Background: There are some issues in screening for cancer, especially breast cancer, which are worthy of further study.
Methods: We reviewed some approaches to the following issues in breast cancer screening: absolute benefit, overdiagnosis, separation of effects of screening from effects of others on deaths from breast cancer over time and determination of which tumours benefit most from early detection. For separation of screening effects from other effects, we developed a simple analysis of survival by tumour size. For the other issues, we identified methods and results from the literature on the randomized trials of screening and on service screening programmes. We also reviewed screening issues pertaining to other selected cancers, including some cancers common or becoming common in Asia.
Results: Published results from the Swedish Two-County Trial showed that for 350 women screened for 10 years, one life would be saved. Results from service screening programmes in Florence and elsewhere suggested that overdiagnosis is a minor phenomenon and mainly confined to ductal carcinoma in situ. Data from before and after the inception of screening in East Anglia suggested that 40-60% of the recent improvement in survival of breast cancer cases was due to early detection. Published data from the Swedish Two-county Trial indicated that the majority of the benefit of breast screening derives from early detection of grades 2 and 3 ductal carcinoma.
Conclusions: Mammographic screening is effective in saving lives from breast cancer. Other cancers, which have strong evidence for a benefit of screening in terms of saving lives from cancer, are cervical and colorectal cancers. There is also evidence of a reduction in mortality associated with ultrasound screening for liver cancer in subjects at very high risk of the disease. There are a number of other cancers that are potential candidates for early detection. There is clearly an opportunity for saving lives from further development and implementation of cancer screening.