Validation of the technique will be performed using a set of HSIL cell samples together with a set of normal cervical cell samples as a control. Our goal is to obtain 100% sensitivity and 80% specificity on this initial study group.