Lead-time in prostate cancer screening (Finland) Anssi Auvinen 1,2, *, Liisa Ma¨ a¨ tta¨ nen 1,2 , Ulf-Ha˚kan Stenman 3 , Teuvo Tammela 4 , Sakari Rannikko 5 , Jussi Aro 6 , Harri Juusela 7 & Matti Hakama 1 1 University of Tampere, School of Public Health, Tampere, Finland; 2 Finnish Cancer Registry, Helsinki, Finland; 3 Helsinki University Hospital, Clinical Laboratory, Helsinki, Finland; 4 Tampere University Hospital, Department of Urology, Tampere, Finland; 5 Helsinki University Hospital, Department of Urology, Helsinki, Finland; 6 Helsinki City Hospital, Department of Surgery, Helsinki, Finland; 7 Jorvi Hospital, Department of Surgery, Espoo, Finland Received 27 June 2001; accepted in revised form 4 December 2001 Key words: epidemiology, prostate neoplasms, prostate-specific antigen, randomized trials, screening. Abstract Objective: Lead-time in prostate cancer screening was estimated using data from the Finnish randomized, population-based trial. Methods: Lead-time was defined as the duration of follow-up needed to accrue the same expected number of incident prostate cancer cases in the absence of screening as detected in the initial screening round. Expected numbers were calculated using an age-cohort model. Results: Based on findings among 10,000 men screened in 1996–1997 with 292 screen-detected cancers, lead-time was estimated as approximately 5–7 years, depending on the reference rates used. This corresponds to a mean duration of the detectable preclinical phase (DPCP) of 10–14 years, given that the cancers were detected on average at the midpoint of the DPCP. Conclusions: The findings suggest that a screening interval substantially longer than the 2 years generally used for mammography screening is unlikely to cause a substantial loss of sensitivity. A long screening interval is further justified in order to diminish the extent of overdiagnosis, until mortality effects can be evaluated. Introduction Prostate cancer is a major public health issue, being the most common malignancy after skin cancer among men in many industrialized countries [1]. As its etiology is poorly understood, hopes for prostate cancer control rely largely on screening for the disease [2]. Currently, it remains unclear if a mortality reduction can be achieved by screening, but large multi-center randomized trials are ongoing to assess this possibility [3]. No valid results pertaining to a possible reduction in prostate cancer mortality are yet available, but preliminary evaluation has to rely on ecological assessment, such as time trends, and on process measures, such as sensitivity. Lead-time is one of the process measures that are necessary, but not sufficient, prerequisites for effective screening. Lead-time is defined as the time interval by which the diagnosis of disease is advanced by screening, i.e. the time interval between detection by screening and expected detection in the absence of screening [4]. Lead- time depends both on the natural history of the disease and on the screening program, e.g. sensitivity and re- screening interval. Lead-time varies between individuals and cannot be directly observed for a single case (without repeated observations in the absence of intervention), but its distribution at the population level can be estimated. The duration of the detectable, preclinical phase (DPCP) of disease (also called sojourn time) is a concept closely related to lead-time, which is used to assess DPCP [5]. DPCP gives the upper limit of the lead-time, i.e. the maximum lead-time that would be achieved if screening was performed at the beginning of the DPCP. In practice, however, screening takes place at random * Address correspondence to: Professor Anssi Auvinen, Tampere School of Public Health, University of Tampere, FIN-33014 Tampere, Finland, Ph.: +358-3-2156 883; Fax: +358-3-2156 057; Email: anssi. auvinen@uta.fi Cancer Causes and Control 13: 279–285, 2002. 279 Ó 2002 Kluwer Academic Publishers. Printed in the Netherlands.