“A plea for individualized prostate cancer screening”

For some years, Vickers, Lilja, and their associates have been arguing that baseline PSA level is able to predict long-term risk for prostate cancer, and now a new paper in European Urology seems to provide support for this argument and a practical clinical strategy for its application.

From 1998 to 2012, Randdazzo et al. have been tracking the PSA levels of a cohort of  > 4,000 men in a population-based screening study. They have now analyzed their data to look specifically into whether re-testing intervals for all men with a PSA level of < 3 ng/ml could be associated with the probability of detection of more aggressive forms of prostate cancer (i.e., those with a Gleason score of 7 to 10).

Here is what they found:

  • Their cohort comprised 4,350 men aged between 55 and 70 years.
  • Men in their cohort could be divided into three groups based on their baseline PSA levels:
    • Group A: Men with a baseline PSA level of < 1.0 ng/ml (N = 2,416, 55 percent)
    • Group B: Men with a baseline PSA level of 1.0-1.9 ng/ml (N = 1,371, 31.6 percent)
    • Group C: Men with a baseline PSA level of 2.0-2.99 ng/ml (N = 563, 12.9 percent)
  • The average (median) follow-up period for the men in this study was 11.6 years.
  • During the 12 years of the study, aggressive prostate cancer was identified in
    • 25/2,416 men in Group A (1.0 percent)
    • 80/1,371 men in Group B (5.8 percent)
    • 34/563 men in Group C (6.0 percent)
  • At 4 years of follow-up, aggressive prostate cancer was identified in
    • 0.0 percent of men in Group A
    • 0.29 percent of men in Group B
    • 1.8 percent of men in Group C
  • At 8 years of follow-up, aggressive prostate cancer was identified in
    • 0.2 percent of men in Group A
    • 1.3 percent of men in Group B
    • 2.5 percent of men in Group C
  • The patients’ baseline PSA level was the only independent risk factor for aggressive prostate cancer.
    • For men in Group B compared to Group A: the hazard ratio (HR) = 6.06; p< 0.0001
    • For men in Group C compared to Group A: HR = 7.33; p< 0.0001.

Randdazzo et al. conclude that:

  • Baseline PSA level was the only predictor of aggressive prostate cancer in their cohort of patients.
  • PSA re-testing intervals can be adapted based on baseline PSA levels as follows:
    • 6-8 years if a man’s baseline PSA is < 1.0 ng/ml
    • 3-4 years if a man’s baseline PSA is 1.0-1.9 ng/ml
    • Annually if a man’s baseline PSA level is 2.0-2.99 ng/ml

They refer to this re-testing model as a “PSA pyramid”.

What is fascinating about this study is how closely it appears to correlate with the theoretical proposals previously suggested by the Vickers, Lilja group in New York and Sweden. The difference in hazard ratios between the groups is very large, and appears to give a very clear rationale for the individualization (“personalization” if you will) of PSA testing based on initial age and baseline PSA level. Indeed the title of this post is the subtitle of the paper by Randdazzo et al.

To be clear, what this study does not mean is that no one with a PSA level of < 3.0 ng/ml needs a biopsy. There are exceptions to every rule, and good clinical care requires that physicians pay attention to those potential exceptions: one more reason why men should be having regular digital rectal exams, for example. What is does mean, however, is that in men with a baseline PSA level < 3.0 ng/ml, a finding of aggressive prostate cancer in the next 12 years is unusual ([25 + 80 +34]/4,350 = 3.2 percent in this study) and that annual PSA testing for most of these men is probably completely unnecessary.

What it would be nice to know is

  • Whether these data are also applicable back to men of (say) 40 or 45 years of age
  • Whether these data are specifically applicable to high-risk subgroups of men, such as African Americans, or whether their “risk pyramid” is somewhat different.


10 Responses

  1. I would say no … it does not apply to those in their 40s. This says that 0% were aggressive at 4 years from baseline of < 1.0 ng/ml My baseline was 0.7 at age 40 and then at age 44 — aggressive cancer. So just my case alone would make it more than 0 percent and I know plenty others in their 40s with similar stories.

  2. Jerry:

    No one is suggesting that there aren’t going to be exceptions. There are exceptions to everything in life: it’s not a perfect world. The question I am asking is what the probability is.

    Vickers et al. suggested that the cut-point for a man of 45 was 0.7 ng/ml. By the time you were 45 you’d had a PSA of 4.03. Yes?

    The question is, what percentage of 4,000 40-year-old men in a randomly selected population (e.g., in Palmyra, PA) with a PSA level of < 0.7 ng/ml will have a PSA level of 3.0 ng/ml or greater within 4 years. If it is only 1/4,000, that's 0.025%; and 3,999 men don't need another PSA test for another 4 years (or more).

  3. I was 44 when diagnosed with a psa of 4.03. I know I am an exception, but that’s the purpose of screening…to filter the exceptions. I personally know several exceptions to this rule.

    PSA testing was completely necessary for 139 men in this study.

  4. Dear Jerry:

    Actually no. The purpose of PSA screening was never thought to be to “filter the exceptions”. The purpose of PSA screening, when it was introduced in 1992 (and I know because I was there), was to find men of 55 and older with a PSA level of 4.0 or higher and then to biopsy them for risk of prostate cancer. No one at that time thought PSA testing would be valuable for most men in their 40s.

    Your assumption is one of the problems with all of the discussions around PSA screening … i.e., that it could be used accurately to detect risk for prostate cancer in men regardless of their age. The truth is that you probably had some prostate cancer even when your PSA was 0.7 ng/ml when you were 40, but it wasn’t being expressed through an elevated PSA level. Also, what most of us thought back then was that any amount of prostate cancer (even with a Gleason score of 5 or 6) on a biopsy core came with some significant degree of risk for metastatic disease. It is now clear that that is not the case too.

    What we know today is that there is a very, very small percentage of men of 45 and less (perhaps with some form of genetic predisposition, but we don’t know that yet) who are at significant risk for aggressive prostate cancer despite the fact that they have a “normal” PSA level. We also know that we have no current test that can detect that risk in those men. That is why you are exception. Because even the model proposed by Randdazzo et al. doesn’t work for those men. It probably works for 99% of us, but then there are the exceptions.

    Exceptions to general rules are always a problem, but one way to look at this is that you don’t have and you never had common or garden “prostate cancer”. You have an undefined and specific type of prostate cancer that doesn’t seem to generate a lot of PSA (at least in the early stages of its development).

  5. Yet, through my family doctor’s efforts (call it screening or whatever) the PSA result of 4.03 prompted a biopsy which found advanced, aggressive disease.

    To me, and a lot of other “exceptional” people, that test was necessary and provided good information.

  6. Jerry:

    No one is suggesting that PSA testing has no value. The point that Randdazzo et al. are making (and I completely agree with them) is that the use of the PSA test has greatest value when it is applied based on the clinical and other factors that are affecting the individual patient (as opposed to a “once size fits all” mentality).


    Thank you so much once again for posting this most interesting study. This should help move the field forward in choosing wise intervals for screening. These are numbers we advocates and educators can use at health fairs to show men and their loved ones who are concerned about screening and how to do it effectively. It should also help reduce health care costs for individuals, insurers, and national health systems.


    We have norms for PSA for healthy men of different ages. (Those norms, such as median PSA values, are probably closely related to the size of the prostate, which typically increases with age, of course.) The findings in this study could be interpreted as saying that a PSA higher than the norm suggests some kind of issue, which in a relatively small percentage of cases is prostate cancer, though more likely early benign growth or an infection. The study also seems to suggest that PSAs further away from the norm indicate risk that is higher than risk for elevated PSAs that are not as far from the norm.

    If these are true views, then it seems likely that the same would hold true for men in their forties. In other words, a PSA above the norm for a man of 40 would suggest some risk, and a PSA further above the norm would suggest a higher risk.

    Hasn’t there been some research reviewed here indicating that a man with a baseline PSA at 40 that is well below 1.0 has very high odds of not developing an elevated PSA throughout all or most of his 40s? It seems to me that the principles suggested by this study could well apply to men in their forties, but perhaps with a lower cut-point than < 1.0, such as, pulling a number from the air, < 0.8.


    I’m wondering if any team has done a similar study based on PSA density. I’m thinking not, but that would be nice information to have. Such a study would be a lot more involved as it would require a reliable number for the size of each participant’s prostate as well as the PSA. Ideally the size would be based on ultrasound or other imaging, but a fairly good rough number based on DREs might be adequate.

  9. Jim:

    The number used by Vickers et al. in their studies has been 0.7 ng/ml for men in their 40s.

  10. Does the Vickers low baseline total PSA hold water? In my case, it looks like it does, but I had my doubts 1 year ago.

    My total PSA at age 53 was 0.5 ng/ml and at age 58 was 1.5. According to Vickers, I had a 1% risk of dying from prostate cancer in the next 10 years.

    Then, a year ago, the total PSA jumped to 8.27 from 4.08 done 6 months previously. I opted for an mpMRI and got one at two major centers, with the latter one showing three possible lesions, 1.0-1.5 cm in diamater, in the transitional zone, which they could biopsy (the first center said they could not biopsy the lesion) by MRI fusion plus systematic and random biopsy cores for a total of 25 cores, all negative for prostate cancer. The prostate size was 100 g, so the PSA density was 0.08.Gleason 6 and is designed to reduce 2nd biopsies.

    My local urologist this week said (before the 4KScore test result was received) he wasn’t worried about the rise in total PSA and the rectal exam was negative except for an enlarged prostate, so a repeat biopsy was not recommended.

    I went ahead and got the 4KScore test anyway, just to play it safe. The result was 3% which was quite low and basically said to me, you don’t need a second biopsy. That was a big relief and confirmation of other information I was hearing.

    One should know the 4KScore test costs $395 and is not [yet] covered by insurance.The specimen drawn in a purple top plasma tube needs to go into the refrigerator immediately and be kept in the refrigerator until just before the overnight FedEx pickup. A frozen sponge is then inserted into the specimen’s styrofoam packaging and sealed up just before the pickup. I had doubts whether my urologist’s office could do all this and so I did the packaging myself. I got a call the next day, saying the specimen arrived cold but I had forgotten to mark down the date and time the specimen was collected. OPKO customer service was quite good and, with a signed consent, faxed me a copy of the report that was also sent to my urologist.

    It is interesting how the Vicker’s baseline total PSA prediction (1% risk) agrees quite well with the subsequent 4Kscore test result (3%). The 4KScore test was also developed by Dr. Vickers .

    Thank you prostatecancerinfolink.net for guiding me to the 4KScore test which you reported as being available now (for the past 2 months).

    FYI, Dr. Lin, from the University of Washington, has a nice video presentation he gave at the recent AUA meeting regarding the 4KScore test.

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 )

Twitter picture

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

Facebook photo

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

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: