New PSA-Based Prognostic Model Improves Prostate Cancer Risk Assessment

Screening relies on blood levels of prostate-specific antigen (PSA), but translating those values into long-term risk is challenging. Although an estimated 10 million PSA tests are performed each year, few tools help clinicians interpret results or guide next steps. Existing calculators are less accurate or rely on biopsy-based testing, which requires tissue sampling and additional processing time. Researchers now present a prognostic model that uses PSA results to estimate prostate cancer-specific mortality and support shared decision-making.

Developed at the University of Michigan (Ann Arbor, MI, USA), the model combines PSA scores with clinical and demographic factors to estimate long-term risk of prostate cancer-specific mortality and identify patients likely to benefit from treatment. In earlier work, the researchers showed that PSA scores can influence both clinician and patient behavior, often leading to biopsy referrals even when the risk of harm is low. By refining risk assessment, the new model aims to better target referrals and focus additional screening and treatment on patients most likely to benefit.

The investigators built the model using data from the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial, which enrolled more than 33,000 participants aged 55 to 74 years from 1993 to 2001. The researchers also took family history of prostate cancer, race, age, body mass index, smoking status and a history of hypertension, diabetes or stroke into account. After building the model, they tested it using PSA scores from more than 200,000 patients who received care in the Veterans Affairs Healthcare System in the same age range from 2002 to 2006. The findings were reported in a study published in Annals of Internal Medicine on January 13. 

“Current tools don’t take into account how long someone may live or the benefit a patient may receive from treatment,” said Kristian Stensland, MD, MPH, MS, assistant professor of urology at the University of Michigan. “Our model is the first to incorporate all these factors and help people understand whether they need further screening or treatment.”

The researchers are now working to implement their model in clinical settings. “It is important to remember that we created and tested the model using data from two decades ago and a lot has changed since then,” said Stensland. “Even though prostate cancer treatment is different now, our model improves on previous tools and can be used to decide how we do PSA screens.”

Related Links
University of Michigan