Between the years 1974-790, 10,400 mammographies were performed in 8,566 women, and 442 prevalent breast cancers were detected. Of these mammographies, 8,400 were performed in 6,556 women for diagnostic purposes, and 2,000 were part of a clinical-radiological screening study. The mammograms were classified according to the patterns described by Wolfe. One-third of our cancer patients presented with the"high risk" P2 and DY pattern; the same pattern was present in 27% of our normal control group. We conclude that patient management and follow-up cannot be based on mammographic pattern alone for this would miss two-thirds of breast cancer cases. The radiologist interpreting mammographies should nevertheless by especially alert in the presence of the high risk pattern.