“Over-diagnosis” and “over-treatment” of prostate cancer — definitions, expectations, and estimations


A conversation between two extremely knowledgeable prostate cancer advocates in comments on an earlier post has highlighted the underlying problem of what we mean by “over-diagnosis” and “over-treatment” of prostate cancer in a highly educational manner.

The way The “New” Prostate Cancer InfoLink has long looked at this is as follows:

  • Over-treatment is the active first-line treatment of any man diagnosed with prostate cancer who would never have progressed to have evident metastatic disease or suffered other “significant” clinical problems as a consequence of prostate cancer during his lifetime. [In other words, in hindsight, the man was demonstrably given unnecessary treatment for a clinically insignificant problem.]
  • Over-diagnosis is therefore the diagnosis of prostate cancer in any man who would never have progressed to have evident metastatic disease or suffered “significant” clinical problems as a consequence of prostate cancer during his lifetime. [In other words, again in hindsight, the man was unnecessarily diagnosed with a clinically insignificant problem, thereby placing him at risk for demonstrably unnecessary treatment.]

So the first problem is that, for any specific individual, knowledge that they were over-diagnosed or over-treated requires hindsight. Science can often provide accurate hindsight … but such accurate hindsight is not yet available in the management of early stage, localized prostate cancer — even from the massive, high-quality Scandinavian databases (as we shall explain below).

The second problem is that, yes, we can quibble over what “significant” might mean. Just as an example, many men who might meet the above definitions could, over time, have some increasing and clinically significant urinary tract problems. Such problems are very common in men as they age. They can be relieved by use of 5α-reductase inhibitors (finasteride, dutasteride) and/or a laser- or other thermally-based form of transurethral resection of the prostate (TURP) and the patients (and their doctors) might still never need to find out that prostate cancer was a contributory cause.

From a clinical point of view, the above two definitions have always seemed to be good ones. On the other hand, from a “scientific” point of view (and on an individual patient basis), given a disease that can take 40+ years to develop from the first potential pathologic sign of some cells that may indicate cancer in the prostate to the actual prostate cancer-specific death of a specific patient from a relatively classic case of slowly growing adenocarcinoma of the prostate, we fully acknowledge that these definitions are not too helpful.

As a consequence, researchers use different definitions and assumptions and models to try to estimate the clinical frequencies of over-treatment and over-diagnosis. These estimates are just that, estimates. We do not have the benefit of real hindsight. Prior to the PSA era, men with very early-stage forms of prostate cancer were rarely diagnosed when their cancer had a high probability of being truly localized to their prostate. So we have no historic 30- or 40-year-long database of the outcomes of men diagnosed with truly early stage prostate cancer. By contrast, with the dawning of the PSA era, it became relatively easy to find and diagnose prostate cancer, but we still have no idea how to differentiate accurately between men with clinically significant and clinically insignificant, early stage, localized prostate cancer.

The question posed by one of the discussants in the above-mentioned “conversation” was (to all intents and purposes) the following:  … What percentage of men actually diagnosed with and treated for localized prostate cancer in the USA today could have gone either for many years (e.g., at least 10 years) or for the rest of their lives without needing active first-line intervention for their cancer (as opposed to some relatively minor symptoms associated with their cancer)?

The “New” Prostate Cancer InfoLink doesn’t actually believe that we can answer that question with scientific rigor. Furthermore, we will probably never have a good answer to that question until we have much better diagnostic tools than the PSA test and the multi-core biopsy. The only way to answer that question today requires early diagnosis and careful monitoring (with or without active treatment) of very large numbers of men, many of whom (at least in the USA) would be utterly unwilling to be just monitored for the next 30+ years … and that’s assuming there was enough money to even consider doing such a study. In the meantime, all of the computer models and past studies come with definable assumptions and/or flaws that limit their practical application. The only ongoing trial that appears to have the potential to answer this question is the ProtecT study, still enrolling patients in the UK. However, that trial is only designed to enroll about 2,000 patients who must be between 50 and 69 years of age … so it will offer us no useful information about the incidence of clinically insignificant disease in men under 50 years of age or of 70 years of age and older; it may not be a big enough study anyway; and there is no clear intent to follow all patients in the study to death (since the question we are asking is not the question the trial was designed to address).

One of the discussants in the prior conversation suggested that (in his personal opinion, after much study)  the true incidence of over-treatment in the USA is somewhere in the 10 to 20 percent range. The “New” Prostate Cancer InfoLink believes that it may well be even lower than that for men in the age range of 40 to 55 years … perhaps as low as 5 to 10 percent. However, once one starts dealing with men diagnosed at 70 to 75 years of age, the incidence of over-treatment is potentially much higher because their life expectancies are much shorter. Age at diagnosis is critical to this discussion. The “New” Prostate Cancer InfoLink thinks that over-treatment of men in the 70s could easily be up as high as 60 to 65 percent (and possibly higher). And in the not too distant past, we have seen evidence that the number of men getting PSA tests and consequent diagnosis and treatment in their 70s is increasing. What percentage of those men are actually getting over-diagnosed? We don’t know, but we suspect it is “way too many” and we think that some of the urology researchers who have built careers on the idea that prostate cancer is being significantly over-treated (e.g., Peter Albertsen, MD) would tend to agree with that perspective.

People who have evidence of prostate cancer in their 30s, 40s, and 50s do indeed need to get diagnosed early and make (commonly) very hard decisions about what to do. Those decisions may be right, and they may be wrong, and we need better tests to be able to better categorize these men into well-defined risk groups because some of them (albeit a small number) really do have clinically insignificant disease. In the case of older men, we need to find a way to better manage people’s “cancer fear” since at least some (and possibly many) of those older men are being diagnosed with (and treated for) forms of prostate cancer that will never be of any clinical significance in their lifetimes.

9 Responses

  1. Why do men between 55 and 70 never benefit from any of this info? It seems like we are the forgotten ones. I hate being in limbo with my decision making. There are too many different options that the proponents tout as the “best” decision regarding prostate cancer treatment. I would really like some more direct info and treatment venue suggestions regarding the recent laser treatment options.

  2. Ninety to 95 percent of men in their 30-50s who are diagnosed require “active first-line treatment”? None of those men might have a small amounts of low grade cancers that could be monitored with active surveillance giving them a year or five years or ten years before being subjected to the brutalities of the “cures”?

  3. Tracy:

    The article above does not say or even imply that men between 30 and 50 who are diagnosed with prostate cancer can’t be managed for a significant period of time with active surveillance. What it says is that on a lifetime basis they are highly unlikely to have been “over-diagnosed.” It also says we have no idea what percentage of such men could be effectively and safely managed for (say) 10 years with active surveillance.

  4. Richard:

    At this time there is minimal information on the application of laser surgery in the treatment of prostate cancer. Is it a possibly effective and safe option? Sure. It has been used for years to treat benign prostatic hyperplasia. Do we have even 5-year follow-up data on 100 patients from anywhere in the world? Not that I am aware of.

    I am not sure what you are implying in your comment about men between 55 and 70 “never benefiting”. The article refers you to one of the very few long-term studies I am aware of that may help to resolve some of the open questions … and it is a study specific to men between the ages of 50 and 69.

  5. The truth is that we need comparative studies that analyze the effectiveness of screening and treatment segmented by age and the morbidity. To my knowledge almost all studies lump all men together. The only exception is the Swedish study (the Goteborg screening trial) that included analysis of the results of screening by age. The conclusion was that the younger cohorts benefit most from screening. This is not a surprising result given the fact that prostate cancer is usually slow growing and that the side effects are less pronounced at a younger age and in a healthier population.

  6. Reuven:

    The Goteborg study did, indeed, break the men down into three age categories. However (once again) all the men in this trial were between 50 and 65 years of age at the start of the trial, so it was of no help in answering the question which is at the heart of the commentary above … and, as we said in the commentary, the chances of us ever being able to carry out a 30-year-long, sufficiently large, screening and comparative effectiveness trial that includes a well-defined opportunity for active surveillance of defined sets of patients at any age seems vanishingly small today (unless someone has solved the debt crisis in the past 30 minutes).

  7. The PIVOT results were presented at the recent AUA conference and will be published in the next few months. This study compared a AS group of over 700 to a matched TX group of over 700 for 12 years and found that low and intermediate risk men did not benefit from treatment while intermediate risk men did.

    The power point presentation is available on the AUA web site.

  8. Dr. Jones is generally correct in his comments about the PIVOT study (see our prior comments on this trial). However, (a) it would be more accurate to describe the control group of patients in this study as being managed with “observation” (i.e., watchful waiting) than “active surveillance,” and (b) a peer-reviewed report on this study has yet to be published. We await that peer-reviewed report with interest.

  9. Your own post “Prostate cancer mortality rates in Denmark: 1998 to 2009” posted on 30 August 2011 05:36 AM PDT stated “To date, no study has categorically demonstrated an absolute extension of life (as opposed to an extension in survival from time of diagnosis) as a consequence of the introduction of the PSA test in the 1980s and improvements in the management of prostate cancer.” Does that not mean that they found no evidence that early diagnosis and primary treatment recover any life expectancy?

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: