Mortality reduction and cumulative excess incidence (CEI) in the prostate-specific antigen (PSA) screening era

Abstract

The extent to which PSA screening is related to prostate cancer mortality reduction in the United States (US) is controversial. US Surveillance, Epidemiology, and End Results Program (SEER) data from 1980 to 2016 were examined to assess the relationship between prostate cancer mortality and cumulative excess incidence (CEI) in the PSA screening era and to clarify the impact of race on this relationship. CEI was considered as a surrogate for the intensity of prostate cancer screening with PSA testing and subsequent biopsy as appropriate. Data from 163,982,733 person-years diagnosed with 544,058 prostate cancers (9 registries, 9% of US population) were examined. Strong inverse linear relationships were noted between CEI and prostate cancer mortality, and 317,356 prostate cancer deaths were avoided. Eight regions of the US demonstrated prostate cancer mortality reduction of 46.0-63.7%. On a per population basis, the lives of more black men than white men were saved in three of four registries with sufficient black populations for comparison. Factor(s) independent of CEI (potential effects of treatment advances) explained 14.6% of the mortality benefit (p-value = 0.3357) while there was a significant main effect of CEI (effect = -0.0064; CI: [-0.0088, -0.0040]; p-value < 0.0001). Therefore, there is a strong relationship between CEI and prostate cancer mortality reduction that was not related to factors independent of screening utilization. Minority populations have experienced large mortality reductions in the context of PSA mass utilization.

Figure 1

Derivation of cumulative excess incidence (CEI). The prostate cancer incidence for each year from 1988 to 2016 for a single SEER registry is depicted (A). The incidence for these years is divided into baseline incidence and excess incidence. Baseline incidence is the average of annual incidence values from 1980 to 1987 (orange horizontal line), and any incidence above this baseline is considered excess incidence for that year (B). The CEI for the year 2016 is depicted as the summation (shading) of excess incidence for 1988 to 2016 (C). The CEI is plotted over time from 1988 to 2016 and constitutes the summing of excess incidence from 1988 to each year from 1988 to 2016 (D).

Figure 2

Prostate cancer mortality and CEI by registry. United States state and city population prostate cancer mortality is plotted as a function of prostate cancer CEI from 1988 to 2016.

Figure 3

Prostate cancer mortality and CEI stratified by race. Prostate cancer mortality is plotted as a function of prostate cancer CEI stratified by race in more diverse US populations (> 5% black).

Figure 4

CEI over time stratified by race. Prostate cancer CEI is plotted over time in more diverse US populations (> 5% black).