Science Spotlight

Harm-benefit tradeoffs for prostate cancer screening in high-risk subgroups

Biostatistics & Biomathematics (Public Health Sciences Division)

Roughly 14% of American men will be diagnosed with prostate cancer during the course of their lifetimes. However, many prostate cancers never become large enough or troublesome enough to be diagnosed clinically. Of those diagnosed, only about 18% will be fatal.

The most common screening measure for prostate cancer is the prostate-specific antigen (PSA) test. This test measures the amount of PSA, a protein produced by both cancerous and non-cancerous tissue in the prostate, in the blood. The PSA test can lead to an earlier diagnosis and treatment and has the potential to improve survival. But when screening/biopsy finds cancers that would never cause symptoms during the patient’s lifetime, it results in over-diagnosis, and subsequent treatment is unnecessary and may cause harm.

Most clinical recommendations about prostate cancer screening are for the average-risk population. Some guidelines suggest starting earlier for African American men who have considerably higher incidence and mortality or for men with a family history of prostate cancer, but few specifics are given. In general, it is not recognized that whether and how to screen more intensively in these populations depends on exactly how their risk is modified relative to the average-risk population.

To investigate the connection between increased risk of disease incidence or mortality and preferred screening strategies, Roman Gulati and Ruth Etzioni (two statisticians in the Public Health Sciences Division) teamed up with medical oncologists at the center. They adapted an existing model of prostate cancer natural history that had been developed for the general US population and used to study benefits (e.g. lives saved) and harms (e.g. over-diagnosis) for a range of possible screening strategies. They used the model as a virtual laboratory to explore how benefit-harm tradeoffs change in populations with early onset/higher prevalence of subclinical disease or faster progression a lethal phenotype. The results of this research were recently published in Cancer Epidemiology, Biomarkers & Prevention.

The investigators conducted computer simulations of subgroups with average and hypothetical increased risk(s) of 1) onset of latent disease, 2) progression, and/or 3) cancer-specific death following diagnosis. For each subgroup, they predicted lifetime probabilities of over-diagnosis and life saved under more and less intensive PSA screening strategies. To investigate how risk might be increased for men with germline BRCA mutations, a recently-identified high-risk subgroup, they estimated the risk of latent onset corresponding to BRCA status using published results from the first screening round of the Identification of Men with a genetic predisposition to ProstAte Cancer: Targeted screening in BRCA1/2 mutation carriers and controls (IMPACT) study.

“Overall, we found that different types of increased risks can have different implications for preferred screening strategies”, concludes Mr. Gulati. “For example, men with faster progression are more likely to benefit from an efficacious screening test, and more frequent screening relative to the general population may be appropriate. In contrast, men with increased risk of developing preclinical disease (proportional to the general population) face higher chances of both benefit and harm by screening due to the larger pool of latent cancers at every age; whether to start screening at an earlier age hinges on the magnitude of increased risk and the associated benefit-harm tradeoffs deemed acceptable.”

In addition, their analysis suggests screening BRCA2 mutation carriers with a lower PSA threshold for biopsy is unlikely to improve outcomes under the assumptions of their model, which included similar PSA growth curves for BRCA2 mutation carriers and non-carriers. 

Graphs showing benefits and harms of 24 possible screening strategies in men with average and increased risks of prostate cancer.
The figure shows benefits and harms of 24 possible screening strategies in men with average risks and in men with hypothetical increased risks of onset of latent disease, diagnosis, and death obtained by applying a hazard ratio of 2 relative to the average-risk population for each risk. The projected outcomes show, for example, that different types of screening strategies achieve fewer than 5 over-diagnoses per life saved in men with increase risks compared to men with average risks. Image provided by Mr. Roman Gulati

“We concluded that screening guidelines for men with increased risks should consider the different types of the increased risks and their magnitudes”, Mr. Gulati notes. “Are they more likely to develop cancer earlier? How much earlier? Does it progress faster? How much faster? Further, while an efficacious screening test is more likely to benefit men with increased risks, it does not always improve benefit-harm tradeoffs.”

What’s in store next? According to Mr. Gulati, “The explorations in this study were mainly hypothetical and the conceptual framework was generic. Additional research is needed to estimate the types and magnitudes of increased risks in African American men or men with inherited genetic mutations in order study expected benefits and harms of targeted screening approaches in these subgroups with greater precision.”

Funding for this study was provided by the National Cancer Institute and the 2015 Prostate Cancer Foundation Young Investigator Award.


Gulati R, Cheng HH, Lange PH, Nelson PS, Etzioni R. 2016. Screening men at increased risk for prostate cancer diagnosis: Model estimates of benefits and harms. Cancer Epidemiology, Biomarkers & Prevention. cebp-0434.