PURPOSE: To evaluate the safety and efficacy of gemcitabine and cisplatin in combination with the immune checkpoint inhibitor pembrolizumab as neoadjuvant therapy before radical cystectomy (RC) in muscle-invasive bladder cancer. METHODS: Patients with clinical T2-4aN0/XM0 muscle-invasive bladder cancer eligible for RC were enrolled. The initial six patients received lead-in pembrolizumab 200 mg once 2 weeks prior to pembrolizumab 200 mg once on day 1, cisplatin 70 mg/m2 once on day 1, and gemcitabine 1,000 mg/m2 once on days 1 and 8 every 21 days for four cycles. This schedule was discontinued for toxicity and subsequent patients received cisplatin 35 mg/m2 once on days 1 and 8 without lead-in pembrolizumab. The primary end point was pathologic downstaging (< pT2N0) with null and alternative hypothesis rates of 35% and 55%, respectively. Secondary end points were toxicity including patient-reported outcomes, complete pathologic response (pT0N0), event-free survival, and overall survival. Association of pathologic downstaging with programmed cell death ligand 1 staining was explored. RESULTS: Thirty-nine patients were enrolled between June 2016 and March 2020 (72% cT2, 23% cT3, and 5% cT4a). Patients received a median of four cycles of therapy. All patients underwent RC except one who declined. Twenty-two of 39 patients (56% [95% CI, 40 to 72]) achieved < pT2N0 and 14 of 39 (36% [95% CI, 21 to 53]) achieved pT0N0. Most common adverse events (AEs) of any grade were thrombocytopenia (74%), anemia (69%), neutropenia (67%), and hypomagnesemia (67%). One patient had new-onset type 1 diabetes mellitus with ketoacidosis related to pembrolizumab and no patients required steroids for immune-related AEs. Clinicians consistently under-reported AEs when compared with patients. CONCLUSION: Neoadjuvant gemcitabine and cisplatin plus pembrolizumab met its primary end point for improved pathologic downstaging and was generally safe. A global study of perioperative chemotherapy plus pembrolizumab or placebo is ongoing.