Simple Summary
Immunotherapy has transformed the treatment of bladder cancer, but not every patient benefits, and treatment carries risks and costs. Doctors have therefore spent a decade searching for a test that identifies in advance who will benefit. The test used most widely measures a protein called PD-L1 in tumor tissue. In this review we examined the main randomized trials and found that PD-L1 testing has largely failed to do what was hoped. Different laboratory tests give different answers, and across trials PD-L1 did not consistently separate patients who benefit from those who do not. By contrast, a blood test that detects fragments of tumor DNA after the bladder is removed identifies patients whose cancer is likely to return. In a randomized trial, such patients lived longer when given immunotherapy. Patients whose test stayed negative were not treated and most remained free of cancer, although about one in nine relapsed within two years, and because they were not randomized the trial could not show whether treatment would have helped them. We summarize what clinicians should and should not test for today.
Abstract
Immune checkpoint inhibitors (ICIs) are established across the treatment continuum of urothelial carcinoma of the bladder, yet most patients derive no durable benefit. This narrative review appraises candidate predictors of ICI benefit in bladder urothelial carcinoma, distinguishing prognostic from predictive biomarkers, applying the same three evidentiary standards to every biomarker, classifying the supporting studies by design and separating evidence from regulatory status. Programmed death-ligand 1 (PD-L1) immunohistochemistry meets the enrichment standard in three co-primary PD-L1-positive populations, in trials also positive overall, but not the predictive standard: its four registrational assays are not interchangeable; randomized comparisons show inconsistent subgroup effects, failed prespecified tests in IMvigor211 and DANUBE and no significant treatment-by-PD-L1 interaction; and the benefit of first-line enfortumab vedotin plus pembrolizumab is independent of combined positive score. Its remaining role is confined to label restrictions, chiefly European. Tumor mutational burden, molecular subtypes, expression signatures, DNA damage response and FGFR3 alterations remain investigational. Tumor-informed circulating tumor DNA (ctDNA) answers a different question: whether residual disease persists after cystectomy. After an exploratory signal in IMvigor010, the prospective enrichment trial IMvigor011 showed that ctDNA-positive patients derive disease-free and overall survival benefit from adjuvant atezolizumab; persistently ctDNA-negative patients were observed without randomization, so their 88% two-year disease-free survival does not exclude a benefit of treatment. This underpins US companion-diagnostic approval and a Category 1 guideline recommendation, not yet European authorization. Post-cystectomy ctDNA thus meets the same enrichment standard as PD-L1 and, like PD-L1, neither the predictive nor the strategy standard; what sets it apart is that the marker was necessary for the result, adjuvant atezolizumab having failed unselected, and that it directs whether to treat rather than which regimen; its role after perioperative immunotherapy is unresolved.