DANUBE
In untreated unresectable locally advanced/metastatic urothelial carcinoma, do durvalumab or durvalumab + tremelimumab improve OS over platinum chemotherapy?
Biomarker / PD-L1
PD-L1 status was a randomization stratification factor. High PD-L1 was defined as ≥25% of tumour cells or tumour-associated immune cells staining for membrane PD-L1, using the trial's composite tumour-cell/immune-cell algorithm. The durvalumab-monotherapy coprimary OS endpoint was tested specifically in the PD-L1-high population; the durvalumab+tremelimumab coprimary OS endpoint was tested in the ITT population regardless of PD-L1 status.
Absolute Effect
No defensible fixed-time absolute OS benefit is used because the coprimary analyses were reported primarily as medians and hazard ratios and neither endpoint met its prespecified criterion.
Safety: Grade 3–4 treatment-related AEs were 14% with durvalumab, 27% with durvalumab+tremelimumab, and 60% with chemotherapy. Serious treatment-related AEs were 9%, 23%, and 16%, respectively.
Deaths from study-drug toxicity: 2 (1%) durvalumab, 2 (1%) combination, and 1 (<1%) chemotherapy.
Interpretation & limitations
DANUBE did not meet either coprimary endpoint. The study nevertheless provides important randomized evidence about the limits of unselected first-line checkpoint blockade in urothelial carcinoma and the additional toxicity of CTLA-4 combination therapy.
Primary source
Powles T, et al. Lancet Oncol. 2020;21:1574–1588.