Who actually dies from prostate cancer? Additional observations

Earlier today we reported on a recent publication based on data from the CaPSURE registry database, which stated that, among men enrolled in that database, men died sooner from their prostate cancer after they progressed to having metastatic disease over time (median survival, 2.4 years from onset of metastasis) than died from their prostate cancer if they were initially diagnosed with metastatic disease (median survival, 5.3 years).

We have now been able to obtain and read the full text of the original article by Borno et al. and we have also been able to do additional research. In our opinion there are some serious issues that will have to be addressed before anyone can generalize from the data presented in this article and apply this thinking more broadly.

First and foremost, we would point out that, in the paper by Borno et al., their research database included just 303 men initially diagnosed with metastatic prostate cancer out of a total of 15,209 patients in the registry (of whom 14,730 were eligible for data analysis according to the criteria applied by the authors). In other words, just 2.1 percent of the patients in the data set used by Borno et al. were actually diagnosed with metastatic prostate cancer and were eligible for inclusion in the study. Even if one were to add in the 62 patients who might have been diagnosed with metastatic disease but were ineligible for inclusion in the study for other reasons, the highest possible percentage of patients enrolled in the registry who were initially diagnosed with metastatic disease would be 365/14,730 or 2.5 percent. However, ….

According to the nationally accumulated data from the SEER database, and used each year by the American Cancer Society to project the numbers of diagnoses of and deaths from prostate cancer each year, the percentage of all the men diagnosed with prostate cancer in any one year who are initially diagnosed as having distant metastatic disease is currently of the order of 6 percent — roughly twice to three times higher that the percentage of such patients represented in the CaPCURE database. This would appear to represent selection bias. One strong possibility is that the design of the CaPSURE study, which was implemented primarily by the urology community, failed to include a representative selection of men being initially diagnosed with metastatic disease because those patients would potentially have been treated by medical oncologists as opposed to urologists. This possibility is further confirmed by the fact that just 16 patients in total were enrolled in this registry as being initially diagnosed with TxNxM1c disease (i.e., having metastatic disease to organs other than bone). And we should note that the authors themselves are careful to point out that selection bias is one of the possible limitations of this study.

A second issue that is utterly unclear is exactly when men initially diagnosed with localized forms of prostate cancer who progressed to having metastatic disease were started on androgen deprivation therapy (ADT) after failure of localized forms to treatment. There is therefore a strong possibility that a high percentage of these men were being treated with ADT long before they showed evidence of distant metastasis and that they may well have become castration-resistant before they had metastatic disease. This, in and of itself, would bias the primary outcome of this study, because it would mean that the two sets of men would have started to receive treatment with ADT based on very different criteria.

The ways in which we treat both early stage, favorable-risk forms of localized prostate cancer and high-risk forms of potentially progressive prostate cancer (not to mention prostate cancer that is metastatic at the time of initial diagnosis) have been changing relatively rapidly over the past 20 years. However, with the exception of some pretty small subsets of men with very specific genetic mutations, there is still very limited, accurate knowledge about how to select the most appropriate form of treatment at a specific point in time for an individual patient. What is more, there is no really substantive evidence to suggest that — despite all the progress — we have really extended the median survival times of men with metastatic prostate cancer from, say, 3 years to 6 years.

Back in February 2018, we noted that Francini et al. reported that, even at the Dana-Farber Cancer Center in Boston, changes in the treatment of men with metastatic, castration-resistant prostate cancer (mCRPC) had only increased their median survival from 2.2 years (in 2004 to 2007) to 2.8 years (in 2010 to 2013). One could hardly expect the average community urology practice to be getting those sorts of results in most of their patients.

The bottom line to all of this is that while the data from the CaPSURE database is intellectually interesting, what it really shows us is that we still have a very long way to go before we can make new, and dramatic forms of impact on the outcomes of most men diagnosed with or progressing to have metastatic forms of prostate cancer. We rarely know “the best” way to treat any of these patients (based on sound prognostic information), and even if we do the expectations for long-term outcomes are still not good.

Now there are exceptions to every rule, and there are certainly some men who do really well even after becoming castration-resistant and metastatic. However, what we still badly need are treatments that can induce much longer survival times (with lower risks for adverse events) for most men with advanced disease, as well as much more accurate prognostic methods to be able to tell an individual patient that he is most likely to respond much better to treatment X as opposed to treatment Y once he has metastatic disease.

In the meantime, we suspect that the idea that most men who are initially diagnosed with metastatic, castration-sensitive prostate cancer are going to live twice as long (on average) as otherwise comparable men who have just progressed to have mCRPC may not be entirely justifiable outside the boundaries of the CaPSURE registry study.

3 Responses

  1. This data set was taken between 2009 and 2016. There is now evidence of increasing levels of advanced prostate cancer and certainly anecdotal evidence of more de novo metastatic disease being diagnosed at a younger age.

    Could this impact the difference between +/- 2% and 6% today?

  2. Lead time bias … those already with cancer have to live long enough to survive to be enrolled … likely have less aggressive cancer

  3. Excellent study analysis, commentary, and responses (so far). Yes, we have a long way to go ‘till true precision medicine in prostate cancer. Practitioner judgment still a huge factor in our survival (and QOL remaining until whatever cause of individual patient demise).

    Paper truly demonstrates limitations of retrospective observational studies. One hopes Dr. Borno recognizes the value of the advocacy community’s observations. She’s got a fine future in more definitive prostate caancer research (as does Allen in advocacy!)

    All good — as seriously exemplary of how challenging characterizing, diagnosing and treatment sequencing as well as new treatment protocols (and better diagnostic tools) are to refining understanding of prostate cancer.

    Is the “natural history” of prostate cancer largely disguised by our diagnosis/treatment environment in the modern world (except perhaps in third-world countries)?

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: