Autopsy and cause of death in randomized mammography studies

Four randomized Swedish studies on the usefulness of mammography in screening for breast cancer have been published. The results have shown a variable but positive effect of screening in women more than 50 years of age but the effect is questionable before this age. The protocol and end points however differ between the studies and in order to summarize the results of the present studies, an unbiased, blinded end-point committee was created. The results of this study have been published recently. The present communication only concerns the role of autopsy in this study. The deaths of 1367 patients with a diagnosis of breast cancer have been studied. The percentage of autopsy varied from 21 to 74% between the four studies. When autopsy was performed, 58% were found to have died from breast cancer, 13% from other malignancies and the rest from other diseases. In cases with no autopsy, 73% were considered to have died of breast cancer and 9% of other malignancies. Deaths due to other cancers included a surprising number of cancers of the stomach, gallbladder and pancreas, two to three times higher than expected.