Knowledge-guided, fully automated image analytic procedures will be applied, and further developed for the extraction of diagnostic and prognostic information from histopathologic sections. It is proposed to develop knowledge files for the grading of solar lesions, for the analysis of prostatic intraepithelial neoplastic lesions (PIN), for benign proliferative epithelial lesions of the breast, and for kidney tumors. Quantitative progression indices will be derived from histometric measurements. These may serve to identify patients at high risk to develop infiltrating disease, to measure rate of lesion progression, and to allow a numeric assessment of the efficacy of chemopreventive intervention. Knowledge files are under development for a quantitative measurement of the vascularization around PIN lesions. For nuclei, lesions and patients, novel methodology is proposed to characterize these entities by identification, rather than by mere classification. This will allow a significantly more precise characterization of the nuceli in a lesion and of the state of lesion progression. The identification methods will be integrated into the current diagnostic decision support system, and be given capabilities to handle missing data, contradictory evidence, atypical diagnostic clue expression. This capability relies on automated reasoning will be developed, and the methodology will be adapted for used in histopathologic diagnosis.

Agency
National Institute of Health (NIH)
Institute
National Cancer Institute (NCI)
Type
Research Project (R01)
Project #
2R01CA053877-08
Application #
2389351
Study Section
Special Emphasis Panel (ZRG7-SSS-X (78))
Program Officer
Aamodt, Roger L
Project Start
1991-03-01
Project End
2000-12-31
Budget Start
1998-01-01
Budget End
1998-12-31
Support Year
8
Fiscal Year
1998
Total Cost
Indirect Cost
Name
University of Arizona
Department
Type
Organized Research Units
DUNS #
City
Tucson
State
AZ
Country
United States
Zip Code
85721