onc brain

The longer read

The question BART answers is narrower than the one the field usually argues about, and that is its strength. Adjuvant radiotherapy after cystectomy has been dismissed for decades on the reasoning that failure in muscle-invasive bladder cancer is overwhelmingly distant, so treating the pelvis buys toxicity and nothing else. The trial takes that premise apart from both ends. It shows a 26% locoregional recurrence rate in the observation arm, which is not a rounding error against the 37% overall recurrence rate, and it shows that irradiating the cystectomy bed and pelvic nodes cuts that to 8%. RT controlled the field it treated. Nobody should be surprised by that, and the honest reading is that the historic dismissal rested on an empirical claim about failure patterns that was simply wrong for the high-risk subset.

Where confidence should be tempered is the leap from local control to survival, and here the trial is silent rather than negative. Eight years of accrual produced 153 patients against a larger target, and the OS confidence interval (0.49-1.26) spans a substantial benefit and a modest harm. Reading the numerically favorable BCSS and DFS as directional support is reasonable; reading them as evidence is not, and the investigators' own framing (an individual patient data meta-analysis as the path to an OS answer) concedes as much. The per-protocol HR of 0.27 deserves the same caution for a different reason: 14 of 77 patients allocated to radiotherapy never received it, and four of those progressed before starting, so the per-protocol arm is enriched for patients whose disease behaved well enough to let them get treated. The intention-to-treat HR of 0.43 is the number that should travel.

The more consequential limitation for a 2026 reader is the control arm's vintage. No patient in either arm received adjuvant immunotherapy, and adjuvant nivolumab is now standard for many of exactly these patients. That does not invalidate the result, because the failure mode nivolumab addresses and the one radiotherapy addresses are not the same, and there is no mechanistic reason systemic immunotherapy should sterilize a positive margin or an incompletely dissected obturator basin. But it does mean the absolute benefit of adding radiotherapy on top of contemporary systemic therapy is unmeasured, and if immunotherapy reduces the competing risk of distant failure, the value of locoregional control could plausibly rise rather than fall.

Generalizability cuts in a specific direction. A median age of 57, a median of 20 nodes dissected, and a 4.6% margin-positive rate describe a young, fit, surgically well-treated cohort. Older patients with less thorough lymphadenectomy carry more residual pelvic disease, which argues the benefit could be larger, but they also tolerate pelvic irradiation less well, and the 17.5% grade 2 GI rate is the number that would grow first. The 28% variant histology component is worth noting for the opposite reason: it makes the cohort more aggressive than a typical Western series, not less.

For the radiation oncologist the operational read is that the technique is unremarkable in the best sense. 50.4Gy in 28 fractions, stoma-sparing IMRT, daily image guidance, standard pelvic nodal volumes, late grade 3+ toxicity of 8.4% against 10.5% in the observation arm. Nothing here requires a specialized program. The barrier the presenter named as a threat, reluctance among surgeons and radiation oncologists, is now the main thing standing between this result and its use.