TARGETED SCREENING FOR CANCER: TAILORING POLICY TO BIOLOGICAL MECHANISM

   Purpose: To investigate how screening policies should be designed for population subgroups with increased risk of cancer incidence and mortality.

   Methods: We construct a modeling framework that accommodates different mechanisms of increased risk, including increased risk of disease onset, progression, and/or cancer-specific death. In each increased-risk setting, we use computer simulation to predict harm-benefit tradeoffs of more versus less intensive screening strategies. We demonstrate the utility of the framework using the example of prostate-specific antigen (PSA) screening for prostate cancer and project lifetime probabilities of overdiagnosis and life saved under 24 strategies. The ratio of these probabilities—the number needed to detect (NND) to prevent 1 prostate cancer death—is used as a measure of harm-benefit tradeoff. Lessons learned are applied to make recommendations for BRCA2 mutation carriers based on data from the IMPACT study.

   Results: Screening is projected to save more lives among subgroups with increased risk, but whether more intensive screening improves harm-benefit tradeoffs depends on the mechanism of increased risk (see Figure). IMPACT data were consistent with increased risks of onset among BRCA2 mutation carriers (HR=1.81, 95% CI 1.14–2.78). This suggests screening BRCA2 mutation carriers earlier and potentially more frequently, but a lower PSA threshold for biopsy is unlikely to improve outcomes if PSA growth is similar across subgroups.

   Figure: Projected lifetime probabilities of overdiagnosis vs life saved in subgroups with average (left) or increased (right) risks of latent onset, progression, and cancer death under 24 screening strategies. Using an NND of 5 (i.e., 5 overdiagnoses per life saved)  as a threshold for preferred strategies, we find that more strategies achieve NND≤5 in the subgroup with increased risks than in the subgroup with average risks, but a lower PSA threshold for biopsy increases overdiagnosis with few additional lives saved.

   Conclusions: Policies for targeted screening should consider the mechanism inducing the increased risk. While the benefit of cancer screening may be greater in individuals with increased risk, more intensive screening is not always optimal.