Prediction of 20- to 30-year risk for advanced prostate cancer

In 2007, Hans Lilja and colleagues first used data from the Swedish cancer registries and blood banks to demonstrate that a single PSA value in men between 44 and 50 years of age was highly predictive of subsequent prostate cancer diagnosis in an unscreened population. New data now support that original finding, but with a 7-year-longer timeframe.

In their new paper, Lilja et al. again describe how blood was collected from 21,277 men aged between 33 and 50 years and living in Malmö, Sweden between 1974 and 1986 Swedish city. (This represented a 74 percent participation rate from the age-appropriate men in Malmö at that time.) However, in this new study, they have been able to identify all known cases of prostate cancer through 2006 (as opposed to just 1999, as in the original study).

By using data from the Swedish cancer registries, Lilja and his colleagues were able to identify 1,408 men whose blood samples were collected between 1974 and 1986 and who had subsequently been diagnosed with prostate cancer through 2006. They were able to go back and measure the PSA levels in the archived blood samples for 1,312/1,408 of these men (93 percent) and for another 3,728 “controls” (other men whose blood samples were collected in Malmö during the same timeframe, but who had never been diagnosed with prostate cancer).

The results of this new study are as follows: 

  • Average (median) patient follow-up was 23 years.
  • At median follow-up, baseline PSA levels were strongly associated with
    • A subsequent diagnosis of prostate cancer 
    • Occurrence of advanced cancer (clinical stage ≥T3 or metastases) at diagnosis
  • 81 percent of patients with advanced prostate cancer were observed in men with a PSA above the median (0.63 ng/ml) at 44 to 50 years of age.
  • This last finding correlates closely with the data from the original 2007 study.

Lilja and his colleagues now conclude that a single PSA value at or before age 50 can potentially predict risk for advanced prostate cancer diagnosed up to 30 years later. They further propose that use of an early PSA test to stratify risk (perhaps between the ages of 35 and 50) would allow a large group of low-risk men to be screened less often but to ensure frequent (annual?) testing of a smaller cohort of men known to be at high risk, thereby improving the risk-benefit ratio for regular testing among an appropriately targeted set of men over time.

The “New” Prostate Cancer InfoLink would suggest that this new paper further substantiates the premise for a risk-based approach to prostate cancer testing that has been described previously on this site.

15 Responses

  1. Mike,

    Does your conclusion mean that you support a “single PSA” screening at an early age (35, 40, 45?) to be followed by annual screenings for men in the higher risk category?

    If so, I would strongly support you and would suggest some follow up studies to cover larger populations than just Malmo.

  2. Reuven … Please see the commentary on this topic published on September 25th this year and referred to above. It does indeed lay out a proposal of the type you appear to be willing to support.

  3. Mike:

    Correct me if I have misread your commentaries, but it would appear we need to know the median PSA levels for each age cohort to determine if the man is at high risk or not. If so, this suggests we need better benchmarks indicated on PSA tests … but what are the indicators of high risk?

    There are certain clear indicators: family history of prostate and other hormonal cancers; ethnicity; rapid PSA velocity; clearly high PSA levels (say > 10 ng/ml); PSA levels higher than median per the Swedish studies (but then we need to test everybody). Did I miss any?

    What concerns me are the significant number of men with advanced disease who may have no high risk markers … like myself, or at least a dozen other men I know who were diagnosed with advanced disease but had no high risk indicators until their PSA got way out of line.

    What am i missing here? How do you suggest we monitor men like myself and others who are first diagnosed with advanced cancer but have no high risk indicators?


  4. Dear Rick:

    Let me be very clear … In the real world, with the information and technology currently at our disposal, we will never be able to identify every case of clinically significant prostate cancer at a time that it is curable. It just ain’t gonna happen. We need tests with much greater selectivity and specificity.

    With regard to your particular case, since I don’t know what your PSA was when you were about 40, I don’t know whether you fall into the risk group that has been defined by Hans Lilja and his colleagues. They have generally defined men of 44 to 50 years of age with a PSA level > 0.7 ng/ml as being in such a high risk group.

    Third, there was extensive work in the 1990s on age-associated PSA levels. That was one of the reasons that people started to consider that 4.0 ng/ml was not an appropriate cut-point for prostate cancer risk. I haven’t done a search, but I suspect that we do have some good insight into age-associated PSA levels, based on that work and on the more recent work of Lilja et al.

  5. Good observation, Mike!

    A couple of times in my early 40s I had high PSAs (5-7 range) that then receded; the question is whether the PSA needs to remain greater than 0.7. I have no history of BPH and a small prostate that measured 19 cc before Tx.

    It has been suggested that PSA from malignancy never declines — I am not so sure.

    I think you know my position. I support widespread testing to catch the 30,000 or more advanced cancers. The problem of over-treatment lies in communication, not testing.


  6. Dear Mike:

    When I read these studies, I often wonder how the results get communicated to family physicians who are on the front line for suggesting a PSA to a male patient and then interperting the PSA results?

    In our case, our family doctor did not pick up on a rise in my husband’s PSA over a 2-year period, and we are paying for his lack of information on prostate cancer screening guidelines now.

    Our lives have been changed forever.



  7. The answer is that they get communicated poorly … if at all! The problem is that: (a) no one has responsibility; (b) broad acceptance of data like these takes years — even among the urology community, let alone the primary care community; and (c) the poor primary care physician is swamped with more information than anyone can handle.

  8. OTOH, Mike, that same “poor PCP” knows to send his female patients for a mammogram where a specialist reads the results.

    With prostate cancer, we are dealing with a disease of epidemic proportions — 230,000 new diagnoses this year; surely there is an obligation on the part of the PCPs to inform themselves. There are too many of us suffering because our PCPs were not paying attention.

    Part of our advocacy is to educate PCPs; as a result of our work, Kaiser Northern California is about to release guidelines to its PCPs, many of whom currently discount the test altogether, never mind how they interpret the result.


  9. Just looking at what little data the abstract gives, I had these random thoughts:

    1. Altthough much of the prostate cancer attention-grabbing media material refers to the one in four lifetime probability of men being diagnosed with prostate cancer, this study shows a less than 7% diagnosis from the sample. Admittedly this is only over a 30-year period and admittedly half the men diagnosed in most countries are over the age of 70, but even if we double the incidence (and a 50% increase would possibly be more accurate), it isn’t 25% by my calculation.

    2. What isn’t given is how many of the less than 7% men diagnosed with prostate cancer developed what is termed advanced cancer (a term defined in the abstract as “clinical stage ≥T3 or metastases at diagnosis“) and, perhaps more importantly what the disease-specific mortality rate was.

    3. 19% of the men with advanced prostate cancer did NOT have a PSA above the median (0.63 ng/mL at ages 44-50)

    As may have become apparent again from this post, I’m not particularly numerate but I seem to have heard of the 80/20 rule (is that the Pareto principle), which seems to have applied here?

  10. Terry:

    The 19% of men who did not get diagnosed with advanced disease still got diagnosed with prostate cancer (presumably localized disease).

    In this study, if 81% of the men diagnosed had advanced disease (as defined) then we know that 81% of 1,312 patients who had a PSA level above the median of 0.63 went on to have advanced disease. This equates to 1,062 patients.

    Actually, because widespread PSA testing was not used in Sweden until comparatively recently, what we really don’t know in this study is how many of these men had undiagnosed, localized disease. That might account for the “missing” patients with localized prostate cancer who would have been increasingly diagnosed in the USA after about 1992 because of increasingly widespread use of the PSA test.

  11. What is interesting is that with reduced use of PSA many men are diagnosed with advanced disease that is potentially more difficult to treat and has a higher mortality.

    The study reported on the stage of 344 of the diagnosed men for which they had records.

    83 patients (24%) were clinically judged to be T1, 130 patients (38%) were T2, 129 patients (38%) were T3-4, with two patients lacking stage data.This is an indication that in the absence of testing most patients were diagnosed with either DRE positive or metastatic disease. What is needed while we wait for a more specific marker is to promote active surveillance for those with more indolent character at diagnosis while understanding that cancers could progress to more aggressive forms given time and possibly poor health habits.

  12. I guess I am destined for advanced prostate cancer. My PSA ar 38 was 2.4. It wasn’t until I was 42 that a girl in my office said she thought it was high when I showed her the results of a physical I had. She worked at Labcorp on the weekends. My PSA was 2.7 by then. None of the three primary care physicians I had seen in the last 4 years thought there was anything wrong with it. Since there is absolutley no evidence that screening reduces mortality, being 45 now, even with my prostate out, in 20 years I am sure I will have advanced stage prostate cancer.

  13. Chris: I don’t think that’s a justifiable mindset at all.

    If you have received successful first-line treatment, and your PSA has subsequently been stable, there is every good reason to think it will remain so. Most of the patients in the Lilja studies were not receiving early first-line treatment.

  14. Mike, I was diagnosed and treated at 42 years old. I truly believe that prostate cancer is a “long term” cancer. Gleason 6 cancer today with the lead time basis that PSA testing gives us is probably a 20- to 30-year cancer. Since most men are diagnosed at age 70, prostate cancer does not kill them in the next 20 to 30 years because they die of something else before that. I think the jury is still way out on the younger men who are diagnosed.

    I hear too many doctors quote survival statistics, and when you look at the studies they are full of old men. Nobody knows how this cancer will “typically” behave in the long run. I read the article you had on here about patients who were tested for circulating tumor cells. Each patient had a tumor volume of less than 0.5 cc; 40% tested positive for CTCs. These patients were all found to have localized prostate cancer as well. Unfortunately, that is the “nail in the coffin” to me. With a life expectancy of 40 more years and an extremely high chance of CTCs, as far as I am concerned, I have little hope of my treatment being successful. By the way, I had 100% organ-confined disease and a radical prostatectomy done by Alan Partin at Johns Hopkins.

  15. Chris:

    Did you see my commentary on the 30+ year data from the Virginia Mason Clinic that was reported at the Genitourinary Cancers Symposium earlier this year?

    Are you aware that Pat Walsh’s very first nerve-sparing patient (treated back in 1982) is still happily alive and thoroughly well?

    If you had organ-confined disease that was 100% removed, you can agonize all you like, but I believe you have an excellent chance of never progressing. Late disease progression at 20+ years does happen, but it is very rare. And the longer patients go without a rising PSA after first-line therapy, the lower the probability that they will ever have a biochemical recurrence, which is in complete contrast to your concerns about CTCs.

Leave a Reply

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

You are commenting using your 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: