The systemic problems of the PCLO trial of PSA screening


A Letter to the Editor published in this week’s issue of the New England Journal of Medicine has, yet again, pointed out that the US-based PCLO trial was badly flawed and was never, in fact, a trial of screening against no screening.

The “New” Prostate Cancer InfoLink is not at all sure what the point of this Letter to the Editor by Shoag et al. actually is. It has been understood for years that a major problem with the PCLO trial was that what it actually ended up testing (from a prostate cancer perspective) was the value of scheduled PSA screenings in the “screening group” as opposed to unscheduled PSA testing in the “control group”. Nearly all of the men in the control group had at least one PSA test during the course of the trial, and so of course there is no way that this was an appropriate evaluation of screening vs. no screening.

It looks to us as though this Letter to the Editor is a “political” publication intended to appear at a time close to that during which the USPSTF’s research group will initiate their review of the available data on PSA screening. The “New” Prostate Cancer InfoLink would like to think that any competent and independent research group would already be well aware of the facts presented by Shoag et al. A media release about this Letter to the Editor of NEJM has also been issued by Weill Cornell Medical College, where two of the three authors are based.

Readers need to appreciate that actually this doesn’t necessarily help the argument that mass, population-based PSA testing is a valid way to screen for prostate cancer. All that it really does is tell us that the data from the PCLO trial is actually irrelevant to any attempt to answer the question. The research group is therefore much more likely, as a consequence, to place more emphasis on data and statements from the coordinators of the ERSPC trial in Europe that the data from that trial also do not justify regular, mass, population-based PSA testing as a method to screen for risk of prostate cancer. (Yes, the coordinators of the ERSCP trial have said this clearly on several occasions.)

Frankly, it is high time we all started to appreciate that testing for risk of prostate cancer needs to be based on a spectrum of known risk factors and not based exclusively on being a male of more than 40 years of age. The PSA test has a role in helping to assess such risk, but not as a stand-alone screening test that we all need once a year.

8 Responses

  1. “it is high time we all started to appreciate that testing for risk of prostate cancer needs to be based on a spectrum of known risk factors”

    It is even higher time we all started to appreciate that testing for risk of prostate cancer under any spectrum of risk factors needs to be based on an RCT that proves such testing will provide an end-point benefit such as significantly improved OS rather than just a surrogate benefit such as reduced prostate cancer death rate. This is why ERSPC says their trial does not justify PSA screening, it failed to prove significantly improved OS.

  2. The Likely Point of this Letter About Disqualification of the PLCO Trial as Evidence Re Screening for Prostate Cancer

    Dear Sitemaster,

    Thank you so much for reporting on this letter, even though to you it seemed redundant and unnecessary! No doubt many of us appreciate the opportunity you provide us to examine the first-hand document and draw our own conclusions. (And thanks for taking the time to read and respond to replies from all of us.)

    As you have helped us see over the years since the report of PLCO trial results regarding screening for prostate cancer were originally published in 2009 in the New England Journal of Medicine (NEJM), the PLCO trial was not evidence for screening versus no-screening. In fact, didn’t you mention once that it was designed as testing a focus on screening versus usual care, which already involved a lot of screening? Nevertheless, as we, your readers, here are well aware, many health care professionals and policy makers have not understood that critical distinction, not to mention their ignorance of other major flaws in the trial as initially published. Having reported so many times on PLCO, it is perhaps natural that the authors’ letter to the NEJM seemed just another piece of publicity.

    However, just last Saturday I witnessed a vivid example that PLCO is still being misrepresented as providing meaningful evidence regarding whether there is a benefit to screening. I was at a health fair where a local urologist gave a fine talk about prostate cancer, with one glaring exception. He had a slide showing PLCO as evidence against the benefit of screening, and so described it in words. Given his evident savvy and experience, I was amazed. I talked to him afterwards and was further stunned that he realized the rate of screening in the control arm was extremely high, and he even understood that this undermined the PLCO evidence. Yet it was in his talk.

    I had a nearly identical experience with Dr. Otis Brawley, MD, then and I believe now the chief medical officer of the American Cancer Society and formerly a medical oncologist with extensive and truly impressive experience with prostate cancer. He was speaking at a Congressional hearing on prostate cancer screening, and my conversation to him before his testimony led me to believe that he understood the disqualifying flaws in PLCO and the severe prematurity of the ERSPC trial at the time. Yet his testimony treated both as sacred writ!

    As another instance of this whack-a-mole phenomenon, I have viewed a presentation about screening in 2013 at a large survivor’s medical conference by Dr. Timothy Wilt, MD, where he misrepresented the PLCO results in the same manner; though a primary care doctor with likely no experience treating prostate cancer, he was probably the member of the US Preventive Services Task Force with the best claim to any significant exposure to the disease, due to his research role. I’m sure that you and many other readers of this blog have had similar surreal experiences: encounters with intelligent people with seeming competence regarding prostate cancer who commit this serious error in drawing the proper logical conclusion, even at times when the flaw is spotlit for them.

    That’s why this letter is important, as is the venue. If I were the NEJM, I would still be smarting over the trouble the NEJM caused by allowing publication of the seriously flawed, grossly premature PLCO and ERSPC reports in 2009. Yes, there was a cautionary editorial, but it was far too gentle. It may help that the NEJM is the site for publication of this letter. Maybe even the USPSTF will get the message, though their inclusion of that obsolete 2009 reference to the PLCO paper as one of the few references provided in their recent “final” research plan to update their prostate cancer screening guidance is not an encouraging sign!

    On a different note, getting back to this letter itself, the graph of screening in the control arm in the linked letter was most helpful. I was aware the rate of such screening was above 70%, which is more than enough to disqualify PLCO as evidence against a benefit for screening, but I did not realize it was as high as just under 90% by the twelfth year, even higher than screening in the group that was supposed to be screened!

  3. In the spirit of disclosure, I am not a clinician but am a healthcare actuary who focuses on health data, outcomes, side effects, complications and cost effectiveness.

    As we all know, the conventional wisdom (pre USPSTF) is to regularly do PSA tests and, when PSA is elevated, use TRUS biopsies. And, based on this approach, many men are upgraded in Gleason score after prostatectomy … so urologists all too often do surgery on many men who could safely be on AS … since they do not have sufficient confidence that conventional tests have properly evaluated risk. So, it should be no shocker that some researchers have concluded rewards of (PSA + TRUS + unnecessary surgery) are disappointing vs. the costs and conclude that PSA tests are the problem.

    With the advent of mpMRIs and guided biopsies, clinicians need to re-think standards of care and re-evaluate what protocols are appropriate and most cost effective. The good news is NIH has already documented the superior safety, efficacy, yield per biopsy core, and cost effectiveness of fusion-guided biopsies. NIH has also evaluated the extra yield having a TRUS biopsy in addition to a guided biopsy (about 4% with fusion-guided biopsy and close to 0% with MRI-guided in-bore biopsies).

    The new mpMRI imaging and guided biopsy technology is far from perfect, primarily due to lack of experience at some imaging locations. But it is clearly superior to TRUS biopsies in sensitivity, specificity, and cost effectiveness in identifying clinically significant tumors. So future analysis should focus more on which combinations of tests, risk factors, etc., are most effective … not simply re-visiting the PSA + TRUS pathway.

  4. That is also true … and the only trial that might be able to show that if the current one in the UK … if the patients are followed for 25+ years.

  5. Response to David’s call for a trial of screening based on overall survival

    Isn’t such an envisioned trial an impractical ideal both because the “signal” from the small percentage of death from prostate cancer would be drowned out by the “noise” from the high percentage of death from all other causes, and also because of the necessary very long length of such a trial necessary due to the long survival period for even lethal prostate cancer?

    Also, we are getting early useful reports from the ERSPC, even though time from diagnosis, even in the screening group, is only 6.4 years, and shorter in the non-screening group — still quite premature. Take a look at the original ERSPC report (follow-up 9 years from enrollment) to the 11-year report and the 13-year report. You will see the benefit from screening originally described as “significant” and moving to being described as “substantial” in the 13-year report. The benefit is enhanced further when factors are considered such as noncompliance with assignment to randomized group, in other words, men in the non-screening group choosing to be screened, and vice versa.

  6. As quoted in the Sitemaster’s article, the fact that the PLCO trial had a huge problem with contaminated data has been known for years. It was obviously known to the Task Force, yet they decided to ignore it and relied on this trial to come up with a D recommendation, ignoring updated data from the ERSPC and subsequent Goteborg trials. Why is it that this is recognized now, but it was not at the time the Task Force made the recommendation? Flawed analysis or intentional misrepresentation of data? Maybe it was a blessing this all happened, as AS is now a more recognized option.

  7. Interestingly, Dr. Shoag and colleagues recently published another article (in JAMA Oncology) about PLCO in which a statistical analysis was conducted using only the screened group and they found that the analysis supported the non-benefit of screening. One wonders why the new letter with no mention of the previous article.

  8. And yet, Dear Jim, the coordinators of the ERSCP have clearly stated, several times — even after the 13-year report — that their data do not justify any recommendation of mass, population-based, PSA-based screening for prostate cancer. And no one could reasonably suggest that the ERSCP research group was in any way biased against that idea. Indeed, it would be realtively easy to get several of the members of the research group to state publicly that they did the trial because they really beleived it would show far better results that have proven to be the case.

    And Dear Wolfram … The USPSTF did not come to their conclusion last time because of the PCLO trial. They came to their conclusion because there were no good and compelling data from anywhere suggesting that, for the average man in good health, with no known risk factors, regular PSA testing led to either a meaningful reduction in risk for death from prostate cancer or an overall survival benefit. That continues to be the case today. My perspective continues to be that this did not justify a D rating. Rather, it justified either a C or an I rating because we know that there are subsets of men at significantly higher risk than the average man in good health with no known risk factors. For those men, the PSA test offers a way to monitor that risk. One group of those men at higher risk are the men who already have a slightly elevated PSA level in their 40s.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

w

Connecting to %s

%d bloggers like this: