Mammographic Screening: Fractal Dimensional Malignancy Identification *

Xiang Yang Ye & Andrew Reilly

The high incidence of breast cancer has lead to the recommendation that women receive yearly radiographic mammograms. The films are graded with high sensitivity in order to catch as many tumors as possible. Consequently there are many false positives; after a decade of screening half of all women will have at least one positive screening. False positives results are accompanied with needless anxiety and followup testing usually involving a biopsy. The recent replacement of film by digitized CCD cameras has made screening images routinely available for computer assisted analysis directed at enhancing screening’s diagnostic accuracy. We report on recent progress in utilizing fractal dimensional analysis (FDA) to differentiate benign and malignant mammograms. Current work is directed at computationally efficient methods for identifying potential tumor sites where FDA should be applied.