onc brain

The longer read

The useful content in this session is not the SBRT-versus-conventional comparison, which PACE-B has already settled to most readers' satisfaction, but the argument about why SBRT's genitourinary signal keeps reappearing and whether it is intrinsic. Two independent lines of evidence in the talk point the same way. PACE-B reported grade 2+ GU events in 5.4% of SBRT patients against 3.7% with conventional or moderately hypofractionated RT, a difference that did not reach significance (p=0.28) but is directionally consistent with the pooled series, where late grade ≥3 GU reached 2.0% (95% CI 1.4-2.8%) and dose correlated with worse late GU (P=.014) at the same time as it correlated with better biochemical control (P=.018). Read naively, that is a therapeutic-ratio ceiling: push dose, buy control, pay in urethra.

The counter-argument offered here is that both datasets share a planning deficiency rather than a biological one. PACE-B did not require urethral contouring, and where the urethra was contoured the constraint applied only if it was visualised, with a target of V44Gy under 20%. Many patients had no MRI to delineate it. If a plan carries hotspots of 120% or more, the urethra can see 48Gy or above, roughly 121Gy EQD2 by the speaker's calculation, which is well outside what anyone would accept deliberately. The same critique extends to the meta-analysis, whose component prospective series largely predate routine urethral constraints. Whether this fully explains the GU excess is not established by anything presented, and it should be read as a hypothesis with a clear test rather than a finding: contemporary SBRT cohorts planned with mandatory MRI-based urethral delineation and a hard constraint should show the excess shrink or disappear. Until such a comparison exists, a reader is choosing between two plausible readings of the same numbers.

What is more solidly grounded is the ADT question, and it is worth separating from the fractionation debate entirely. The RTOG 9408 secondary analysis is the cleanest thing cited: in unfavorable intermediate-risk patients, four months of ADT improved distant metastasis (HR 0.48, 95% CI 0.28-0.83, P=.008) and prostate cancer-specific mortality (HR 0.40, 95% CI 0.26-0.60, P<.001), while favorable intermediate-risk patients gained nothing. That is a post-hoc stratification of a trial designed before the favorable/unfavorable split existed, so the effect sizes should be treated as generous, but the direction is consistent with how the field has behaved since. It also establishes that the risk-group label is doing real work, given that grade group 3 alone carries an HR of 3.49 for distant metastasis.

The forward-looking claim, that the field is moving toward whole-gland de-escalation with a boost to the dominant nodule, rests on early feasibility data rather than comparative outcomes. A phase 1a/b design escalating the dominant lesion to 45, 47.5, and 50Gy over five fractions on a 36.25Gy base reported no grade 3 or higher toxicity, and Loblaw's regimen of 35Gy to prostate, 25Gy to pelvis, and 50Gy to the lesion produced urinary and bowel quality of life comparable to unboosted cohorts. Both are safety signals, not efficacy ones, and neither answers whether de-escalating the uninvolved gland costs control. The FLAME 2.0 and FORT trials are the relevant tests, and this excerpt does not carry their results.