On 2014 May 28, David Keller commented:

Bravo - PSA screening guidelines should be issued by urologists & prostate cancer experts

According to the USPSTF website, their volunteer members devote an average of 200 hours per year to their unpaid duties on that committee, and those 200 hours are divided between consideration of several separate and vastly different issues. In addition, the USPSTF includes specialists who do not treat or screen for prostate cancer, such as pediatricians and gynecologists. Contrast that with professors of urology, who have devoted their careers to study of the complex issues related to prostate cancer screening. As a primary care physician, I place far more credence in the recommendations of urology professors when it comes to PSA screening, than to (let us say) a USPSTF gynecologist who has spent (let us say) 100 hours studying the topic, and has never examined a prostate as an attending physician. I predict that future evidence will force the USPSTF to retract their "D" recommendation against PSA screening. The harms done, though, will not be easy to retract.

The AUA guidelines suggest increasing the routine PSA screening interval to 2 years (instead of annual screening) in order to "reduce the harms of screening". However, PSA is a noisy signal, and decreasing the rate at which it is sampled will lower the effective signal-to-noise ratio and decrease the accuracy of the PSA measurement. Further, increasing the interval between PSA samples will reduce the bandwidth of PSA signal we can detect. In other words, we may miss the opportunity to detect prostate cancers with high Gleason scores and fast doubling times while they are still confined to the prostate. Increasing the gap between PSA samples to 2 years will reduce our ability to cure the most aggressive cancers with prostatectomy.

One of the oft-cited "harms" of screening is false positive PSA results triggering too many unnecessary biopsies. But none of the clinical trials used even the most basic PSA signal processing techniques to improve specificity, such as simply waiting a week and repeating the PSA before rushing to biopsy. This eliminates a great many false-positive PSA spikes. Another technique which seems likely to improve the validity of PSA screening is to include PSA velocity (the rate of increase of PSA) into biopsy decisions (1). For example, a man whose PSA has been stable at 4.0 for one year seems less likely to have a threatening prostate malignancy than a man whose PSA has tripled from 1.0 to 3.0 during that same year. Yet, every major study has relied on static PSA biopsy thresholds which would biopsy the patient with the stable PSA of 4.0 but not the man whose PSA tripled.

Other oft-cited harms of screening, such as patient anxiety, are easily overcome by physician reassurances regarding the benign nature of most spikes and glitches in the PSA, and the indolent nature of most prostate cancers.

The PSA test gives us information; it does not require a reflex response. More information is usually better, and information cannot hurt us unless we misuse it.

Reference

1: Bjurlin MA, Loeb S. PSA Velocity in Risk Stratification of Prostate Cancer. Rev Urol. 2013;15(4):204-6. Review. PubMed PMID: 24659919; PubMed Central PMCID: PMC3922327.

On 2014 May 28, David Keller commented:

Bravo - PSA screening guidelines should be issued by urologists & prostate cancer experts

According to the USPSTF website, their volunteer members devote an average of 200 hours per year to their unpaid duties on that committee, and those 200 hours are divided between consideration of several separate and vastly different issues. In addition, the USPSTF includes specialists who do not treat or screen for prostate cancer, such as pediatricians and gynecologists. Contrast that with professors of urology, who have devoted their careers to study of the complex issues related to prostate cancer screening. As a primary care physician, I place far more credence in the recommendations of urology professors when it comes to PSA screening, than to (let us say) a USPSTF gynecologist who has spent (let us say) 100 hours studying the topic, and has never examined a prostate as an attending physician. I predict that future evidence will force the USPSTF to retract their "D" recommendation against PSA screening. The harms done, though, will not be easy to retract.

The AUA guidelines suggest increasing the routine PSA screening interval to 2 years (instead of annual screening) in order to "reduce the harms of screening". However, PSA is a noisy signal, and decreasing the rate at which it is sampled will lower the effective signal-to-noise ratio and decrease the accuracy of the PSA measurement. Further, increasing the interval between PSA samples will reduce the bandwidth of PSA signal we can detect. In other words, we may miss the opportunity to detect prostate cancers with high Gleason scores and fast doubling times while they are still confined to the prostate. Increasing the gap between PSA samples to 2 years will reduce our ability to cure the most aggressive cancers with prostatectomy.

One of the oft-cited "harms" of screening is false positive PSA results triggering too many unnecessary biopsies. But none of the clinical trials used even the most basic PSA signal processing techniques to improve specificity, such as simply waiting a week and repeating the PSA before rushing to biopsy. This eliminates a great many false-positive PSA spikes. Another technique which seems likely to improve the validity of PSA screening is to include PSA velocity (the rate of increase of PSA) into biopsy decisions (1). For example, a man whose PSA has been stable at 4.0 for one year seems less likely to have a threatening prostate malignancy than a man whose PSA has tripled from 1.0 to 3.0 during that same year. Yet, every major study has relied on static PSA biopsy thresholds which would biopsy the patient with the stable PSA of 4.0 but not the man whose PSA tripled.

Other oft-cited harms of screening, such as patient anxiety, are easily overcome by physician reassurances regarding the benign nature of most spikes and glitches in the PSA, and the indolent nature of most prostate cancers.

The PSA test gives us information; it does not require a reflex response. More information is usually better, and information cannot hurt us unless we misuse it.

Reference

1: Bjurlin MA, Loeb S. PSA Velocity in Risk Stratification of Prostate Cancer. Rev Urol. 2013;15(4):204-6. Review. PubMed PMID: 24659919; PubMed Central PMCID: PMC3922327.