ESMO Congress 2026

In cisplatin-eligible muscle-invasive bladder cancer (MIBC), the evidence base for perioperative immuno-oncology plus ADC strategies rests on published phase III trials: KEYNOTE-B15/EV-304 randomized patients to 4 cycles of neoadjuvant enfortumab vedotin (EV) plus pembrolizumab followed by radical cystectomy with pelvic lymph-node dissection and adjuvant therapy, versus neoadjuvant cisplatin/gemcitabine followed by observation, with event-free survival (EFS) as the primary endpoint ^[1]; a meta-analysis pooling NIAGARA and KEYNOTE-B15 favored such perioperative immune-based regimens for EFS (HR 0.61) and overall survival ^[5]; the Australian practical guide likewise treats NIAGARA, KEYNOTE-905 and KEYNOTE-B15 as markers of a new perioperative era and lists ctDNA monitoring as a precision-medicine priority ^[6]. Reviews also note that while liquid biopsy has potential to guide MIBC treatment selection, clinical translation remains slow ^[2].

What was reported at ESMO26 is the exploratory ctDNA analysis of KEYNOTE-B15 (Abstract 4734O, presented by Jonathan Chatzkel of Moffitt, with Min Teo as discussant): 808 patients were randomized to perioperative EV + pembrolizumab (n=405) or neoadjuvant gemcitabine/cisplatin (n=403), all with planned radical cystectomy plus pelvic lymph-node dissection, using the tumor-informed Signatera assay at baseline, before surgery, after surgery and during adjuvant therapy. Presurgical ctDNA clearance was 82.1% vs 66.7% (absolute difference 15.4 percentage points), and only 4.8% vs 13.5% remained ctDNA-positive after surgery. The EFS HR in baseline ctDNA-positive patients was 0.54 (95% CI 0.37–0.80) and 0.68 (0.37–1.26, crossing 1) in negative patients. pCR among presurgical ctDNA-negative patients was 72.4% vs 43.7%, versus only 3.7% vs 2.2% among those remaining positive; postsurgical ctDNA-negative HR was 0.48 (0.29–0.81) and positive 0.63 (0.30–1.35). 24-month EFS after clearance was 81.1% vs 63.2%, only about 33–35% with persistent positivity, and postsurgical ctDNA-positive patients had only about 10% event-free rate at 24 months in either arm.

Relative to existing knowledge, the efficacy endpoints of the main trial are already established ^[1][5]; the step forward here is moving assessment from pathological response toward molecular response: ctDNA clearance correlated with higher pCR and better EFS, while persistent postsurgical positivity almost ruled out pCR and marked very poor prognosis; one live post noted that the postsurgical findings replicate NIAGARA. This echoes the literature view that liquid biopsy is promising but slow to translate ^[2] and guideline expectations for ctDNA monitoring ^[6]. Yet commentators stressed that prognostic does not equal predictive: current data cannot yet justify using ctDNA to select EV + pembrolizumab, omit cystectomy or safely de-escalate adjuvant therapy, and translating molecular clearance into clinical decisions will require prospective ctDNA-guided trials.

As for limitations, this was an exploratory analysis; the postsurgical ctDNA-positive subgroup was small, sample availability decreased over time, and several confidence intervals crossed 1, so the figures above are presented only as reported and do not yet warrant changing clinical decisions.