Does early salvage radiation save lives?


Several studies have tried to address the issue of whether adjuvant radiation actually increases prostate cancer survival in the long term compared to waiting. They all showed that biochemical progression-free survival could be improved by earlier treatment, but it remained questionable whether that would eventually translate into a survival benefit.

After 10 years of follow-up, the randomized clinical trial ARO 96-02 found that neither metastasis-free survival nor overall survival was significantly improved by adjuvant radiation. However, the study was underpowered to reliably detect those results.

After 10 years of follow-up, the randomized clinical trial EORTC 22911 found that adjuvant radiation did not significantly improve overall or metastasis-free survival, although there may be benefit for men less than 70 years of age, or those with positive margins.

There was only one study, SWOG S8794, which, after 12.6 years of follow-up, demonstrated a significant improvement in both overall and metastasis-free survival among patients who had adjuvant radiation compared to patients who waited. However, the difference did not hold up when patients who received adjuvant radiation were compared to those who received salvage radiation. The difference was possibly attributable to the fact that very few patients in the wait-and-see arm ever got salvage radiation.

None of those three studies used radiation doses that are now considered to be adequate for curative adjuvant or salvage radiation.

So, with highly equivocal findings from the best studies we have available so far, how is the patient to make a decision as to whether it is worthwhile to undergo the potential side effects of early salvage radiation? This is the question that Hsu et al. at the University of California San Francisco set out to answer by looking for evidence in their large CaPSURE database. They identified 305 patients who had radiation after surgery, and who had such high-risk features on their surgery pathology report as:

  • Positive surgical margins, or
  • Gleason score 8-10, or
  • Stage T3 or T4

In that group, they found:

  • 76 men who had undetectable PSA and received adjuvant radiation within 6 months of surgery
  • 229 who had salvage radiation after reaching a PSA > 0.1 ng/ml, or after 6 months post-surgery, of whom
    • 180 who had early salvage radiation before PSA reached 1.0 ng/ml
    • 49 who had late salvage radiation after PSA reached 1.0 ng/ml

After a median elapsed time of over 6 years after surgery, the researchers found:

  • Overall, 98 percent were still alive.
  • Overall, 12 percent had progressed to metastases or death.
  • Adjuvant and salvage radiation patients had comparable high-risk features.
  • Men who had salvage radiation had an all-cause mortality rate 2.7 times higher than men who had adjuvant radiation.
  • Men who had salvage radiation had a prostate cancer-specific mortality rate 4.0 times higher than men who had adjuvant radiation.
  • 10-year estimated prostate cancer-specific mortality was
    • 12 percent among men who had adjuvant radiation.
    • 16 percent among men who had early salvage radiation.
    • 29 percent among men who had late salvage radiation.
  • Late salvage radiation carried increased risk of prostate cancer-specific mortality and all-cause mortality.
  • Early salvage radiation had about the same mortality risk as adjuvant radiation.

The conclusion is that early salvage radiation, while PSA is still below 1.0 ng/ml, had the same survival benefit as adjuvant radiation, but without the risk of over-treatment. However, waiting until after PSA reached 1.0 ng/ml significantly increased the risk of metastases and prostate cancer mortality.

The PSA threshold used in this study, 1.0 ng/ml, is quite high and well beyond the limit used for the definition of biochemical recurrence (0.2 ng/ml). It may turn out to be the case that this higher threshold is a more useful definition of biochemical recurrence than the current definition. However, the follow-up period here is short (median 6 years), as it may require 20 or 30 years for a survival benefit to show up when a lower threshold is used. It should also be noted that the definition of “adjuvant” radiation used in this study included therapy begun any time before PSA reached 0.1 ng/ml, while there may be greater survival advantage when adjuvant radiation is begun at a lower level measured on an ultrasensitive PSA test.

While this study provides evidence for early salvage radiation, its retrospective nature makes it subject to selection bias: there may have been specific reasons why the patients were selected to receive adjuvant, early salvage, or late salvage radiation. The abstract of the study makes no mention of the radiation doses used, whether androgen deprivation was used along with the radiation, the use of radiation to pelvic lymph nodes, or whether age and co-morbidities were significantly different. We await the results of ongoing randomized clinical trials to provide more reliable information.

Editorial note: This commentary was written for The “New” Prostate Cancer InfoLink by Allen Edel.

4 Responses

  1. Good summary of an important issue in light of the most recent retrospective study by Hsa et al. (2015).

    I like the author’s definition of aRT which specifies both timing and PSA level. Studies to date have not been consistent about defining aRT vs early sRT. For instance, of the three RCTs cited, ARO 96-02 trial is the only trial in which all patients had an undetectable PSA at the time of RT. This makes a broad interpretation difficult.

    I am surprised the authors did not find more survival advantage to the aRT cohort, which likely includes a higher percentage of patients that would never have progressed without treatment than the sRT cohorts.

    The distinction between early and late sRT is always going to be a bit arbitrary, and the authors picked 1.0 ng/ml. In this day and age, it would be very difficult for an informed patient to choose to wait till 1.0 ng/ml to initiate sRT. He would have to ignore the advice of most clinicians, and would have to accept the risk that he might be missing his last chance for a “cure”. Some support might be found in the available literature (e.g., Amling et al., 2001) for a somewhat lower break point of 0.4 or 0.5 ng/ml, for an appropriate patient who decides to wait past currently defined BCR (0.2 ng/ml).

    Many of us will have to make our personal decision without the benefit of results from the ongoing RCTs, but can learn much from good blogs such as this one.

  2. The CaPSURE database is multi-institutional, which is a benefit in that it does not reflect the particular characteristics of the patients seen and treatments given at a single institution, but the opposite side of that coin is that the testings and the treatments are not uniform. I don’t know how many of the salvage treatments were given using the dose-escalated radiation doses that are now considered to be effective — a recent quantum improvement in our knowledge. They report only normal PSAs with a detectable threshold of 0.1 ng/ml, so, as you rightly point out, many of the patients considered “adjuvant” in this study, might be considered “early salvage” in a study that used ultrasensitive PSA tests.

    I agree that few competent radiation oncologists would recommend patients with high-risk features wait until their PSA reached 1.0 ng/ml before initiating sRT. I think we have to look at 1.0 ng/ml as maximum rather than a reasonable or optimal PSA threshold. Even the 0.1 ng/ml as a lower limit is suspect because of the short follow-up of 6 years — a slowly growing cancer will take a lot longer than that to become lethal. All we can infer (subject to confirmation someday by randomized clinical trials) is that already, by 6 years, treating before PSA reaches (at most) 1.0 ng/ml affords confers a survival advantage.

  3. Agreed on all counts. The research by Hsu et al. is pretty much consistent with the major findings of SWOG 8794. It seems that there is reasonable evidence supporting survival advantage of RT when administered prior to 1.0 ng/ml.

    In my very humble layman’s opinion, AUA/ASTRO guidelines may have strayed a bit from their evidence-based methodology when they cited grade A evidence in favor of aRT, when the studies were not powered to discriminate between aRT and sRT as they are now commonly defined. Likewise for citing grade C evidence for their definition of BCR at 0.2 ng/ml, when there is no such evidence. As you correctly point out, even 0.1 is suspect. I credit them for trying to make sense out of a large and unruly data base and translate this into practical guidelines, but this remains a grey area.

  4. Thanks for posting this article, and thanks for the comments.

    Regarding the first three studies, even including the last one that showed some degree of benefit for adjuvant radiation, follow-up of 10 years or even 12.6 years seems just too short, as you indicated, though counting metastasis helps give an earlier sign of ultimate results. The American Cancer Society is now reporting survival of nearly 100% at the 10-year point, and a remarkable 94% at the 15-year point for all patients, with all approaches. With success like that, I’m looking for extended follow-up with these trials to buttress the results of the retrospective trial.

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: