AUA 2026: Consolidative local therapy after EV+pembro response (surgery vs radiation vs surveillance)
TL;DRRT-based preservation, cystectomy and surveillance each argued for cCR after neoadjuvant EV+pembro in MIBC; no randomized data compare the three options.
Reported via UroToday →
This point-counterpoint session asked how to manage the bladder in MIBC after cCR on neoadjuvant EV+pembro. The RT case held that clinical restaging misses residual disease, and that RT-based preservation keeps the bladder with better sexual function and body image than cystectomy. The surgical case held that all level 1 EV+pembro data include cystectomy, and that cCR predicts pCR only 51% of the time. The surveillance case shifted the question from pCR to MFS. It rested on a 97% PPV of cCR for 2-year MFS and on local recurrence coming before metastasis, which would allow salvage. ctDNA was presented as promising but not yet able to select pts.
- No data were shown on RT consolidation after EV+pembro; the RT case rests on chemo-era cohorts and observational TMT vs RC QOL comparisons, where pts chose their modality.
- The 97% PPV comes from GC + nivo in HCRN GU16-257; applying it to EV+pembro assumes a cCR means the same thing under a different regimen.
- The 51% cCR-pCR concordance measures residual pathology, not whether that residual changes MFS once surveillance and salvage are in place, so the surgical case treats pCR as if it were an outcome.
- RETAIN-1 vs RETAIN-2 is a cross-trial comparison in small surveillance cohorts (N=22 in RETAIN-2) after AMVAC-based therapy, not EV+pembro, and no trial compares RT, cystectomy and surveillance head to head.
KEYNOTE-905/EV-303KEYNOTE-B15/EV-304SWOG 0219RETAIN-1RETAIN-2HCRN GU16-257ARCHERBRIGHT
For RT, the key observation is the failure pattern: in RETAIN-2, no pt developed mets without a prior local recurrence, which is the premise behind both salvage RT and surveillance. BRIGHT (55 Gy/20 fx whole bladder, then pembro) is the trial that will test whether RT consolidation after cT0–T1 response adds bladder control beyond surveillance. No EV-era RT data exist yet.
In MIBC with cCR after neoadjuvant EV+pembro, the evidence leaves cystectomy, RT-based preservation and surveillance all open, each supported only by non-randomized or chemo-era data. It does not apply to pts with residual MIBC on restaging TURBT.
No EV-era data support RT consolidation after cCR yet. BRIGHT (55 Gy/20 fx whole bladder, then pembro for 12 mo) is the test. The local-before-metastatic failure pattern in RETAIN is what makes salvage RT a credible backstop for pts on surveillance.
Both phase III EV+pembro trials mandated cystectomy, and pooled cCR-pCR concordance is 51%, so omitting RC means leaving level 1 evidence behind. ctDNA negativity does not rule out ≥pT2 disease (67% residual), so a negative test cannot yet justify skipping surgery.
11 details 4 trials watching
AUA 2026 point-counterpoint session with three advocates (RT, cystectomy, surveillance), reported in a UroToday write-up. It argues across published and presented data and reports no new trial result.
MIBC (cT2–T4aN0M0 in the cited trials) with clinical complete response after neoadjuvant EV+pembro, i.e. the population treated per KEYNOTE-905/EV-303 and KEYNOTE-B15/EV-304.
The slide shows the BRIGHT schema: RT 55 Gy/20 fx to the whole bladder, then pembro 200 mg q21 days for 12 months, in pts with a cT0–T1 response confirmed by TURBT with systematic biopsies after ≥3 cycles of NAT. The ARCHER arm regimens were not legible on the slide.
In EV-303 and EV-304, every pt was planned for cystectomy, so the EFS and OS benefits (EV-303 EFS HR 0.40, OS HR 0.50) reflect EV+pembro plus RC. Organ-sparing evidence comes from RETAIN-1/2 and HCRN GU16-257, which used chemo or chemo-IO backbones, not EV+pembro.
The disagreement comes down to which endpoint cCR has to predict. Against pCR, cCR fails often (51% concordance); against MFS, it looked accurate in one GC + nivo trial (PPV 97%). If local failure reliably comes before metastasis, salvage keeps that miss from costing a cure, but this has not been shown after EV+pembro.
The trial IDs for ARCHER and BRIGHT differ between the write-up and the slide. The QOL comparison mixes TMT cohorts where pts chose their modality, and it shows worse pain and emotional functioning with TMT.
| Option | Core argument | Evidence cited |
|---|---|---|
| Bladder-preserving RT | cCR does not guarantee eradication; RT gives durable bladder control with a QOL advantage | SWOG 0219, RETAIN-2, TMT vs RC patient-reported outcomes |
| Radical cystectomy | Cystectomy is the backbone of all level 1 EV+pembro data; cCR is not pCR | KEYNOTE-905/EV-303, KEYNOTE-B15/EV-304, Roessler meta-analysis, RETAIN ctDNA |
| Surveillance | Most pts had pCR at cystectomy anyway; the real question is whether cCR predicts MFS | HCRN GU16-257, RETAIN failure patterns, ctDNA (Epstein, ASCO GU 2026) |
| Outcome | RETAIN-1 | RETAIN-2 |
|---|---|---|
| Local recurrence | 62% | 36% |
| Metastatic recurrence | 36% | 14% |
| MFS with intact non-irradiated bladder | 48% | 68% |
- Does cCR after EV+pembro predict MFS as it did after GC + nivo? not yet A Phase 2 Study Evaluating Bladder Preservation With Enfortumab Vedotin Plus Pembrolizumab in Patients With Muscle-Invasive Urothelial Cancer Phase 2n=25 · primary completion 2034-11 · EV+P clinical response-based sparing, no RC or RT
- Does RT consolidation improve bladder control over surveillance after cCR? recruiting Enfortumab Vedotin in Combination With Pembrolizumab vs. Concurrent Chemoradiotherapy (cCRT) in People With Muscle Invasive Bladder Cancer (EV-309) Phase 3n=390 · primary completion 2030-03 · phase 3 EV+P alone vs chemoRT for bladder preservationn=56 · primary completion 2031-04 · response-adapted EV+P induction, then RT, CRT or RC
- Can ctDNA select pts to safely omit cystectomy? recruiting NEO-BLAST: Neoadjuvant Therapy for Bladder Cancer Followed by Active Surveillance vs Treatment Phase 2/3n=688 · primary completion 2028-12 · ctDNA/MRI-selected surveillance vs definitive therapy
📚 Sources · 📄 1 paper
Abstract
The longer read
The session is useful less for any one advocate's conclusion than for showing that the three positions disagree about which endpoint matters. The surgical case measures cCR against pCR and finds it unreliable: a pooled 51% of cCR pts were truly pT0 in the Roessler meta-analysis, and SWOG 0219 found residual disease in 6 of 10 cT0 pts who went to immediate cystectomy. The surveillance case concedes that miss rate and changes the endpoint to MFS. On that measure HCRN GU16-257 reported a 97% PPV of cCR for 2-year MFS. That is an argument about salvage: residual or recurrent disease is acceptable as long as it can be caught locally before it spreads. Which framing a reader adopts largely decides the answer.
The salvage argument therefore rests on the failure pattern, and the data behind it are thin. In RETAIN-2 no metastasis appeared without a prior local recurrence, and in RETAIN-1 local recurrence came first in 8 of 9. These are small surveillance cohorts after AMVAC-based therapy, and the RETAIN-1 vs RETAIN-2 improvement is a cross-trial comparison, not a randomized one. A local-first pattern after chemo-IO may well hold after EV+pembro, but the session presented no EV-era surveillance data to show it.
For an RT reader, the gap is plainer still. The RT case was built on the limits of clinical restaging, on chemo-era preservation cohorts, and on patient-reported outcomes from TMT vs cystectomy comparisons. Those comparisons are observational, with pts who chose their modality, and they cut both ways: better sexual function and body image, worse pain and emotional functioning. None of it isolates what RT adds over surveillance in a pt who is already cT0 after an ADC-IO regimen. If systemic therapy has cleared most of the invasive disease, consolidative RT has to earn its toxicity by lowering local recurrence enough to matter, and that is exactly what BRIGHT's single-arm whole-bladder 55 Gy/20 fx plus pembro design will measure.
The surgical position has the strongest evidence base but argues past it slightly. The EV-303 and EV-304 benefits are real, and they were generated with cystectomy in every pt, so omitting surgery means leaving the conditions of level 1 evidence. But those trials cannot tell us how much of the survival benefit comes from extirpation in pts who were already pT0, and the session's own figures suggest that group is large. The ctDNA data sharpen the problem without resolving it: a positive result flags residual disease, but 67% of ctDNA-negative pts still had ≥pT2 disease, so a negative test cannot yet justify skipping cystectomy.
What would settle the question is a randomized, or at least a prospectively response-adapted, comparison in EV+pembro-treated pts, with MFS and bladder-intact survival as endpoints and a standardized cCR definition. Until ARCHER and BRIGHT report, cystectomy remains the default the phase III data support, and preservation or surveillance remains a reasonable but unproven option to discuss with a well-staged, motivated pt.