Evaluation of computed tomography in pancreatic cancer

A study of the clinical efficacy of computed body tomography (CBT) is ongoing at the Massachusetts General Hospital. This study attempts to measure the contribution of CBT to diagnostic understanding, the use of other diagnostic tests, and the choice of therapy. During the first nineteen months of the study, 185 patients were examined on the pancreas protocol, a number approximating 30 per cent of all patients in the study. Physicians have some suspicion of pancreatic cancer in nearly all patients examined on the pancreas protocol. CBT improved the physician's diagnostic understanding in 54 per cent of patients examined on the pancreas protocol and contributed to a change in therapy in 18 per cent. CBT also contributed to a net reduction of 54 per cent in the number of angiograms performed compared to the number planned prior to CBT examination. In a subset of seventy-three patients, a definitive diagnosis of pancreatic cancer, inflammation, or normal pancreas was reached in each case. For patients with inflammation, CBT was 1.5 times as likely to increase as to decrease the physician's estimate of the correct diagnosis; for cancer patients CBT was 3.5 times as likely to aid as to mislead the physician; for patients with a normal pancreas, CBT was 6 times as likely to help as to hinder correct diagnosis. CBT can be a valuable aid to the diagnosis and treatment of patients suspected of having pancreatic cancer, but the physician should interpret its results in the light of all available clinical information.