Age-restricted PSA screening could reduce preventable deaths, overdiagnosis

The introduction of PSA into the clinic and its rapid adoption since its FDA approval in the early 1990s led to a dramatic change in the way we diagnose and treat men with prostate cancer. Importantly, the consequences of its widespread use have ignited a broader discussion surrounding cancer biomarker regulatory approval in the United States.
There is no question that PSA screening was/is overutilized. Nevertheless, it is because of PSA screening that we have had a significant stage migration toward earlier-stage disease, and it is quite likely that the significant reduction in prostate cancer mortality we have observed over the past 25 years is similarly due in large part to its use. It is also true, however, that PSA overutilization has led to the overdiagnosis of prostate cancer, subjecting large numbers of men to the potential physical, mental and financial harms associated with the treatment of indolent disease (ie, cancer that does not cause harm).
The problem has been further complicated by the results of two large PSA screening trials. The findings from these two studies led the US Preventive Services Task Force to conclude that the potential risks of PSA screening in healthy men outweigh the potential benefits. Notably, the American Urological Association and American Cancer Society recommend that men be informed of the risks and benefits of screening before PSA screening is performed. ASCO recommends screening be discouraged in those who are expected to live <10 years. ASCO also recommends a joint discussion between patient and physician take place before PSA screening is employed in men with a life expectancy of >10 years. Although easily said, the practical application of these guidelines in the clinic can be challenging. The development of robust patient decision aids that capture patient factors and preferences and reconcile them with best evidence is a significant clinical need.
What would happen if we eliminated PSA screening altogether? Gulati and colleagues report the results of an analysis based on two models of prostate cancer natural history and diagnosis that were previously developed using reconstructed PSA screening patterns and incidence data in the United States. The authors’ objective was to use the models to predict incidence and mortality rates for the period from 2013 through 2025 under continued PSA screening and under discontinued PSA screening for all men or for men aged >70 years.
Their models predicted that continuation of recent screening rates will overdiagnose 710,000 to 1.1 million men but will avoid 36,000 to 57,000 cancer deaths over the period 2013 through 2025, and that discontinued screening for all men would eliminate 100% of overdiagnoses but fail to prevent 100% of avoidable cancer deaths. Continued screening for men aged <70 years would eliminate 64% to 66% of overdiagnoses but fail to prevent 36% to 39% of avoidable cancer deaths.
Thus, the authors conclude that “discontinuing PSA screening for all men may generate many avoidable cancer deaths while continuing PSA screening for men aged <70 years could prevent greater than one-half of these avoidable cancer deaths while dramatically reducing overdiagnoses compared with continued PSA screening for all ages.”
Although there is no question that these findings will be debated and discussed, what is clear is that — despite recent improvements in systemic therapy for advanced disease — metastatic prostate cancer is still lethal in virtually all afflicted. As many in the field have said, we cannot afford to go back to the “old days.”  This is true in many ways — we also cannot go back to “indiscriminate” PSA screening. What is needed is a more patient-centric approach coupled with a more nuanced and rational national policy that takes into account both the good and the bad of PSA screening. PSA screening does not have to be an all-or-none proposition and, after all, “deadly cancer” is the enemy.

I completely agree with Drs. Hussain and Palapattu, but I think that we also should recall that analyses of big sets of data often overlook important nuances of patient and case selection bias, misreporting of registry information, and unreported crossover of patterns of surveillance. Although the hypotheses advanced are important and perfectly reasonable, we need to anchor them with primary level one data that are validated before attempting to define new national policy approaches to screening of prostate cancer. Although modeling studies are useful, they are models and not reality. Our new approach, whatever it is, needs to produce a clear and sustained OS benefit, and the data need to apply equally to African-Americans and to men with a positive family history of prostate cancer.