On 2016 Apr 09, David Keller commented:

Screening for prostate cancer with PSA has never been properly evaluated

The two randomized trials of PSA screening upon which the USPSTF based their D recommendation were ERSPC and PLCO. PLCO suffered from such high rates of statistical "contamination" (including off-protocol and pre-protocol screening of control subjects) that its negative result is not considered a valid assessment of PSA screening. [1]

ERSPC reported a reduction in the prostate cancer mortality rate for screened men of 27% at 13 years by per-protocol analysis [2], versus a 21% reduction by intention-to-treat analysis. There is a 6% discrepancy between these mortality reduction estimates because the latter counts men as having been screened based only on their initial randomization, even if they never had a single PSA test. Many consider the former figure a more realistic estimate for patients who actually get screened as directed.

The high false-positive rate of PSA screening led to many unnecessary negative biopsies in the randomized trials. In clinical practice, the common-sense response to a high PSA is to repeat the test a week later for confirmation, because there are many benign causes of transient PSA elevation. If the repeated PSA is normal, the patient can be spared a biopsy. This approach has been demonstrated to reduce the harms of PSA screening compared with reflex biopsy based on a single elevated PSA level, as practiced in the randomized trials. [3]

PSA velocity was not considered in the biopsy decision. For example, in centers using a biopsy threshold PSA of 3, a man whose PSA rose from 2.9 to 3.1 (a 6% increase) was biopsied, but a man whose PSA rose from 0.5 to 2.5 (a 400% increase) was not biopsied. A rapidly rising PSA is more likely to signal an aggressive prostate cancer than a higher but essentially stable PSA.

Most subjects were screened with a PSA test about every 4 years in ERSPC, an interval long enough to allow aggressive tumors to metastasize before being detected. In clinical practice, PSA should be measured annually, or even more frequently, to better distinguish its inherent signal from its noise. The additional PSA data points can be used to establish a baseline PSA range, calculate PSA velocity, and to detect (and confirm) worrisome increases earlier, with the goal of intervening before metastasis occurs. Harms falsely attributed to frequent PSA measurements are actually caused by inappropriate reflex biopsies, which, as we have seen, are actually reduced by additional PSA data [3].

PSA screening has been associated with a greater than 40% decrease in mortality from prostate cancer [4], for which no other combination of interventions or population trends can account. This substantial decrease in prostate cancer mortality observed with PSA screening cannot be dismissed based on questionable results from flawed randomized trials. We should not abandon the intervention (PSA screening) most likely to have caused the bulk of the observed decrease in prostate cancer mortality. After all, we advise against smoking based on observational data, in the complete absence of randomized trial data.

It was unwise for the USPSTF to issue their anti-PSA recommendation without a single urologist on their panel. As health systems implement bans on PSA testing, conservative models predict that "discontinuing PSA screening for all men may generate many avoidable cancer deaths. Continuing PSA screening for men aged <70 years could prevent greater than one-half of these avoidable cancer deaths while dramatically reducing over-diagnosis compared with continued PSA screening for all ages." [5] If these models are correct, the USPSTF recommendation will be responsible for many preventable prostate cancer deaths.

Harms associated with prostate cancer treatments, such as erectile dysfunction and urinary incontinence, have been steadily reduced by advances in conformal radiation, brachytherapy, robotic surgery, imaging and watchful waiting. Men should be given a choice whether to have PSA screening. Such a choice requires thorough discussions with patients and more effort by clinicians to carefully track PSA levels over time, with the goal of maintaining or improving the reductions we have achieved in prostate cancer mortality by means of PSA screening.

References

1: Vickers AJ. Does Prostate-Specific Antigen Screening Do More Good Than Harm?: Depends on How You Do It. JAMA Oncol. 2016 Mar 24. doi: 10.1001/jamaoncol.2015.6276. [Epub ahead of print] PubMed PMID: 27010733.

2: Schröder FH and ERSPC Investigators. Screening and prostate cancer mortality: results of the European Randomised Study of Screening for Prostate Cancer (ERSPC) at 13 years of follow-up. Lancet. 2014 Dec 6;384(9959):2027-35. doi:10.1016/S0140-6736(14)60525-0. Epub 2014 Aug 6. PubMed PMID: 25108889; PubMed Central PMCID: PMC4427906.

3: Lavallée LT, Binette A, Witiuk K, Cnossen S, Mallick R, Fergusson DA, Momoli F, Morash C, Cagiannos I, Breau RH. Reducing the Harm of Prostate Cancer Screening: Repeated Prostate-Specific Antigen Testing. Mayo Clin Proc. 2016 Jan;91(1):17-22. doi: 10.1016/j.mayocp.2015.07.030. Epub 2015 Dec 10. PubMed PMID: 26688045.

4: Howlader N, Noone AM, Krapcho M et al: SEER Cancer Statistics Review, 1975-2009 (Vintage 2009 Populations), National Cancer Institute. Bethesda, MD, http://seer.cancer.gov/csr/1975_2009_pops09/, based on November 2011 SEER data submission, posted to the SEER web site, April 2012

5: Gulati R, Tsodikov A, Etzioni R, Hunter-Merrill RA, Gore JL, Mariotto AB, Cooperberg MR. Expected population impacts of discontinued prostate-specific antigen screening. Cancer. 2014 Nov 15;120(22):3519-26. doi: 10.1002/cncr.28932. Epub 2014 Jul 25. PubMed PMID: 25065910; PubMed Central PMCID: PMC4221407.

On 2016 Apr 09, David Keller commented:

Screening for prostate cancer with PSA has never been properly evaluated

The two randomized trials of PSA screening upon which the USPSTF based their D recommendation were ERSPC and PLCO. PLCO suffered from such high rates of statistical "contamination" (including off-protocol and pre-protocol screening of control subjects) that its negative result is not considered a valid assessment of PSA screening. [1]

ERSPC reported a reduction in the prostate cancer mortality rate for screened men of 27% at 13 years by per-protocol analysis [2], versus a 21% reduction by intention-to-treat analysis. There is a 6% discrepancy between these mortality reduction estimates because the latter counts men as having been screened based only on their initial randomization, even if they never had a single PSA test. Many consider the former figure a more realistic estimate for patients who actually get screened as directed.

The high false-positive rate of PSA screening led to many unnecessary negative biopsies in the randomized trials. In clinical practice, the common-sense response to a high PSA is to repeat the test a week later for confirmation, because there are many benign causes of transient PSA elevation. If the repeated PSA is normal, the patient can be spared a biopsy. This approach has been demonstrated to reduce the harms of PSA screening compared with reflex biopsy based on a single elevated PSA level, as practiced in the randomized trials. [3]

PSA velocity was not considered in the biopsy decision. For example, in centers using a biopsy threshold PSA of 3, a man whose PSA rose from 2.9 to 3.1 (a 6% increase) was biopsied, but a man whose PSA rose from 0.5 to 2.5 (a 400% increase) was not biopsied. A rapidly rising PSA is more likely to signal an aggressive prostate cancer than a higher but essentially stable PSA.

Most subjects were screened with a PSA test about every 4 years in ERSPC, an interval long enough to allow aggressive tumors to metastasize before being detected. In clinical practice, PSA should be measured annually, or even more frequently, to better distinguish its inherent signal from its noise. The additional PSA data points can be used to establish a baseline PSA range, calculate PSA velocity, and to detect (and confirm) worrisome increases earlier, with the goal of intervening before metastasis occurs. Harms falsely attributed to frequent PSA measurements are actually caused by inappropriate reflex biopsies, which, as we have seen, are actually reduced by additional PSA data [3].

PSA screening has been associated with a greater than 40% decrease in mortality from prostate cancer [4], for which no other combination of interventions or population trends can account. This substantial decrease in prostate cancer mortality observed with PSA screening cannot be dismissed based on questionable results from flawed randomized trials. We should not abandon the intervention (PSA screening) most likely to have caused the bulk of the observed decrease in prostate cancer mortality. After all, we advise against smoking based on observational data, in the complete absence of randomized trial data.

It was unwise for the USPSTF to issue their anti-PSA recommendation without a single urologist on their panel. As health systems implement bans on PSA testing, conservative models predict that "discontinuing PSA screening for all men may generate many avoidable cancer deaths. Continuing PSA screening for men aged <70 years could prevent greater than one-half of these avoidable cancer deaths while dramatically reducing over-diagnosis compared with continued PSA screening for all ages." [5] If these models are correct, the USPSTF recommendation will be responsible for many preventable prostate cancer deaths.

Harms associated with prostate cancer treatments, such as erectile dysfunction and urinary incontinence, have been steadily reduced by advances in conformal radiation, brachytherapy, robotic surgery, imaging and watchful waiting. Men should be given a choice whether to have PSA screening. Such a choice requires thorough discussions with patients and more effort by clinicians to carefully track PSA levels over time, with the goal of maintaining or improving the reductions we have achieved in prostate cancer mortality by means of PSA screening.

References

1: Vickers AJ. Does Prostate-Specific Antigen Screening Do More Good Than Harm?: Depends on How You Do It. JAMA Oncol. 2016 Mar 24. doi: 10.1001/jamaoncol.2015.6276. [Epub ahead of print] PubMed PMID: 27010733.

2: Schröder FH and ERSPC Investigators. Screening and prostate cancer mortality: results of the European Randomised Study of Screening for Prostate Cancer (ERSPC) at 13 years of follow-up. Lancet. 2014 Dec 6;384(9959):2027-35. doi:10.1016/S0140-6736(14)60525-0. Epub 2014 Aug 6. PubMed PMID: 25108889; PubMed Central PMCID: PMC4427906.

3: Lavallée LT, Binette A, Witiuk K, Cnossen S, Mallick R, Fergusson DA, Momoli F, Morash C, Cagiannos I, Breau RH. Reducing the Harm of Prostate Cancer Screening: Repeated Prostate-Specific Antigen Testing. Mayo Clin Proc. 2016 Jan;91(1):17-22. doi: 10.1016/j.mayocp.2015.07.030. Epub 2015 Dec 10. PubMed PMID: 26688045.

4: Howlader N, Noone AM, Krapcho M et al: SEER Cancer Statistics Review, 1975-2009 (Vintage 2009 Populations), National Cancer Institute. Bethesda, MD, http://seer.cancer.gov/csr/1975_2009_pops09/, based on November 2011 SEER data submission, posted to the SEER web site, April 2012

5: Gulati R, Tsodikov A, Etzioni R, Hunter-Merrill RA, Gore JL, Mariotto AB, Cooperberg MR. Expected population impacts of discontinued prostate-specific antigen screening. Cancer. 2014 Nov 15;120(22):3519-26. doi: 10.1002/cncr.28932. Epub 2014 Jul 25. PubMed PMID: 25065910; PubMed Central PMCID: PMC4221407.