onc brain

The longer read

Trimodality therapy earned its guideline place on a thin prospective evidence base plus one large institutional experience, and the standing objection to it has never been the complete-response rate, it is the selection of the pts who get offered it. This cohort speaks to that gap: 369 pts across multiple Spanish centers over 2010 to 2022, median age 76, treated off protocol. A 63.7% complete local response in that population is the number worth carrying, because it comes from ordinary practice rather than from the trial-eligible slice of it, and because the 9.7% salvage cystectomy rate sits well below the 28.8% who progressed, which in a cohort this old says as much about fitness as about oncology.

The interesting claim is not the response rate, it is the pair of predictors. Weekly portal imaging carried OR 0.35 (0.20-0.60) for complete local response, and 5-FU-based chemoradiotherapy OR 4.9 (1.1-22.1). Read causally, the first says image-guidance quality drives local control in an organ that moves and fills between fractions, which is biologically plausible and is exactly why daily volumetric imaging and plan-of-the-day adaptation became standard. Read skeptically, weekly portal imaging is a marker of when and where a pt was treated. A 2010 to 2022 accrual window spans the shift from 2D verification to daily cone-beam, and it spans the parallel shift in planning technique, which is presumably why VMAT trended favorable without reaching significance. Multivariable adjustment cannot separate a technique from the era that used it when the two are close to collinear, so the honest reading is that the imaging variable stands in for a bundle of modernization, not that changing a verification schedule delivers that odds ratio.

The chemotherapy finding needs the same discipline. An interval running from 1.1 to 22.1 is compatible with a marginal effect and with a very large one, and in a retrospective series regimen choice tracks renal function, performance status and institutional habit, each of which independently predicts response. The direction is consistent with the trial that defines this space, BC2001 having shown that adding chemotherapy to radiotherapy improves locoregional control, and that consistency is the value here rather than the point estimate.

What the analysis cannot do is what its conclusion reaches for. There is no cystectomy arm, no matching, and no reported follow-up or survival estimate, so bladder preservation as an alternative to surgery is asserted against external series rather than tested, and the tweet's claim that complete response predicts better survival arrives without a magnitude. The recurrence pattern data are the more useful part for a radiation oncologist: systemic failure at 10.7% ran at or above isolated local failure at 10.1%, which argues that the limiting factor in a median-age-76 MIBC population is micrometastatic disease rather than the bladder. That reframes where the next gain lives. Sharper image guidance and an active concurrent backbone are worth defending because they protect the local half of the problem cheaply, but they are unlikely to move the survival curve much on their own, and a trial that intensifies local therapy further in this population would be answering a question the failure pattern says is already near its ceiling.