Results of two phase 3 studies — IMvigor011 and DYNAMIC-III — presented at the European Society for Medical Oncology (ESMO) Annual Meeting 2025 suggest that circulating tumor DNA (ctDNA) could help guide adjuvant therapy decisions in bladder and colon cancer, potentially improving outcomes and sparing many patients from unnecessary treatment.

While many studies have shown that ctDNA has prognostic value, in the IMvigor011 trial, ctDNA positivity was strongly predictive of a clinical benefit from adjuvant atezolizumab in patients with muscle-invasive bladder cancer (MIBC).

IMvigor011 provides “the first level 1 evidence” for intervening in MIBC based on a positive plasma ctDNA test in the adjuvant setting, said study discussant Alexander Wyatt, PhD, from the University of British Columbia, Vancouver, Canada.

Patients who were consistently ctDNA-negative had “excellent outcomes with surveillance alone,” while those who were consistently ctDNA-positive had significantly longer overall survival with atezolizumab than placebo, Wyatt explained. In other words, serial plasma ctDNA testing can help tailor adjuvant treatment in MIBC, “from treat most to treat those with molecular evidence of disease,” Wyatt added.

Results of the DYNAMIC-III trial in stage III colon cancer were more nuanced. Patients with negative post-surgery ctDNA had a low risk for recurrence, and ctDNA-guided de-escalated therapy yielded outcomes approaching standard therapy, especially in low-risk tumors. However, the noninferiority of ctDNA-guided de-escalation was not confirmed in the trial. 

“Although the trial is not practice-changing for all stage III colon cancers, it is highly important and will likely influence clinical practice,” Marco Gerlinger, MD, with Barts Cancer Institute, London, UK, said in a statement from the UK Science Media Centre. 

“For low-risk stage III tumors, patients may make an informed decision to omit or reduce chemotherapy intensity if the test comes back negative, as the trial showed that this reduces severe side effects,” Gerlinger said.

IMvigor011: ctDNA and Personalized Bladder Cancer Care?

Radical cystectomy with or without neoadjuvant therapy can cure many patients with MIBC, but roughly half will develop disease recurrence, which is associated with poor prognosis, explained lead author Thomas Powles, MD, with Barts Cancer Institute.

Given variable outcomes after cystectomy, it’s important to differentiate patients more likely to relapse from those who are not.

A growing body of evidence suggests that detecting molecular residual disease (MRD) via ctDNA is “strongly prognostic,” Powles said.

In the earlier IMvigor010 trial, adjuvant atezolizumab did not lead to a significant disease-free survival (DFS) or overall survival benefit in unselected patients with MIBC. However, exploratory analyses showed that patients who were ctDNA-positive after surgery did have an overall survival benefit with adjuvant atezolizumab (hazard ratio [HR], 0.59), while ctDNA-negative patients did not (HR, 1.38). 

These results led to the IMvigor011 trial, presented at ESMO and simultaneously published in The New England Journal of Medicine, which assessed a ctDNA-guided approach to adjuvant atezolizumab.

The trial enrolled patients with MIBC and no radiographic evidence of disease within 6-24 weeks of radical cystectomy. All patients underwent serial tumor-informed ctDNA monitoring every 6 weeks for up to 1 year after surgery.

A total of 250 ctDNA-positive patients were randomly assigned (2:1) to intravenous atezolizumab (n = 167) or placebo (n = 83) every 4 weeks for up to 1 year, while 357 ctDNA-negative patients were surveilled without adjuvant therapy.

Compared with the placebo group, in ctDNA-positive patients, at a median follow-up of 16 months, median DFS was about two times higher in the atezolizumab group (9.9 months vs 4.8 months; HR, 0.64; P = .0047), and median overall survival was about 1 year longer (32.8 months vs 21.1 months; HR, 0.59; P = .0131). 

Clinical benefit with atezolizumab in ctDNA-positive patients was “generally consistent” across key subgroups, including patients excluded from prior adjuvant trials (those with pT2N0 disease), which suggests that ctDNA status enhances risk determination beyond classical surgical pathological staging, Powles said.

Similar efficacy was observed in patients with ctDNA-positive status at baseline and those who became ctDNA-positive with repeat testing, supporting serial ctDNA monitoring. 

As for toxicities, 28.5% of those in the atezolizumab group and 21.7% of those in the placebo group had adverse events of grade 3 or 4, with 3% and 2.4%, respectively, having a fatal adverse event (related to atezolizumab in 1.8% and to placebo in none).

Notably, said Powles, the persistently ctDNA-negative patients did exceptionally well without adjuvant treatment. DFS was 95% at the end of the 1-year surveillance period and 88% at 2 years, meaning ctDNA-negative patients “can be spared unnecessary and toxic and expensive therapy,” he said.

Looking ahead, Powles said he sees a world “in the not-too-distant future where ctDNA turns out to be more accurate than radiology, and so we’ll be tracking patients with blood tests rather than CT scans, and that’s going to be broadly applicable.”

Wyatt agreed that this trial may “change the paradigm.” 

“Typically, we would want to see the disease on a scan to trigger treatment,” Wyatt said during an ESMO press briefing. But this trial says it’s “molecular evidence of disease that could guide these decisions.” 

DYNAMIC-III: Safe to Skip Chemo in ctDNA-Negative Colon Cancer?

The utility of a ctDNA-guided approach in colon cancer is less clear.

In stage II disease, a ctDNA-guided approach has been shown to reduce adjuvant chemotherapy use without compromising recurrence-free survival.

However, in stage III disease, escalating therapy in ctDNA-positive patients did not improve recurrence-free survival over standard management, according to an earlier DYNAMIC-III analysis reported during ASCO 2025. In fact, the ctDNA-informed approach resulted in worse recurrence-free survival at 2 years — 52% in the escalated therapy group vs 61% in the standard care group (HR, 1.11; P = .6).

The latest findings from DYNAMIC-III, presented at ESMO and simultaneously published in Nature Medicine, explored whether the ctDNA-negative cohort would benefit from treatment de-escalation compared with the standard of care.

In this trial, 702 patients with stage III colon cancer who were ctDNA-negative 5-6 weeks following surgery were randomly assigned to standard clinician-directed adjuvant therapy or ctDNA-guided de-escalated therapy. The trial assessed several de-escalated approaches, including going from 6 months of oxaliplatin plus fluoropyrimidine to 3 months of the doublet regimen or 6 months of fluoropyrimidine alone, as well as 3-6 months of fluoropyrimidine or no chemotherapy and 3 months of the doublet regimen to 3-6 months of fluoropyrimidine. 

ctDNA-guided treatment de-escalation significantly reduced use of oxaliplatin from 89% to 35%, leading to fewer hospitalizations (8.5% vs 13.2% on standard of care; P = .047) and fewer grade 3 or higher adverse events (6.2% vs 10.6%; P = .037).

However, ctDNA-guided de-escalation yielded slightly lower recurrence-free survival at 3 years — 85.3% vs 88.1% — which missed the margin for noninferiority.

Although the overall recurrence-free survival difference was “very small,” noted lead author Jeanne Tie, MD, MBChB, with the Peter MacCallum Cancer Centre in Melbourne, Australia, “unfortunately, statistically this study did not meet its primary endpoint, meaning that we are not 100% certain that de-escalation definitely would not be leading to detriment in outcome.”

Notably, a pre-planned subgroup analysis suggested that de-escalation may be noninferior in clinical low-risk tumors (T1-3N1) — the 3-year recurrence-free survival was 91.0% vs 93.2% — though not for clinical high-risk tumors (T4 and/or N2) — 72.8% vs 78.6%.

So where does that leave clinicians today? 

“This is complicated,” said Tanios Bekaii-Saab, MD, with Mayo Clinic, Phoenix, who served as discussant for DYNAMIC-III. There’s still predictive uncertainty for MRD-negative and MRD-positive patients with stage III colon cancer, he said.

Bekaii-Saab explained that noninferiority trial designs can be problematic since “you need a very large sample size to be able to adjust for many of the confounding factors that ultimately may limit the robustness of conclusions.”

He highlighted baseline imbalances between the study groups, which should have favored the experimental arm overall. The standard arm had a higher percentage of T4 lesions, which do worse, and the experimental arm had more microsatellite instability high cases, which typically do better, and “yet the experimental arm did worse,” he said. 

He also noted that recurrence-free survival at 2 or 3 years within the DYNAMIC framework “remains to be validated as a reliable surrogate for overall survival.”

“And then, of course, the big elephant in the room is the cost-value equation of MRD assessment,” given that it’s a prognostic tool right now, Bekaii-Saab said. 

Tie agreed that “more work will be required to further demonstrate the clear signal of clinical utility,” but believes the findings from DYNAMIC-III suggest that ctDNA “could help further personalize chemotherapy decisions.”

The IMvigor011 study was funded by F. Hoffmann-La Roche. Powles reported receiving support from Roche and other pharmaceutical companies. Wyatt declared having no relevant financial relationships. The DYNAMIC-III study had no commercial funding. Tie and Bekaii-Saab disclosed having relationships with various pharmaceutical companies. Disclosures for Gerlinger were not available.