Blood Test for Circulating Tumor DNA Predicts Metastasis Risk After Bladder-Sparing Treatment

Bladder cancer treatment has long forced patients into a stark choice: remove the bladder and live without it, or preserve it and accept uncertainty about whether the cancer is truly gone. Surgery carries serious quality-of-life consequences - patients who undergo cystectomy must manage urine collection for the rest of their lives through an external bag. Bladder-sparing approaches, when they work, avoid that outcome entirely.

The challenge is identifying who can safely skip surgery. New data from the phase 2 RETAIN-2 clinical trial, presented at the 2026 ASCO Genitourinary Cancers Symposium in San Francisco, offer a partial answer - and a clear boundary to what circulating tumor DNA tests can tell us.

The RETAIN-2 trial design

More than 70 patients with muscle-invasive bladder cancer (MIBC) received combination chemotherapy plus nivolumab, an immune checkpoint inhibitor. Patients who showed a complete response - no detectable cancer after treatment - entered active surveillance rather than proceeding to immediate surgery. Blood samples were analyzed for circulating tumor DNA, or ctDNA, at multiple time points throughout the study. Circulating tumor DNA consists of tiny DNA fragments shed by dying cancer cells into the bloodstream; its presence or absence reflects whether active tumor activity is occurring.

The headline efficacy result was strong: 80 percent of complete responders who entered active surveillance remained free of metastatic disease after two years. That outcome compares favorably to historical results from chemotherapy alone in earlier trials, supporting the case for adding immunotherapy to the bladder-preservation approach.

What the ctDNA test can and cannot detect

The specific question the researchers addressed is whether ctDNA can guide decisions about who needs closer monitoring or intervention. The answer is split in a way that matters clinically.

ctDNA performed well as a predictor of metastatic risk. Patients who tested positive for ctDNA after treatment were substantially more likely to eventually develop metastases. The absence of ctDNA predicted favorable outcomes regardless of whether the patient had their bladder removed or kept it - suggesting the test captures something real about systemic disease burden.

For local recurrence within the bladder, however, ctDNA was not informative. Of the 22 patients who developed a recurrence of cancer in their bladder during surveillance, 19 had no detectable increase in ctDNA. The tumor was growing locally while leaving essentially no detectable signal in the blood.

"This tells us ctDNA can be incorporated into the decision-making of who should keep their bladder and who should not, knowing that we also need additional tests or biomarkers to detect local recurrence early in patients who undergo active surveillance," said Dr. Pooja Ghatalia of Fox Chase Cancer Center, the study's first author.

Why local and systemic recurrence behave differently

The divergence between ctDNA's performance for metastatic versus local recurrence is not surprising from a biological standpoint, though it has important practical implications. Metastatic disease requires cancer cells to enter the bloodstream and establish new tumors at distant sites - a process that generates ctDNA almost by definition. A tumor that recurs locally within the bladder wall may shed very little DNA into the circulation, particularly if it is small and early-stage. This means ctDNA surveillance would likely miss these cases until the local recurrence became large enough to produce detectable systemic signal.

The implication is that bladder-sparing strategies require complementary monitoring tools. Imaging and endoscopic surveillance of the bladder itself will likely remain necessary alongside ctDNA testing, rather than being replaced by it.

What comes next

Researchers will continue following RETAIN-2 participants for five years to assess long-term outcomes. A follow-on trial, RETAIN-3, is in design with plans to use ctDNA as a prospective biomarker informing treatment decisions in real time rather than retrospectively analyzing it after the fact. That design, if successful, would represent a more direct test of whether ctDNA-guided decision-making improves patient outcomes compared to standard surveillance protocols.