A Multi-Institutional Analysis of Adjuvant and Salvage Radiotherapy After Radical Prostatectomy

Reviewer: John Wilson, MD
Abramson Cancer Center of the University of Pennsylvania
Last Modified: October 5, 2004

Presenter: Alan Pollack MD
Presenter's Affiliation: Fox Chase Cancer Center
Type of Session: Scientific


  • Radical prostatectomy is a definitive treatment option for patients with prostate cancer that appears localized to the prostate (T1-2N0M0)
  • Prostate cancer occasionally recurs after surgery, and some patients are recommended to have ART (adjuvant radiotherapy), or radiation before they have any sign of recurrence to prevent metastatic spread
  • Others are treated with SRT (salvage radiotherapy), receiving radiation when there is evidence of local recurrence in the absence of metastatic spread
  • This study was a multivariate analysis to help decide which factors predict the outcome for men treated with ART or SRT
  • The advantage of this study was that it was multi-institutional, unlike most series which have been from a single institution

Materials and Methods

  • 1168 patients were studied who were treated with prostatectomy followed by either ART or SRT
  • A minimum of 12 months follow-up after radiation treatment was required
  • 22% received ART and 78% received SRT
  • 16% of patients also received adjuvant androgen deprivation for a median period of 6 months
  • 34% had negative margins at surgery, 26% had SVI (seminal vesicle involvement), and 70% had pathologic T3-4 disease (cancer extending outside of the prostate capsule at surgery)
  • 49% had an intermediate GS (Gleason Score) of 7, with 28% in the low Gleason group (2-6), and 23% in the high Gleason group (8-10)
  • Patients were defined as receiving ART if their Prostate Specific Antigen (PSA) was below 0.2 and they received radiation within a year of surgery
  • The rest of the patients were placed into the SRT group and were subdivided into: PD-PSA (persistently detectable PSA, 37%) and DR-PSA (delayed rise in PSA, 63%)
  • Androgen deprivation was given to 16%, for a median length of 6 months
  • Biochemical Failure (BF) was defined as a PSA > or = 0.2 after RT (radiotherapy)
  • The covariates that were analyzed were: SRT, negative margin, SVI, capsule invasion, Gleason score, pre-RP PSA, pre-RT PSA, time to BF, RT dose, and PSA doubling time (for the DR-PSA subgroup)
  • The endpoints that were analyzed were: BF, distant metastasis (DM), and overall mortality (OM)
  • Recursive partitioning analysis (RPA) was used to identify prognostic groups


  • Here are the results for the ART and SRT groups:

Median RT dose
60 Gy
64 Gy
Median time from RP to RT
4 months
20 months
Median follow-up
73 months
53 months
5 yr BF rate
5 yr DM rate
5 yr OM rate
10 yr BF rate

  • The median time to nadir after SRT was 8 months, and in 78% the PSA dropped below 0.2
  • The BF curves for the SRT group continue to increase even beyond 10 years
  • Overall, the most important determinants of BF (biochemical failure) were SRT, SVI, high pre-RT PSA, GS 8-10, margin negativity, conformal RT (vs conventional), and pelvic RT
  • For the ART group, the only BF predictor was SVI
  • For the SRT group, the BF predictors were: SVI, pre-RT PSA, GS 8-10, conformal RT
  • Endpoints not significant included RT dose and androgen deprivation
  • The following risk groups were identified using RPA:

RPA Group
> or = 1.0

  • Significant predictors of distant metastases were SVI, SRT, margin negativity, and GS 8-10
  • Overall mortality was predicted by SVI and margin negativity

Author's Conclusions

  • The main predictors of failure after post-RP RT are SRT, SVI, high pre-RT PSA, GS 8-10, and margin negativity
  • Margin negativity was also a predictor of distant metastases and overall mortality
  • RT dose was not related to outcome, which suggests that patients were treated with adequate doses
  • Androgen deprivation was also not significant for outcome, but it was usually given for short periods
  • Group 1 should be treated with RT alone
  • Group 2 should be considered for hormone therapy
  • Groups 3-5 should be considered to not receive RT and just receive long term hormone therapy, and we need new trials to investigate new treatment strategies
  • Extended androgen deprivation may be useful for RPA groups 2 and 3, and groups 4 and 5 require more aggressive systemic therapy

Clinical/Scientific Implications

Dr. Pollack presents here a large multi-institutional retrospective analysis determining risk factors for outcome with adjuvant and salvage radiotherapy.  One of the important things to keep in mind is that these groups are not equal and therefore cannot be compared as such.  The salvage group likely has a worse prognosis because they are the patients who failed from a selected  group of "watch and wait" patients, whereas the adjuvant group was treated shortly after surgery before they had a sign of recurrence.  It should be stated that margin negativity comes out as a strong predictor for biochemical failure, distant metastases, and overall mortality because the PSA rise is more likely to be coming from distant metastases, whereas patients with positive margins can have a PSA rise from residual local disease left at surgery.  Dr. Pollack says pelvic radiotherapy was seen as a risk factor probably because the higher risk patients are assigned to it.  However, since he was using multivariate analysis with many risk factors, if patients receiving pelvic radiotherapy had higher risk factors, that should be controlled for, unless there were other risk factors associate with pelvic RT not in the model.  It would be hard to imagine pelvic RT causing more biochemical failures, although it can cause more morbidity.  Also, it must be kept in mind that one doesn't see a benefit for pelvic RT, at least for definitive patients, unless they are given neoadjuvant hormone therapy as well per the RTOG 94-13 study.  Interestingly, conformal radiotherapy was a risk factor for biochemical failure (as compared to conventional therapy) indicating that perhaps the tighter margins used with conformal radiotherapy are a bit too tight.  Using a low PSA cutoff of 0.2 for biochemical failure might be including those who have residual prostate glands from surgery but have not failed.  However, when Dr. Pollack was asked about this, he said that looking at a PSA cutoff of 0.4 did not change the shape of the biochemical failure curves, but just shifted them down temporally.  The take home message is that salvage radiotherapy appears to be relatively ineffective.  The biochemical failure rate at 10 years is over 80%, and the curves for the salvage radiotherapy group continue to increase beyond 10 years, which suggests that very few people are being cured.  The adjuvant group has a biochemical failure rate of 40% at 10 years, and the biocehmical failure curves do not increase as much, suggesting that some patients are cured with post-prostatectomy radiation.