When access was equalised by trial protocol, the racial outcome gap reversed

Design

Secondary analysis of a randomised controlled trial of imaging-guided salvage radiotherapy

Population

157 post-prostatectomy men with detectable PSA completing per-protocol salvage radiotherapy (55 African American, 102 of other races)

Primary outcome

Failure-free survival at up to 48 months

Effect

Across arms 72.8% (95% CI 53.8–85.0) vs 58.7% (46.6–68.9), P = .002; in the PET-guided arm 81.5% vs 73.0%, P = .131

Post-prostatectomy men with detectable prostate-specific antigen were randomised to salvage radiotherapy guided by conventional imaging or by 18F-fluciclovine PET/CT, and followed for up to 48 months. This analysis asked how race, socioeconomic status and metabolic dysregulation affected failure-free survival within that trial.

Across both arms, African American men had higher failure-free survival than men of other races: 72.8% (95% CI 53.8–85.0) against 58.7% (46.6–68.9), P = .002. In the conventional imaging arm the gap was 64.0% versus 45.3% (P = .008); in the PET-guided arm both groups improved and the gap narrowed to 81.5% versus 73.0%, no longer significant (P = .131). African American men had lower socioeconomic status scores and higher allostatic load in both arms.

The direction is the opposite of what population data on prostate cancer outcomes show, and the authors attribute it to a setting that controls for access to care. That is a plausible reading and a limited one: 55 African American men across two arms is a small number, the confidence intervals are very wide, and a per-protocol analysis of a trial not designed for this question cannot separate access from selection into the trial.

• Do not read a worse population-level outcome as a biological one — this trial points the other way when access is held constant.
• Offer molecular imaging-guided salvage radiotherapy where available; both groups did better in that arm.
• Treat allostatic load and socioeconomic barriers as modifiable parts of cancer care, not background noise.
• Note the small numbers: 55 African American men in total, with confidence intervals spanning more than 30 percentage points.
• The specific racial categories here are US ones and do not transfer to Indian practice; the underlying point about access does.

Why it matters

Outcome gaps attributed to tumour biology look different when the trial protocol removes the access differences.

Don't overread it

This is a secondary per-protocol analysis of a trial designed to compare imaging strategies, not to compare outcomes by race.

The statistics, in plain English

The confidence interval for African American men in the conventional imaging arm runs from 34.4% to 82.9% — a span so wide that the point estimate of 64.0% carries little weight on its own. The P value of .008 reflects the comparison between groups rather than the precision of either estimate. With 29 and 26 men in the two arms, this analysis can suggest a direction and cannot establish a magnitude.

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.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.