PACE-A (sexual function, 5-year) NCT01584258
ForLocalised prostate cT1-T2, Gleason ≤3+4, PSA ≤20, surgical candidates, no ADT
TL;DR5y median EPIC-26 sexual score change -42 after prostatectomy vs -17 after SBRT (p=0.02); IIEF-5 median 5 vs 19 (p<0.001).
No ADT was permitted, so the SBRT trajectory (median EPIC-26 79 to 58) is attributable to 36.25Gy/5 fractions alone, not a hormonal confound. By the trial's own p<0.01 rule only IIEF-5 (median 19 vs 5) is positive, and clinician-graded ED was similar (27% vs 35%). Counselling data, not a claim of preserved function.
In men with cT1-T2, Gleason ≤3+4 localised prostate cancer who are prostatectomy candidates and weight erectile function heavily, this informs the SBRT vs surgery discussion with randomised 5-year PRO data; it does not extend to disease treated with ADT or address cancer control.
No ADT was permitted, so the SBRT trajectory (median EPIC-26 79 to 58) is attributable to 36.25Gy/5 fractions alone. By the trial's own p<0.01 rule only IIEF-5 (median 19 vs 5) is positive, and clinician-graded ED was similar (27% vs 35%). Counselling data, not a claim of preserved function.
Median EPIC-26 sexual score fell from 62 to 17 after prostatectomy, with 18/22 (82%) crossing the MCID. Approach was mixed (open, laparoscopic, robot assisted) and nerve-sparing use is not reported, so the figure may not reflect a contemporary nerve-sparing robotic series. Analysis was by treatment received.
Also covered Sep 28
12 details
PACE is a phase III open-label multiple-cohort RCT; PACE-A randomised 1:1 to SBRT or prostatectomy. 123 randomised (60 prostatectomy, 63 SBRT) between August 2012 and February 2022. Sexual outcomes analysed by treatment received.
Early localised prostate cancer, cT1-T2, Gleason ≤3+4, PSA ≤20 ng/mL, and suitable for prostatectomy. Baseline median EPIC-26 sexual domain score was 62 in the prostatectomy group and 79 in the SBRT group.
36.25Gy/5 fractions delivered in 1-2 weeks. ADT not permitted. Platform, margins and target volume not stated in source.
Comparator was prostatectomy by open, laparoscopic or robot-assisted approach. Nerve-sparing status not stated in source.
Co-primary (previously reported): 2-year bowel and urinary symptoms. Secondary, reported here through 5 years: EPIC-26 sexual domain score, proportion with MCID (≥11-point reduction), IIEF-5, clinician-reported CTCAE grade ≥2 erectile dysfunction. Significance threshold p<0.01.
All patient-reported measures favoured SBRT in direction; only IIEF-5 met the prespecified p<0.01 threshold. Clinician-graded erectile dysfunction was similar between groups.
| Endpoint | SBRT | Prostatectomy | p |
|---|---|---|---|
| EPIC-26 sexual domain, median baseline → 5y | 79 → 58 | 62 → 17 | n/a |
| EPIC-26 median change from baseline | -17 (IQR -43 to 0) | -42 (IQR -58 to -24) | 0.02 |
| MCID (≥11-point reduction) | 19/35 (54%) | 18/22 (82%) | 0.03 |
| IIEF-5 median | 19 | 5 | <0.001 |
| CTCAE G≥2 erectile dysfunction | 13/48 (27%) | 13/37 (35%) | 0.48 |
ProtecT randomised prostatectomy against external beam RT given with ADT, so its sexual-function comparison carried a hormonal confound that PACE-A avoids. PACE-A is described by its authors as the first randomised comparison of SBRT and prostatectomy with long-term sexual outcomes.
Baseline sexual scores were unbalanced (median 62 vs 79), and the MCID analysis rests on 22 and 35 men. Surgical approach was mixed and nerve-sparing use is not reported. Accrual ran from 2012 to 2022, so technique on both sides spans a decade.
The direction is consistent across three patient-reported measures, which carries more weight than any single p-value here. SBRT still produced a decline (median change -17), so the result is a relative advantage, not preserved function. Cancer control is not reported in this abstract.
CONSORT flow
Secondary PRO endpoint, open-label, analysed as treated; EPIC-26 and MCID p-values miss the trial's own p<0.01 threshold. PRO return 60% and 63% at 5y.
- Does the sexual-function gap hold against nerve-sparing robotic prostatectomy alone?
- Does SBRT erectile function keep declining beyond 5 years?
- Oncologic outcomes for SBRT vs prostatectomy in PACE-A
📚 Sources · 📄 1 paper
Abstract
The longer read
The value of this report is randomisation. Sexual function after surgery versus radiotherapy for localised prostate cancer has mostly been compared in observational cohorts, where the men who choose surgery differ at baseline from those who choose radiation, and in ProtecT, where the radiotherapy arm received ADT and so carried a hormonal confound on exactly this endpoint. PACE-A removes that confound by design. ADT was not permitted, so whatever happened to erectile function in the radiation arm reflects 36.25Gy in 5 fractions and nothing else.
The direction of the result is what most readers would predict. The confidence it deserves is lower than the abstract's language suggests. The authors set p<0.01 as their threshold for statistical significance and then describe sexual function as significantly worse after prostatectomy on the strength of an EPIC-26 change with p = 0.02 and an MCID comparison with p=0.03. By the trial's own rule only the IIEF-5 difference (p<0.001) qualifies. The clinician-graded endpoint did not separate at all (p=0.48). That discordance is familiar: CTCAE erectile dysfunction grading is a blunt instrument next to a validated questionnaire. It still means one of four measures is formally positive.
The analysis was by treatment received, not by allocation. In a trial asking men to accept a coin flip between an operation and five outpatient fractions, departures from allocation are not random. In the prostatectomy arm 50 men had the operation, 3 had chemoradiotherapy and 6 withdrew; in the SBRT arm 60 were treated as allocated. An as-treated comparison in that setting loses part of the protection randomisation was supposed to provide. Baseline medians of 62 and 79 show the analysed groups already differed on the outcome itself. Change from baseline partly addresses that, and a lower starting score leaves less room to fall, which if anything works against the surgical arm showing the larger decline. Participant-reported data at 5 years came from 30/50 and 38/60, and the MCID analysis from 22 and 35 men. Whether the non-responders were the men doing well or badly is unknown, and it could differ by arm.
Transferability has its own gaps. Accrual ran from August 2012 to February 2022, a long window for 123 men, which points to a selected population. Prostatectomy could be open, laparoscopic or robot assisted, and the abstract is silent on nerve sparing. On the radiation side, platform, margins and any vessel- or bulb-sparing constraints are not given. A median EPIC-26 of 58 after SBRT is still a fall from 79. Radiation-related erectile decline is typically gradual while surgical decline is immediate with partial recovery, so the gap at 5 years is one cross-section of two different curves.
What this changes is the counselling conversation for a surgical candidate with low or favourable intermediate risk disease for whom erectile function is a deciding factor: there are now randomised patient-reported data, free of ADT, that favour SBRT. It leaves cancer control untouched, since none is reported here, and it has to be read alongside the bowel and urinary co-primaries reported earlier. It would be wrong mainly if missing responders differed systematically between arms.