Why Ki-67 may not be USEFUL in predicting aggressive prostate cancer


A paper to be presented on Sunday at the upcoming annual meeting of the American Association for Cancer Research suggests that “the proliferative index as assessed by Ki-67 immunoreactivity is an independent predictor of aggressive prostate cancer.”

As The “New” Prostate Cancer InfoLink regularly emphasizes (probably to the point of boredom for regular readers), there are no established biomarkers that allow us to differentiate aggressive from indolent disease in early-stage prostate cancer, with the single exception of the Gleason grade, which can only be determined with accuracy after radical prostatectomy because even saturation biopsies cannot reliably identify the highest grades of tissue in a cancerous prostate.

Weinmann et al. have used data from 600+ prostate cancer patients managed within the Kaiser health care systems to evaluate the predictive value of immunohistochemical expression of four molecular markers isolated from radical prostatectomy specimens. This was a case-controlled study in which data from 286 men whose prostate cancers led to their to deaths between 1971 and 2001 was compared to data from 344 “control” patients with non-fatal prostate cancer who were closely matched to the men with aggressive disease on the basis of their age, race, tumor stage, year of diagnosis, and patterns of health plan membership. However, it has to be stated immediately that data from all such case control studies have to be interpreted with great caution.

The pathologic slides of the prostatectomy specimens from all 630 patients were carefully re-examined by a single, expert prostate cancer pathologist who first re-determined the pathologic Gleason grade and who then selected the tumor specimen with the highest grade from each tumor for analysis. The formalin-fixed, paraffin-embedded tissue sections from each of the 630 highest-grade tumor specimens were then immunostained using antibodies against four known markers: Ki-67, p53, p27kip1 and BCl-2.

The immunoreactivity of each tumor specimen was then scored for each marker and the results were correlated with clinicopathologic variables (things like clinical stage, PSA level, etc.).

The results of the analysis showed the following:

  • Based on univariate analysis (in which the predictive ability of each marker was independently assessed), Ki-67, P53, P27kip1, and BCl-2 were all shown to be  statistically significant predictors of prostate cancer mortality after adjustment for age, group, race, health plan, diagnosis year, and tumor stage.
  • However, based on multivariate analysis (in which the predictive ability of the markers are all assessed together), only the immunoreactivity of the Ki-67 marker reached statistical significance (with an odds ratio = 10.6) when data from all four markers were taken into account.

Now if we take these data at face value, it would suggest that an assessment of the immunoreactivity of the Ki-67 marker ought to be capable of distinguishing between men with clinically significant, high-risk prostate cancer and men with indolent disease. Unfortunately, however, The “New” Prostate Cancer InfoLink suspects that these data are not nearly as compelling as one might immediately think.

For starters, people have been studying the potential of Ki-67 as a prostate cancer marker for 15 years or more. It seems most unlikely to The “New” Prostate Cancer InfoLink that the degree of immunoreactivity of Ki-67 is suddenly going to become a critical prognostic factor compared to Gleason grade. Why? Because to be sure you are measuring the immunoreactivity of the highest levels of Ki-67 present, you may still need to be able to identify tissue from the highest grade of cancer in the prostate — and you can’t be sure you are doing that until you can examine tissue from the prostatectomy specimen!

If there was a way to be able to measure the immunoreactivity of Ki-67 with accuracy from the amounts of Ki-67 expressed in a urine sample or similar, then we would be looking at a whole different potential opportunity, but The “New” Prostate Cancer InfoLink does not believe that that is possible. So the ability of Ki-67 to act as a marker that can predict the aggressiveness of prostate cancer before you even had to do a biopsy is very definitely in doubt, and it would seem to be unclear whether it is possible to assess the immunoreactivity of Ki-67 based on biopsy specimens, which might be helpful, but is still far from the ideal. After all, if you can’t be sure that you have found the highest grade of tumor when you do the biopsy, how can you be sure you are accurately assessing the immunoreactivity of the Ki-67?

All of this is not intended to imply that the results of this study are of no interest. That would be untrue. These data are of considerable scientific interest, but that does not necessarily imply that they have practical clinical value in initial diagnosis and prognosis of prostate cancer, or in the decision-making process when it comes to first-line treatment. In the longer term, however, these data may be of some interest in the assessment of risk for progression of men treated surgically for prostate cancer. In that case, the question will be whether the immunoreactivity of K-67 adds meaningfully to the data already available from the pathologic Gleason score in combination with the patient’s PSA and at the time of diagnosis and his pathologic stage at treatment.

2 Responses

  1. I’m going to be one of about 40 survivor representatives as part of the Scientist↔Survivor Program at this upcoming AACR meeting, and I appreciate your pointing out this work on Ki-67 and the other markers.

    However, I’m thinking we can take a more optimistic look at the results. The Sitemaster was thinking the usefulness of Ki-67 in a practical sense would be limited, “Because to be sure you are measuring the immunoreactivity of the highest levels of Ki-67 present, you may still need to be able to identify tissue from the highest grade of cancer in the prostate — and you can’t be sure you are doing that until you can examine tissue from the prostatectomy specimen!” Yes, it would be ideal to have the specimen and be able to determine the highest grade, but that was not what was done in the study, and the results were still very impressive. Just because something is not ideal does not mean that it is not outstanding! I suppose time will tell, but I’m enthusiastic about this result.

    I’m puzzled by the Sitemaster’s statement that, “there are no established biomarkers that allow us to differentiate aggressive from indolent disease in early-stage prostate cancer, with the single exception of the Gleason grade, which can only be determined with accuracy after radical prostatectomy because even saturation biopsies cannot reliably identify the highest grades of tissue in a cancerous prostate.” Again, that seems to me a confusion of the non-ideal and no-good, both for Gleason grade, but also for PSA velocity in the year prior to diagnosis. As the D’Amico teams discovered (and others have verified), a cut point at a PSAV greater than 2.0 portends a substantially less favorable outcome than would otherwise be estimated based especially on the PSA level, Gleason score, and stage.

    For me, information does not have to be perfect for it to be useful.

  2. Dear Jim:

    1. With respect to Ki-67, the selection of the highest grade of tissue from the RP specimen appears to be exactly what was done in this study. Per the abstract, (a) “we conducted a case-control study to evaluate the predictive value of immunohistochemical expression of four molecular markers in formalin-fixed paraffin-embedded tumor tissue from prostatectomy specimens” and “One expert prostate cancer pathologist reviewed all pathology slides to determine Gleason grade and select the highest grade tumor specimen for analysis.”

    2. With respect to the value of PSA velocity as a marker of the risk for cancer, Vickers et al. have reported multiple times that there are absolutely no data to suggest that PSAV is any more accurate than a single PSA value, and Catalona’s group has provided no statistically sound analysis that refutes the analyses by Vickers et al.

    3. There are most certainly markers for prognosis after a man has received first-line treatment (PSA doubling time, CTC numbers, time to presence of evident metastasis, etc.). However, such markers were not the focus of this commentary. This commentary was focused on the need for markers that are clinically accurate and clinically useful in helping a man to know: (a) whether he needs a biopsy at all and (b) whether needs first-line treatment or whether he can simply be monitored over time. The utilility of accurate markers for those two uses affects hundreds of thousands of men every year who either get unnecessary biopsies or get unnecessary treatment.

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 )

Connecting to %s

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

%d bloggers like this: