Previous prostate cancer incidence projections estimated tha
Previous prostate cancer incidence projections estimated that the incidence rates will remain stable from the period of 2003–2007 to 2028–2032 at 123.3/100,000 . However, these projections are based on a long-term projection of constant rates, since the authors of the previous projections were unsatisfied with the extreme increase in incidence projected using the Nordpred model in their analysis. Our estimates suggest that after a slight decrease in the incidence rate of prostate cancer up to 2013, the rate will stabilize. The peaks in prostate cancer incidence in 1993 and 2001 can be explained by the two waves of intensified prostate-specific antigen (PSA) screening . Population-based PSA screening is no longer recommended in Canada .
As with any long-term projections, projecting cancer incidence to 2042 involves uncertainty and should be interpreted with caution. To reduce the uncertainty, we used high quality cancer incidence data from the CCR for 1992–2012, which misses very few incident cancer cases in Canada (>99% case ascertainment) . However, data for the province of Quebec for the years 2011 and 2012 were carried forward by Statistics Canada, since data after 2010 are not available for Quebec. Ethnicity is not included in incidence data released by the CCR, to maintain the confidentiality of cases. The projections also depend on the population projections, which are based on assumptions about fertility, mortality and AR-13324 at both the national and provincial levels.
Cancer incidence projections from statistical models rely on the assumption that past trends in the data will continue into the future, which further highlights the importance of continuously updating projections. To reduce some of the errors associated with failing to account for future changes in risk behaviours and screening practices, we used expert opinion when the statistical model produced implausible results that were possibly driven by artifacts in data collection (e.g. changes in registration practices, over diagnosis due to screening, etc.). We, as content experts in cancer epidemiology, reviewed the projections and evaluated the findings independently of goodness-of-fit, to increase the plausibility of the projections. To further validate our results, we compared our projections for 2015 with data released from the CCR after our analyses were completed. A comparison of our projections and actual incident frequencies is presented in Supplementary Table 1. In general, our projections are within 1,500 cases of the actual number of incident cases in 2015. However, our projected incidence estimates for prostate cancer was over 8,000 cases higher than the actual number of cases diagnosed in 2015. This difference was also seen for the previous Canadian estimates  and is likely due to past spikes in incidence caused by PSA testing, as previously mentioned. Although radiometric time spikes were taken into consideration, the residual effects of the PSA testing is leading to an overestimation of future prostate cancer. These discrepancies show the potential impact of a change in screening practice or technology on future cancer rates. Our estimates assume no additional changes in these practices which may be unreasonable, given recent technologies focus on screening and early detection of pre-cancerous lesions.
Conclusions These estimates were used in our larger ComPARe study , in which we were estimating the future burden of cancer attributable to various lifestyle, environmental and infection risk factors. In addition, our projected substantial increase in cancer cases at the top five sites by 2042 highlights the importance of and need for cancer control strategies and prevention programs aimed at reducing exposure to cancer risk factors. Planning for the future of cancer control in Canada needs to account for the increasing numbers of cancer patients expected each year.