Relative 10-year survival after surgery, EBRT, or brachytherapy in the PSA era

A new study on relative rates of survival at 10 years, just published on line, is probably going to get differing reactions from patients and from physicians depending on their individual points of view.

Kibel et al. have reported data from an analysis of the 10-year, overall survival and the prostate cancer-specific mortality rates of more than 10,000 men treated at tweo major institutions with any one of three, standard first-line therapies for clinically localized prostate cancer between 1995 and 2005.

The analysis is based on data from 10,429 consecutive patients treated at either Barnes-Jewish Hospital (associated with Washington University in St. Louis, Missouri) or at the Cleveland Clinic (in Cleveland, Ohio). The actual breakdown of the number of patients treated was:

  • 6,485 received a radical prostatectomy of some type
  • 2,264 were treated with some form of external beam radiation therapy (EBRT)
  • 1,680 were treated with permanent seed implant brachytherapy.

The research team also collected information on the race, age, and comorbidities of all the patients; on and disease-specific parameters (e.g., PSA, biopsy Gleason score, and clinical stage); and on relevant details of the patients’ treatments. It is worth noting up front that this study was carried out under the oversight of Drs. Michael Kattan and Andrew Stephenson of the Cleveland Clinic, who probably have more experience of this type of analysis than almost anyone else conducting research of this type today.

So, without trying to get into a lot of the details, here are the fundamental findings of their research:

  • Adjusted 10-year overall survival of the patients by type of treatment were
    • 88.9 percent for those who had radical prostatectomy
    • 82.6 percent for those who had EBRT
    • 81.7 percent for those who had brachytherapy
  • Adjusted 10-year prostate cancer specific mortality rates for the patients by treatment type were
    • 1.8 percent for those who had radical prostatectomy
    • 2.9 percent for those who had EBRT
    • 2.3 percent for those who had brachytherapy, respectively.
  • Based on propensity score analysis (the type of statistical analysis used in this study)
    • EBRT was associated with decreased overall survival (hazard ratio [HR] = 1.6; p < 0.001) compared to radical prostatectomy.
    • EBRT was associated with increased prostate cancer-specific mortality (HR = 1.5; p = 0.041) compared to radical prostatectomy.
    • Brachytherapy was associated with decreased overall survival (HR = 1.7; p <0.001) compared to radical prostatectomy.
    • Brachytherapy was not associated with increased prostate cancer-specific mortality (HR = 1.3; p = 0.5) compared to radical prostatectomy.

However, the authors go to considerable lengths in the full text of their paper to point out that — even though radical surgery was associated with improved overall survival and prostate cancer-specific mortality compared to EBRT and brachytherapy in this particular study — 

 … the absolute improvement in overall and [prostate cancer-specific mortality] among groups at 10 years was modest.

They further note that,

These survival differences may arise from an imbalance of patient or disease related confounders despite our efforts to adjust for them, from improved cancer control when [radical prostatectomy] is performed as initial therapy, and/or from differences in treatment related mortality.

It seems likely to The “New” Prostate Cancer InfoLink that the “best” way to look at this study is to see it as showing that the overall 10-year survival rates from these three widely used types of first-line therapy (when carried out at apporopriately expert facilities) is high, at > 80 percent, and that the 10-year prostate cancer-s[pecific mortality rates are low, at < 3 percent. This perspective is put into context by the first paragraph of the “Discussion” section of this paper by Kibel et al.:

A man with clinically localized [prostate cancer] is faced with a challenging process of selecting the optimal treatment. He must choose between radical therapy and surveillance for a cancer that may pose an uncertain threat to his longevity or quality of life.If he chooses treatment, none has been proven superior in terms of quantity or quality of life. In the absence of data from randomized trials, one must rely on observational studies, although these may be biased. In this study, to our knowledge the largest comparative analysis of contemporary patients treated according to current treatment standards, [radical prostatectomy] was associated with improved overall survival compared to EBRT and brachytherapy after adjusting for major confounders, and improved [prostate cancer-specific mortality] compared to EBRT. However, the differences in adjusted survival among treatment modalities at 10 years were small, particularly for [prostate cancer-specific mortality].

10 Responses

  1. Let the discussion commence! As a person familiar with EBRT, I would not define 1995-2005 as the ‘modern era’ for that treatment modality. IMRT didn’t come online until late in that period, and further, dose-escalation studies had not yet definitively pinpointed the proper dose of radiation to give. There’s a good chance that right up until the end of that period, prostate cancer patients were receiving a total of 68-74 Gy, when we know today that about 78 Gy is ideal (and indeed is the standard of care). While the dosage delta may seem small, I’m willing to hazard that if 78 Gy using IMRT/PBRT is put up head-to-head against RP, you’ll see absolutely no difference in outcome, either for overall survival or PCa-specific mortality.

    The RP procedure from 1995-2005 is substantially similar to that still in use today (even considering the laproscopic and robotic advances). That’s simply not the case for EBRT, which has come a long way from the 1990s and early 2000s.

  2. Mike was there any breakdown for age medians in these three modality classes?

  3. Tony:

    Only 182 out of 10,000+ patients (< 2%) died of prostate cancer within the follow-up period. I doubt if a breakdown by age would have shown anything significant within that timeframe. If the authors did such a breakdown, it is not reported in the full text of the paper.

  4. Dear Nerf:

    I think you might be surprised to discover just how much the technical quality of most radical prostatectomies improved between 1995 and 2005. And think about it … If the authors were going to get 10-year data, they pretty much had to study this data set. If they’d studied data from 2000 to 2010, they’d only have been able to study a median 5-year survival … and almost no one would have died at all, so the study would have been pretty pointless!

    I think the point the authors are really making is that over 10 years of follow-up it really makes very little difference which form of treatment you select today — assuming you go to a high-quality center where people do actually know what they are doing.

  5. The question I had about age was whether the average age differed between the groups, and if so was this an adjustment factor. One might reasonably expect the age amongst those receiving radiation, and in particular EBRT, to be higher than for RP. The higher the average age, the lower should be the expected 10-year survival rate.

  6. I’ll just add to nerf’s comment about advances in EBRT technology with the point that brachytherapy also advanced as experts learned how to avoid and compensate for initial cold spots from sub-obtimal seed placement or seed movement.

    Thanks for reporting this study.

  7. Rick:

    Yes there were age differences. They are documented in detail in the paper. However, the paper does not go into detail about the way all of the adjustment factors were applied. You’d need to talk to Dr. Kattan or Dr. Stephenson if you wanted to understand that.

    The statistical analysis section of the paper reads as follows:

    “Multivariable Cox proportional hazards regression analysis was used to determine the association of baseline patient and disease specific parameters with overall survival. Fine and Gray competing risk regression analysis was used to determine the association of clinical parameters with PCSM. In an attempt to address imbalances in the distribution of covariates among treatment groups, we used propensity scores to minimize bias from nonrandom assignment of treatments.

    “For the multivariable models internal validation using tenfold cross-validation was performed to obtain an unbiased measure of their performance. Discrimination was quantified using the c-index, and calibration was assessed by visual inspection of plots comparing observed and predicted outcomes.12 All statistical analyses were performed using R v2.8.1 software (R Foundation for Statistical Computing) with additional packages (Design and cmprsk) added.”

    Maybe that will mean something specific to someone like Matt Cooperberg … It’s beyond my pay grade! I would, however, tend to assume that Dr. Kattan and Dr. Stephenson (who have been doing this stuff for years) didn’t “miss” anything as obvious as the age differences or the many other clinical factors that are clearly listed by treatment group in Table 1 of the paper. Ask Dr. Cooperberg to print you a copy. Copyright law prevents me from distributing copies of a PDF of the paper to everyone who probably wants one.


  8. When I got my EBRT 10 years ago, it was because my cancer was higher risk. They said I needed EBRT, because RP would not catch it all. At that time, maybe also today, higher-risk patients had radiation. This aspect may define the higher mortality rate.

  9. What would you expect the 10 year survival and quality of life to be without any treatment?

  10. Barry:

    Have a look at the most recent data from Sweden on watchful waiting. They closely followed 223 patients for more than 30 years. More than half those patients never needed treatment at all. However, it does all depend on the precise initial diagnosis of the patient and how his personal condition progresses over time.

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 )

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: