onc brain

The longer read

The main message for a radiation oncologist is that abiraterone works on top of the local therapy, not instead of it. RT was planned in 85% of pts and mandated for node-negative disease, so the pooled population is largely men getting prostate RT with 3 years of ADT. The prespecified subgroup split by planned RT gives nearly identical MFS hazard ratios (0.54 with RT, 0.51 without, interaction p=0.67). That argues against reading systemic intensification as a replacement for RT or as a reason to shorten ADT. The trial did not test either of those, and the backbone stayed fixed.

The primary endpoint change deserves scrutiny because it is the obvious line of attack. The trial group moved the non-metastatic primary from OS to MFS and pooled the two trials. The authors say this was decided before any efficacy was inspected by arm and was published as a declaration beforehand, which removes most of the concern about outcome switching. It also turns out not to matter much, because OS improved as well (HR 0.60, 6-yr 86% vs 77%). A surrogate-only result would have faced the usual question of whether MFS predicts survival. This one does not have to rely on that argument.

The enzalutamide comparison is indirect: the interaction HR was 1.02 across two separately randomised trials, not a head-to-head test. The confidence interval (0.70-1.50) cannot rule out a small benefit, as the authors say. But grade 3+ AEs nearly doubled in the enzalutamide trial versus its control, so the toxicity cost is clear. For practice this is a de-escalation read: abiraterone + prednisolone, not the doublet.

Three points should temper how far the reader extends this. First, staging was conventional imaging. A population restaged on PSMA-PET would move some of these men into oligometastatic categories, and the effect size in that more selected group is unknown. Second, the pelvic RT volume for node-positive men is not described in the source. Only 71% of N+ men had RT planned at all, so the N+ subgroup result (HR 0.49) is not a clean answer on nodal RT combined with intensification. Third, the performance status interaction (HR 0.86 in PS 1-2) and the unexplained NSAID/aspirin interaction are both fragile subgroup findings. The authors advise against acting on the latter. The former is a reasonable prompt to weigh competing mortality in older or comorbid men.

The open-label design matters more for OS than it might seem. Control pts could receive second-generation hormonal therapy for non-metastatic castration-resistant disease, either on trial or, after 2019, as standard care. That would pull the OS estimate toward the null, so the survival benefit is more likely understated than overstated. Data on subsequent therapy were unreliable, so this cannot be checked.

The decision this moves is systemic intensification alongside definitive RT for high-risk N0 and N+ disease. The questions it leaves open are the benefit after prostatectomy, the optimal duration beyond the pragmatic 2 years, and the benefit in men relapsing after prior local therapy, who were underrepresented.