The ESCAPE project — who gets early metastasis and why?


Researchers at the Department of Urology at Erasmus University Medical Center in Rotterdam appear to have initiated a new — and potentially international — initiative out of findings from the European Randomized Screening Study for Prostate Cancer (ERSCP). The new initiative is known as the ESCAPE project.

It has long been known that despite widespread application of efforts to ensure early detection of prostate cancer and the application of curative therapy, a subset of men are commonly identified who already have progressive disease and even metastatic disease at time of diagnosis, even if they have undergone prior testing with PSA and DRE for risk of prostate cancer. The long-term goal of the ESCAPE project is to improve methods of prostate cancer testing in order to better identify such patients and to treat them as early and as appropriately as possible in the hope of avoiding risk for metastasis and subsequent prostate cancer mortality.

In what appears to be the first publication to come out of the ESCAPE initiative, Zhu et al. have attempted to characterize “escapes” from within the participants in the Dutch (Rotterdam) section of the ERSCP. They define “escapes” as men who developed metastasis and/or died from prostate cancer specifically after being tested according to the protocol specified for participants in the Rotterdam section of the ERSCP.

A retrospective analysis of data from the Rotterdam section of the ERSCP showed the following:

  • A total of 21,210 men were initially randomized to the screening arm of the study.
  • 19,950/21,210 men actually received screening according to the study protocol.
  • 2,317/19,950 screened men (11.6 percent) were diagnosed with prostate cancer.
  • Of these 2,317 cancers
    • 1,946 (84.0 percent) were detected at a screening round.
    • 371 (16.0 percent) were detected during intervals between screening rounds.
  • Average (median) follow-up was 11.1 years for the whole cohort and 7.3 years for men diagnosed with prostate cancer.
  • There were 168 “escapes” among the 2,317 cancers, which represents 0.8 percent of the total screened cohort of 19,950 men.
  • 94/168 “escapes” (56.0 percent) were identified in the initial screening round.

The authors offer a variety of possible reasons to explain why individuals might have become “escapes,” including the following:

  • Non-attendance at specific rounds of screening
  • Inadequacy of the screening tests and processes
  • The relative long screening interval (4 years in the Rotterdam section)
  • The age cut-off at 75 years
  • Under-treatment after identification of risk and subsequent diagnosis.

The authors draw no conclusions from these data since their primary objective was to initiate discussion and build consensus about the size of the potential problem. Indeed, their primary request is for cooperation with other groups in order to be able to compare data on the “escapes” from the Rotterdam section of the ERSCP with similar data from other screening cohorts that might have different and more aggressive screening strategies. Examples of such cohorts might be those from the Göteborg screening cohort in Sweden (who were tested every 3 years) and from the PLCO group in the USA (who were tested annually).

The authors initial goal is to make improvements to current screening algorithms (presumably based on specific risk characteristics) that can lead to a decrease in risk for metastasis  and for prostate cancer-specific mortality without any unnecessary increase in diagnosis and treatment.

5 Responses

  1. I’m glad the researchers are working the data from the trial to gain insight into early metastasis.

    This abstract also helps us toward an insight into the original report from the ERSPC. The follow-up information they present is as follows from Sitemaster’s article: “Average (median) follow-up was 11.1 years for the whole cohort and 7.3 years for men diagnosed with prostate cancer.” This may help answer a key question regarding the ERSPC results initially published in the New England Journal of Medicine: How long was the follow-up for the patients in the screening arm who were diagnosed with prostate cancer?

    One of the key criticisms of the ERSCP results was that the follow-up was too short to get at the likely impact of screening on survival. Based on the clue from this study, we have an indication that follow-up for diagnosed men in the screening group after their diagnosis was about 4 years less than follow-up for the entire group (11.1 years minus 7.3 years = 3.8 years. Of course, these figures come from only one part of the ERSCP results, the Dutch cohort, so they may not be representative of the whole, but that group of screened men was large at 21,210 (or 19,950 who actually received screening as stated above).

    As published in the original report, the median follow-up was 9 years for all men in the entire ERSPC. Therefore, if we subtract 3.8 years from 9 years, we get 5.2 years as a rough approximation of follow-up post-diagnosis for the initial ERSPC report. That is extremely short follow-up for assessing true survival impact of screening versus non-screening as a very high percentage of patients will be surviving at that point. As a point for comparison, in the US, survival at the 10-year point for low- and intermediate-risk patients is virtually 100%. Even so, with much bigger gaps between screening than in the US and with some contamination, some benefit for screening was indicated even in the very preliminary ERSPC results as initially published.

    As the ERSPC researchers report progressively more mature results, we are seeing an expanding benefit for screening, as we have expected.

  2. Dear Jim:

    Respectfully, you can’t “do statistics” like that.

    For example, the “survival at the 10-year point” in the PCLO trial was a statistical estimate. The median follow-up on men in that study when reported was actually only about 7 years.

    Also, you can’t subtract the 3.8 year difference from the Rotterdam group from the 9 year average of the entire ERSPC cohort. You are mixing apples and oranges.

    Mike

  3. Jim:

    In addition, I am not aware of any “more mature” data from the ERSPC than that originally published in the NEJM back in early 2009. There has been no subsequent analysis based on longer follow-up as yet.

    All subsequent data from the ERSPC to date has been subset analysis of the 2009 data. It is therefore post hoc analysis. Any implied “expanding benefit” of screening in such publications is not actually based on new data. It is based in interpretation of the original data, which is open to question because it is based on subjective assumptions.

  4. I concur that you cannot get valid statistics from the figures in the first response, but I believe you can use such numbers to do a tentative rough estimate. Given the limitations of research and our need to make decisions, those tentative rough estimates are often the best we have to go on at the time we need to make decisions.

    I thought longer follow-up had been used in some of the post-analysis reporting, but I see I was over-enthusiastic on that point, mistaking the sub-analyses for longer follow-up. Do you know if such more mature follow-up is planned?

    Thanks to you for keeping a cool head and being accurate.

  5. Jim: It is not completely clear at this time whether longer follow-up will be available for either the ERSPC or the PCLO study. I would like to think it would be, but there have been no definitive statements from either group that I am aware of.

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: