Whether to escalate radiotherapy dose in high-risk prostate cancer treated with long-term androgen deprivation has been unresolved for two decades, because the trials that tested it were either underpowered or read out too early. GETUG AFU 18 has the follow-up to answer it.
This French multicentre phase 3 trial randomised 505 men with high-risk prostate cancer — PSA 20 ng/mL or above, Gleason 8 or above, or clinical stage T3-T4 — to 80 Gy in 40 fractions or 70 Gy in 35 fractions, both with long-term androgen deprivation. Median follow-up reached 9.5 years.
At the prespecified five-year endpoint the difference was small: progression-free survival 91.4% with 80 Gy versus 88.1% with 70 Gy. At ten years, reported post hoc, the curves had separated substantially — 83.6% (95% CI 77.8 to 88.0) versus 72.2% (65.3 to 78.0), stratified hazard ratio 0.56 (95% CI 0.40 to 0.78, p<0.0001).
Toxicity did not follow the dose. Grade 3 or worse acute events were 24% versus 25%, and grade 3 or worse late toxicity at five years was 8% versus 7%. The commonest late grade 3 event was bladder or urethral disorder, 4% versus 2%. There were no treatment-related deaths.
The interpretive discipline matters here. The five-year endpoint was the prespecified one and it was unimpressive; the ten-year figure is post hoc and would ordinarily be treated with suspicion. What makes it credible is that the direction never changed and the mechanism is plausible — biochemical failure after prostate radiotherapy takes years to declare. The authors are appropriately cautious that cancer-specific and overall survival remain unproven.
- Offer 80 Gy rather than 70 Gy to men with high-risk prostate cancer receiving long-term androgen deprivation, where modern conformal planning is available
- Counsel on the timescale: the difference is small at five years and substantial at ten, so this matters most for men with long life expectancy
- Reassure on toxicity — grade 3 or worse late events were 8% versus 7%, essentially unchanged
- Note that cancer-specific and overall survival benefit remain unproven; the endpoint here is progression-free survival
- The ten-year result is post hoc, and the prespecified five-year endpoint showed only a 3.3 percentage point difference
The statistics, in plain English
The gap between the five-year and ten-year results is the whole story, and it is a lesson in endpoint timing rather than a statistical trick. At five years there had been too few events for a difference to emerge — the trial required 197 events and had not accumulated them — so the prespecified analysis was effectively underpowered by biology rather than by design error. The ten-year hazard ratio of 0.56 with an interval of 0.40 to 0.78 is convincing, but it is post hoc, which means the analysis time was chosen after seeing data. Treat the direction as established and the precise magnitude as provisional.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for oncology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free