Background Novel second-line treatments are needed for patients with advanced urothelial cancer (UC). Interim analysis of the phase III KEYNOTE-045 study showed a superior overall survival (OS) benefit of pembrolizumab, a programmed death 1 inhibitor, versus chemotherapy in patients with advanced UC that progressed on platinum-based chemotherapy. Here we report the long-term safety and efficacy outcomes of KEYNOTE-045. Patients and methods Adult patients with histologically/cytologically confirmed UC whose disease progressed after first-line, platinum-containing chemotherapy were enrolled. Patients were randomly assigned 1 : 1 to receive pembrolizumab [200 mg every 3 weeks (Q3W)] or investigator’s choice of paclitaxel (175 mg/m 2 Q3W), docetaxel (75 mg/m 2 Q3W), or vinflunine (320 mg/m 2 Q3W). Primary end points were OS and progression-free survival (PFS) per Response Evaluation Criteria in Solid Tumors, version 1.1 (RECIST v1.1) by blinded independent central radiology review (BICR). A key secondary end point was objective response rate per RECIST v1.1 by BICR. Results A total of 542 patients were enrolled (pembrolizumab, n = 270; chemotherapy, n = 272). Median follow-up as of 26 October 2017 was 27.7 months. Median 1- and 2-year OS rates were higher with pembrolizumab (44.2% and 26.9%, respectively) than chemotherapy (29.8% and 14.3%, respectively). PFS rates did not differ between treatment arms; however, 1- and 2-year PFS rates were higher with pembrolizumab. The objective response rate was also higher with pembrolizumab (21.1% versus 11.0%). Median duration of response to pembrolizumab was not reached (range 1.6+ to 30.0+ months) versus chemotherapy (4.4 months; range 1.4+ to 29.9+ months). Pembrolizumab had lower rates of any grade (62.0% versus 90.6%) and grade ≥3 (16.5% versus 50.2%) treatment-related adverse events than chemotherapy. Conclusions Long-term results (>2 years’ follow-up) were consistent with those of previously reported analyses, demonstrating continued clinical benefit of pembrolizumab over chemotherapy for efficacy and safety for treatment of locally advanced/metastatic, platinum-refractory UC. Trial registration ClinicalTrials.gov: NCT02256436.

Patients with mrTRG1 without tumour spread may be suitable for TAE.