Lead-time in the European Randomised Study of Screening for Prostate Cancer

Abstract

Background: Lead-time is defined as the time by which screening advances the diagnosis compared with absence of screening. A sufficiently long lead-time needs to be achieved so that cancer can be detected while still curable. A very short lead-time may indicate poor sensitivity of the screening test, while a very long lead-time suggests overdiagnosis.

Material and methods: In the first screening round, a total of 56,294 men aged 55-74 years were screened with serum prostate specific antigen (PSA) in five countries of the European Randomised Study of Screening for Prostate Cancer (ERSPC) with an overall detection rate (prevalence) of 2.8% (1972 prostate cancers). Prostate cancer incidence among 92,142 men randomly allocated to the control arm of the trial was also assessed. Lead-time was estimated as the time required to accumulate a similar cumulative risk of prostate cancer in the control arm to the detection rate in the intervention arm, i.e. from the ratio of detection rate (prevalence of screen-detected cases) and expected incidence (cumulative risk).

Results: Using a serum PSA cut-off of 4 ng/ml, the mean lead-time in the whole study population was estimated as 6.8 years (95% confidence interval (95% CI) 7.9-8.4). It was 8 years in The Netherlands, 6 in Sweden and Finland, 5 in Italy and 4 in Belgium. The mean lead-time was similar, 6-7 years, at ages 50-64 years, but close to 8 years among men aged 65-74 years. A lower PSA cut-off level of 3 ng/ml used in Sweden and The Netherlands prolonged the mean lead-time by approximately 1 year. Lead-time based on advanced prostate cancer only was slightly shorter, mean 5.3 years (95% CI 4.6-6.0). The lead-time for the second screening round was slightly shorter than that for the first (5.9, 95% CI 5.4-6.4), reflecting a similar relation between detection rate and control group incidence.

Conclusion: The lead-time for prostate cancer found in ERSPC substantially exceeded that found for breast, cervical and colorectal cancer screening. One round of prostate cancer screening can advance clinical diagnosis by 4-8 years. Overdiagnosis or detection of non-progressive tumours may contribute substantially to the lead-time.