Alas, … the surgery vs. radiation therapy war continues …

In what certainly appears to be the largest and most thorough meta-analysis published to date, the authors conclude that surgery is a more effective first-line treatment for localized prostate cancer than radiation therapy. However, …

Every retrospective meta-analysis that addresses this topic has two fundamental problems. First, there is no way to allow for selection bias with regard to which patients were treated surgically and which patients were treated with radiation therapy of some type. Second, the precise cause of death is often either uncertain or at least not pathologically confirmed. As a consequence, it is always very hard to know what to make of the results of this type of meta-analysis … however well it has been conducted.

In this particular paper, which is freely available on line as a full-text article (and a thank you to European Urology for that), Wallis et al. report that, based on their detailed meta-analysis:

  • Men with localized prostate cancer (118,830 patients from 15 studies) treated with radiotherapy had an increased risk of prostate cancer-specific mortality compared to those treated with surgery.
  • The adjusted hazard ratio (aHR) for the various sets of there patients were
    • aHR = 2.08 for all patients regardless of risk level
    • aHR = 1.70 for the low-risk patients
    • aHR = 1.80 for the intermediate-risk patients
    • aHR = 1.83 for the high-risk patients
  • Men with localized prostate cancer (95,791 patients from 10 studies) treated with radiotherapy had an increased risk of overall mortality compared to those treated with surgery
    • aHR = 1.63 for all patients regardless of risk level
    • aHR = 1.47 for the low-risk patients
    • aHR = 1.50 for the intermediate-risk patients
    • aHR = 1.88 for the high-risk patients
  • All these results were highly statistically significant.

Basically, what this means is that among all the men in this study, the risk of prostate cancer-specific mortality was about twice as high among the radiation therapy patients compared to the surgical patients, and the risk of overall mortality was about 1.6 times as high.

Now we wish to be very clear that Wallis et al. have carried out an extremely thorough meta-analysis on this highly controversial topic. The authors include two radiation oncologists. And it may well be that their findings really are accurate … but … it would not be correct to conclude, on the basis of this study, that surgery is a “better” form of treatment for all men with localized prostate cancer than radiation therapy, and even if it was, this study takes no account of what the customer (i.e., the patient) actually wants to do about his cancer.

We also wish to be clear that Wallis et al. acknowledge the problem of selection bias in the conduct of this study, stating clearly that:

While the studies used have a potential for bias due to their observational design, we demonstrated consistently higher mortality for patients treated with radiotherapy rather than surgery.

Based on all of the currently available data, The “New” Prostate Cancer InfoLink remains of the opinion that:

  1. Every patient with low-risk, localized prostate cancer should be advised, first and foremost, that some form of expectant management (i.e., monitoring of some type) is a perfectly reasonable first-line form of management that comes with minimal risk for side effects and for prostate cancer-specific death over the next 10 to 15 years.
  2. Any patient started on active surveillance can always have active treatment later on, and in something like 99 percent of patients that treatment can still be given with curative intent.
  3. Every patient with low-risk disease who actually wants to have immediate active treatment should consult (at a minimum) with both a urologist and a radiation oncologist before deciding on treatment.
  4. Every patient who is thinking about having surgery as a first-line option needs to be clearly told by the surgeon about the significant risks for short- and long-term incontinence and for short- and long-term impact on erectile and sexual dysfunction.
  5. Every patient who is thinking about having radiation therapy as a first-line option needs to be clearly told by the radiation oncologist about the risks for gastointestinal and genitourinary side effects (which are not as high as the risks for incontinence associated with surgery) and about the medium- to long-term effects on erectile and sexual dysfunction (which are not usually as debilitating as those associated with surgery).
  6. Every patient who is thinking about having radiation therapy as a first-line option needs to be advised by the surgeon and the radiation oncologist that if the radiation therapy is unsuccessful, second-line surgical treatment is rarely a reasonable or a simple option.

From a patient perspective, we remain optimistic that data from the ongoing ProtecT trial will help to truly resolve this problem once and for all … but it is very possible that it will not. It is never going to be possible to carry out a large enough and a sufficiently well-structured, randomized trial, probably requiring 20 years of follow-up, here in the USA, to answer the question of whether first-line surgery or first-line radiation therapy is actually “better” for the treatment of localized prostate cancer. And it is already clear that, for many, many men with low-risk prostate cancer, the answer is that neither is better than initial active surveillance.

What is needed is not this continuing fight over which therapy is “better” but rather a complete sense of honesty among the members of the clinical community that this continuing attempt to seek a definitive answer to a potentially unanswerable question is entirely ignoring the fundamental needs of patients to be given accurate data that will allow them to make the best possible decision they can for themselves. Sometimes the simplest answer to complex questions is a straightforward, “We just don’t know that.”

7 Responses

  1. I think your assessment of the Wallis meta-analysis is spot on, but either it (or perhaps the meta-analysis) fails to mention a factor that seems to me far more important then the second decimal place in the hazard ratios:

    Which kind(s) of radiation therapy, conducted how long ago, with what kind of equipment and software?

    Any retrospective meta-analysis will have to choose how to weight early results against later results. In other words, if:

    A is the cohort of men treated in 1990–2005 with { open RP vs. (brachy or CRT or IMRT), with/without ADT }


    B is the cohort of men treated in 2005–2015 with { (ORP or RALP) vs. (brachy(plain|boosted) or (old|new)EBRT or SBRT), with/without ADT }

    then should A and B be given equal weight in calculating results in the meta-analysis?

    And the comparison gets even harder if the only endpoints used are risk of death-by-any-cause and risk of death-by-prostate cancer, because the treatment of recurrent prostate cancer has changed enormously over the last 30 years.

    And, finally, the men treated with radiation tended to be older than those treated with surgery, but I see no corresponding adjustment made to the mortality risk (but perhaps I missed it).

    From my vantage point, results from radiation seem to be improving significantly from year to year, whereas results from surgery seem to have plateaued. Therefore, if equal weighting is given to cohorts from 1990 and from 2015, then surgery will seem to be better-by-comparison than is actually the case in 2016 and, presumably, beyond.

  2. Paul:

    With respect to the types of radiation therapy being used, see section 2.3 of the paper, but then you’d also have to look at the individual data (see Tables 1 and 2) to get the dates of treatment.

    You are, of course, correct, that this is another of the problems with this and other meta-analyses in the past. I was deliberately avoiding digging into the details because I was trying to make what I considered to be a rather more important point.

  3. I’m wondering why the Sitemaster thought this problem-ridden analysis was worth two commentaries on this site (see this link). ;-)

    The biggest problem is the inclusion of all those studies (almost all of them, in fact) where the radiation dose was insufficient to kill the cancer. It would be correct to conclude that radiation therapy, as commonly practiced in the last century, was inferior to surgery. That was not a secret: the equipment used then could not safely deliver curative doses, and patients selected were those who could not undergo surgery due to age or comorbidities. As I said before Anthony Zietman’s response was excellent, and can be read in full (at this link) by any who are interested. The only point on which I’d take issue with Zietman is that the ProtecT clinical trial will solve the issue – the radiation dose used was too low by today’s standards.

  4. The Sitemaster didn’t think it deserved two commentaries. He had simply forgotten that he had seen the paper 8 months ago when it was first published on line! It took 8 months to get in to print. :O)

    However, The sitemaster would just point out that Dr. Zeitman and his colleagues did not say the the ProtecT study will resolve this issue. What they actually said was that the ProtecT study would “have far greater meaning and influence” than this or any other meta-analysis (which the Sitemaster agrees with).

  5. As indicated, quality of life factors are quite critical too. Using mortality as an endpoint (no pun) is a given but also overlooks the morbidity of most treatments that some men have and will endure. Also, do we bias other treatment considerations while focusing on just 2 longstanding treatments?

  6. This analysis, on the surface, seems flawed in the sense that it does not define “radiation”. There are many forms of radiation therapy and proton therapy seems to have been ignored completely — though there is much evidence that it is highly effective, with the lowest short- and long-term side effects. Loma Linda was a pioneer in this field and there are now dozens of facilities effectively applying this technology. I completed my proton treatment in 2015 at the Chicago Proton Center in Warrenville, IL, now part of Northwestern Medicine.

  7. Dear Steve:

    The study did not actually “ignore” proton beam radiation therapy (PBRT) at all — any more than it “ignored” stereotactic body radiation therapy. The problem is that no one has ever conducted a high-quality, retrospective analysis of the outcomes of patients treated surgically compared to those treated by PBRT that met the standards for inclusion in this meta-analysis. The study was very clear about the forms of radiation therapy that were included.

    No one has ever suggested that PBRT isn’t effective as a treatment for prostate cancer — and particularly for low-risk forms of prostate cancer. The problems with PBRT have always been: (a) the cost; (b) the degree to which is demonstrably “safer” than other forms of modern, high-quality, carefully targeted radiation therapy; and (c) the effectiveness in men with unfavorable, intermediate-risk and high-risk forms of localized prostate cancer (i.e., the ones who we are quite sure need to be cured early). There is a major trial going on to try to resolve some of these issues, but a lot of patients are unwilling to participate in this trial (and a lot of radiation oncologists are unwilling to recommend enrollment to their patients).

    Part of the problem here is that Loma Linda never did what they told me they were going to do back in 1996, which was collect and publish high quality outcomes data on all of their prostate cancer patients. What they actually published was a rather less than convincing set of data on some of their patients which omitted all sorts of information that would have been needed to truly evaluate the quality of those patients’ outcomes over time. So we now have a dozen PBRT centers around the country that cost us a fortune to build, several of which are actually losing money, and we are still no wiser whether they are actually “better” for treating clinically significant prostate cancer than any other form of radiation therapy.

    PBRT is a very valuable form of radiation therapy for some forms of cancer. Establishment of its actual value in the treatment of clinically significant prostate cancer is something we are still waiting for. Belief is one thing; actual data are quite another.

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: