The radiologic diagnosis of pancreatic cancer. The effect of new imaging techniques.

The objective of this study was to evaluate how the introduction of radiologic studies affected the diagnostic workup for pancreatic cancer, from 1955 through 1979. For 961 patients diagnosed as having pancreatic cancer at three teaching hospitals, we reviewed medical records, autopsy reports, and death certificates for results from all radiologic studies, surgical and pathologic procedures, and for the final diagnosis. The number of radiologic studies per patient increased as new studies were introduced; 1.6 for 1955-1959, 3.5 for 1975-1979 (P less than 0.0001). Depending on the cutoff level chosen, the sensitivity of the overall radiologic diagnosis increased over time, 0.17-0.43 for 1955-1959, to 0.54-0.78 for 1975-1979; CT, US, and ERCP accounted for much of the increase. As newer radiologic studies are introduced, continued use of previously accepted studies should be carefully evaluated.