In March, a randomized clinical trial from the United Kingdom made headlines suggesting that PSA testing does not reduce death from prostate cancer. The authors of the Cluster Randomized Trial of PSA Testing, also called CAP, evaluated low intensity screening for prostate cancer, specifically a single PSA test. The authors concluded that giving men a single PSA test did not reduce rates of prostate cancer death over a median follow-up period of 10 years. However, the trial's hypothesis of a single PSA test as a screening strategy is problematic. Furthermore, the study's design was flawed to adequately test the strategy. Please see Dr. Catalona's response below for further details.
CAP trial design
Men in the CAP trial were randomly assigned to either a screening arm or a control arm. Men in the screening arm were offered a single PSA test, and those with a PSA of 3 ng/mL or greater were offered a biopsy. Men diagnosed with localized prostate cancer were recruited for the ProtecT trial for treatment, which randomized them to radical prostatectomy, radiation therapy combined with androgen-deprivation therapy, or observation without active surveillance biopsies.
By contrast, men in the control group were only given information about PSA testing if they requested it. This is standard practice for the UK's National Health Service that currently does not advocate population-based screening.
The CAP analysis included more than 408,000 men aged 50 to 69 years. After a median 10 years of follow-up (i.e., some men were followed for more than 10 years and some were followed for less), the authors found that approximately the same number of men died of prostate cancer in each group.
An Unjustified Null Result:
My Response to the CAP Trialby Willian J. Catalona, M.D.
The authors of the CAP trial concluded that a single PSA test versus no screening did not reduce prostate cancer-specific mortality after a median follow-up of 10 years. However, the CAP trial had many limitations and included questionable assumptions that diluted the benefits of PSA screening for reducing suffering and death from prostate cancer to an unjustified null result.
Here, in my opinion, are some of the important issues with the CAP trial:
A screening strategy using a single PSA test for men between 50 and 69 years old is unrealistic.
Cancer screening strategies do not rely on a single test performed once to be effective. A persistently rising PSA is a better predictor of clinically- significant prostate cancer.
Furthermore, men in the CAP trial were 50 years or older. There is strong evidence that baseline PSA levels measured in men in their 40s provides powerful predictive information about the subsequent risk of prostate cancer, aggressive disease, metastases, and prostate cancer-specific mortality.
Moreover, currently available refinements of PSA testing, such as the Prostate Health Index, are more reliable than PSA alone.
A critical omission of the CAP trial was the lack of data on metastases.
A major benefit of screening is that it decreases metastases. An analogy is wearing seat belts in cars: Is the benefit only the deaths prevented, or should they also include the catastrophic injuries prevented that did not result in death? Reducing metastases further shifts the balance between benefits and harms.
Men in the screening arm could have had higher rates of death from prostate cancer due to being treated with observation.
CAP authors asserted that randomization of their screening arm patients to various treatment options in the ProtecT trial did not affect the prostate cancer-mortality results. This was based on the assumption that treatment with surgery or radiation was not better at reducing prostate cancer death than observation. However, this has not yet been validated. In fact, in the ProtecT trial, half the patients randomized to monitoring were treated with surgery or radiation. Also, men in the monitoring group had twice the rate of clinical progression and distant metastases and a higher prostate cancer death rate.
A median follow-up of 10 years is insufficient to evaluate prostate cancer-specific mortality.
Trials conducted over a limited time period do not reveal true information about absolute benefits of screening over a man's lifetime. Early trial data underestimate benefits and exaggerate harms. Assuming steady rates of screening over time, there should be a greater decrease in prostate cancer-specific mortality at longer follow-up in the screening arm.
CAP authors assumed that early detection of what appears to be clinically-localized, low-risk prostate cancer based on biopsy findings always represents overdiagnosis of harmless cancer.
More men in the screening arm were diagnosed with prostate cancer, which is typically considered a proxy for overdiagnosis. However, in reality it consists of a mixture of overdiagnosed cases and true early-detected life- threatening cases.
Biopsy results have substantial sampling errors. In patients appearing to have low-risk prostate cancer on biopsy, 20%-40% actually harbor more threatening disease.
In the CAP screening group, the men diagnosed with cancer were significantly younger, had significantly more localized disease and significantly less advanced-stage cancer. In many of these cases, early detection of localized disease was probably fortunate. Prostate cancer is generally more often advanced and more aggressive in older men.
Future efforts to improve screening
New biomarkers and imaging techniques are now available to improve the performance of PSA screening and help stratify potential disease aggressiveness and select patients for active surveillance or active treatment. New research aims to validate these techniques for clinical use, including the focus of some of our ongoing research projects.