Gleason score and metastatic prostate cancer; risk for 4-year mortality

It is difficult to accurately interpret a set of data just published that was designed to assess the relationship between 4-year survival and Gleason score at diagnosis in men initially diagnosed with metastatic prostate cancer.

According to this new paper by Rusthoven et al. (published on line in Urologic Oncology this month), there is a close association between the presence of any amount of Gleason pattern 5 in the biopsy specimen at diagnosis and death within 4 years among men initially diagnosed with metastatic prostate cancer. This is hardly a major surprise. Gleason pattern 5 has long been known to put patients at high risk for prostate cancer-specific mortality.

The full text of the paper will be available on line for a few weeks. After that, only the abstract will be available.

Rusthoven et al. used the Surveillance, Epidemiology, and End Results (SEER) database to identify men with metastatic prostate cancer, diagnosed between 2006 and 2008. They then classified these men by Gleason score, including the possible variants thereof:

  • Gleason 6 (i.e., Gleason 3 + 3)
  • Gleason 7 (i.e., Gleason 3 + 4 and Gleason 4 + 3)
  • Gleason 8 (i.e., Gleason 4 + 4, Gleason 3 + 5, and Gleason 5 + 3)
  • Gleason 9 (i.e., Gleason 4 + 5 and Gleason 5 + 4)
  • Gleason 10 (i.e., Gleason 5 + 5)

They then analyzed the impact of Gleason score on overall and prostate cancer-specific survival.

Here is what they found:

  • 4,654 patients were evaluable.
  • At 4 years of follow-up, the overall survival rates were
    • 51 percent among patients with Gleason scores of 6
    • 45 percent among patients with Gleason scores of 7
    • 34 percent among patients with Gleason scores of 8
    • 25 percent among patients with Gleason scores of 9
    • 15 percent among patients with Gleason scores of 10
  • Also at 4 years of follow-up, the prostate cancer-specific survival rates were
    • 69 percent among patients with Gleason scores of 6
    • 57 percent among patients with Gleason scores of 7
    • 44 percent among patients with Gleason scores of 8
    • 33 percent among patients with Gleason scores of 9
    • 21 percent among patients with Gleason scores of 10
  • Survival differences for Gleason scores of 7 vs. 8, 8 vs. 9, and 9 vs. 10 at 4 years of follow-up were
    • Highly significant on univariate analysis
    • Highly significant on multivariate analysis (accounting for age, PSA level, and T stage)
  • The presence of Gleason pattern 5 was an independent prognostic factor
    • Overall for all patients with any Gleason score from 6 to 10
    • On primary/secondary Gleason pattern comparisons within the Gleason 8 and Gleason 9 subgroups.
  • No survival differences were evident between men diagnosed with Gleason scores of 3 + 4 = 7 vs. 4 + 3 = 7.
  • Lower PSA level, younger age, and lower Gleason score were all associated with improved survival.
  • Gleason score was the strongest prognostic factor for prostate cancer-specific survival.

Rusthoven et al. conclude that

In this large population-based cohort, stratified survival outcomes were observed for GS 6 to 10, with sequential comparisons of GS 7 to 10, and the presence and extent of Gleason pattern 5 representing independent prognostic factors in the metastatic setting.

However, here’s the problem …

We know that men diagnosed with true Gleason 6 disease have a very, very low risk for metastasis. And the authors acknowledge this is the full text of their article. Since all the men in this study who were diagnosed with Gleason 6 disease (about 6 percent of the total) went on to become metastatic, one pretty much has to conclude that all of these men (or at least very nearly all of them) actually had occult prostate cancer with a higher Gleason score than the one identified at diagnosis.

However, what the authors do not acknowledge is that this fact potentially affects all of the other data in this study! If the accuracy of the Gleason scores in the biopsy data in just 6 percent of the patients in this study is that questionable, then what does it say about the accuracy of the biopsy data for the other 94 percent?

Thus, while the data provided by Rusthoven et al. are certainly of intellectual interest, one has to be careful about taking these data too literally. Does the paper reasonably imply that higher Gleason scores are associated with higher risk for prostate cancer-specific mortality within 4 years among men initially diagnosed with metastatic prostate cancer? Sure it does, and that general statement is probably accurate. It is probably also reasonable to conclude that the presence and extent of Gleason pattern 5 is independently predictive of risk for prostate cancer mortality within 5 years. However, can we take all the actual percentages in this paper at face value? Probably not, because they are clearly impacted by the accuracy of the actual biopsy data available.

7 Responses

  1. Two questions come immediately to mind. First, how was metastatic disease diagnosed? Secondly in a patient such as myself who did have pattern 5 on biopsy (Gleason 9) and who had no clinical evidence of metastatic disease, but whose radiation oncologist was absolutely convinced that metastatic disease was present, does the mere fact that I am still alive 3.5 years later (and presumably will be alive in 6 months since my PSA 2 weeks ago was 0.04 ng/ml), argue against the presence of occult metastatic disease? I think I know the answer to the second question, but I always find your responses enlightening.

  2. Michael:

    The paper is not specific about how the patients were diagnosed, but the standard way to diagnose metastatic disease is on the basis of a positive bone scan or CT scan. Without such a positive scan, micrometastasis may well be suspected (i.e., M0 disease) but actual metastatic (M1) disease is not proven.

    With regard to your own case, if you were only suspected of having metastatic disease at diagnosis (i.e., a negative bone scan and a negative CT scan), then the data from this study are not applicable to you (in my opinion).

  3. I followed your link to the full text of the paper and found this statement: “Data regarding tertiary Gleason patterns and systemic therapy administration, including androgen deprivation therapy and chemotherapy, were not available for analysis.”

    Thus, aren’t the conclusions also blurred because some Gleason 6 and 7 men probably had tertiary pattern 5, and no doubt some men in the study were on hormone therapy or other treatments which, presumably, would affect survival (or if the treatments didn’t, that would be useful to know too)?

    Thanks Sitemaster for your insight!



    Hi MHHJ and Sitemaster,

    This study shows very substantial lethality for high Gleason cancer that had become metastatic (“microscopically confirmed”) at the time of diagnosis in the 2006-2008 timeframe. While the survival curves in Figure 1 show some leveling off, declines in overall and disease-specific survival were still steadily continuing at about the 4.5-year point shown in the figure. It’s clear that well under 50% of men with Gleason scores 8-10 and metastatic disease at diagnosis in that time period were alive at the 4.5 year point. It seems reasonable that patients like these would account for the bulk of the deaths caught in the ERSPC and PLCO trials that both featured very short follow-up from diagnosis (viewed against extremely high rates of survival in the US at 5 and 10 years at that time) at first publication in 2009.

    Is it not also reasonable that a key group benefiting from screening would be men like you, MHHJ, and me (bPSA 113.6 in December 1999, Gleason 4 + 3 = 7, all cores positive, most 100% cancer, perineural invasion, PSA doubling time likely between 3 and 4 months, but bone and CT scans negative with ProstaScint — one suspicious spot in an unusual area — also likely actually negative)? Men like us who were screened in the trials were picked up in time before our cancers went metastatic (hopefully), perhaps just in time for effective treatment. Men like us who were not screened likely would have a high probability of contributing to the mortality totals in the trials.

    It also seems reasonable that the ERSPC, with its average 4-year gap between PSA tests, would allow some of us very high-risk guys to look fine at initial screening, develop aggressive disease the next year, and be well on our way toward serious jeopardy by the next screening 4 years after the initial screen, thereby diluting the overall, long-term impact of screening by failure to catch as many of those men as could be caught with a “smart” screening approach featuring more frequent (such as annual) screening for those at higher risk, as indicated in part, perhaps by a prior PSA greater than 1.

    I’m thinking that information in this paper will help us better understand results from the screening trials as years of follow-up tick by, especially the ERSPC that had superior compliance compared to PLCO and seems much more meaningful, despite its serious flaws.

    Congratulations on that wonderfully low PSA at the 3.5 year point, and wishing you many happy returns!


    The authors chose their time period to capture the first full year when the revised Gleason grading system was available for M1 disease, 2006. The revision is the updated 2005 International Society of Urologic Pathology (ISUP) consensus recommendations for Gleason grading of prostate carcinoma (see Epstein et al.; no useful abstract) I’m thinking that system would also have applied to cancer within the gland itself, and I’m a bit puzzled by the statement relating the revised system only to the determination of M1 disease.

    My impression is that that revised Gleason grading system calls for the first grade to be the grade of the predominant cancer in the sample, as in the previous Gleason protocol, but with the second grade being the higher of grade 5 or 4 if present at all, and if not, the most prominent grade present. Moreover, some patterns previously considered grade 3 are now reported as grade 4. This later paper by Epsetin presents some detail on the changes. That means a higher likelihood of reflecting a second grade of 5 or 4 if present at all, even if in a relatively tiny proportion.

    Therefore, for the same group of SEER patients evaluated under the new system, as was done in this study, compared to the old system, a higher proportion of Gleason grade 4 and 5 patients would be picked up for the study. Of course, it’s likely compliance by pathologists with the new scheme was not perfect in that first year (or now).

  6. Richard:

    There are actually all sorts of small details that would affect the interpretation of the data presented in this paper. The lack of tertiary Gleason data is certainly one of those. However, I doubt if this is as important as the probable need to upgrade the Gleason 6 and some other patients based on the failure to identify the presence of primary or secondary Gleason pattern 4 tumors.

    With regard to the treatment issue, I would be surprised if most of the men in this analysis were not on androgen deprivation therapy and (probably) chemotherapy too at some point if they were initially diagnosed with metastatic disease. However, the authors’ point is about their survival regardless of type of management.

  7. Dear Jim:

    One of the things that everyone seems to ignore when they talk about “screening” is that — regardless of the evidence in its favor (e.g., in colorectal cancer, where we know with absolute certainty that screening does extend cancer-specific survival) — only about 50% of persons who should be tested actually go to get tested.

    In other words, one has to ask to what extent human nature undermines the possibility that we would find half the patients who need treatment early enough to be treated appropriately. There has certainly been a significant decline in the risk for prostate cancer mortality over trhe past 25 years, and some of this is undoubtedly due to the availability of PSA testing. On the other hand, the absolute number of deaths from prostate cancer has barely budged at all, and a surprisingly high number of men are still diagnosed each year with true metastatic disease or with M0 disease that is clearly more aggressive than yours (because nothing stops some of those patients’ PSAs from rising). One has to assume that many of those men reached those stages for the simple reason that they never or rarely took any action to monitor their health in conjunction with a qualified clinician. Thus, they were diagnosed based on symptomatology.

    Which of course leads me to ask what actually stimulated you, Jim, to have that PSA test in 1999? Had you not had one any earlier? And if not, why not? I know (because I helped to facilitate it) that there was extensive public information about PSA testing available for at least 7 years before that. Did you know of that and ignore it, or were you, like an awful lot of men, simply not paying enough attention to your risk for cancers of any type? I understand why you are ardent in your advocacy for screening now … but what about before your own diagnosis?

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 )

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 )

Connecting to %s

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

%d bloggers like this: