Background: Immune checkpoint inhibition (ICI) is a standard of care therapy in metastatic urothelial carcinoma (mUC) but its role in the muscle-invasive setting remains unclear. Prior neoadjuvant ICI studies have demonstrated promising antitumor activity. We report results of a multi-institutional phase Ib/II two-stage study investigating chemo-immunotherapy in patients (pts) with cisplatin eligible or ineligible muscle-invasive UC (MIUC) (NCT02365766). Methods: Eligible pts were surgical candidates with clinical stage T2-4aN0M0 UC. Pts were enrolled on either the cisplatin eligible (CE) or cisplatin ineligible (CI) cohort. Both cohorts received 5 doses neoadjuvant pembrolizumab 200 mg every 3 weeks. CE chemo comprised four 21-day cycles of gemcitabine 1000 mg/m2 day (d) 1,8 and cisplatin 70 mg/m2 d 1 or 35 mg/m2 d 1,8. CI chemo comprised three 28-day cycles of gemcitabine 1000 mg/m2 d 1,8,15. Chemo-immunotherapy was followed by definitive surgery (radical cystectomy or nephroureterectomy). Primary endpoint was investigator assessed pathologic muscle-invasive response rate (PaIR, ≤ypT1N0) assessed at definitive surgery. Null hypothesis was PaIR rates of 23% (CE) and 18% (CI) for type I error rate 4% and power 86%. Secondary endpoints were rate of definitive surgery, ypT0 rate, 18 month (mo) relapse free survival (RFS), and 36 mo overall survival (OS). Results: 82 pts (42 CE, 40 CI) enrolled. 1 CI pt withdrew before cycle 1. 88.1% (CE) and 87.2% (CI) pts received definitive surgery, leaving 37 CE and 34 CI pts evaluable for the primary endpoint. 81% CE pts completed all 4 chemo cycles (median 4), and 92% CI pts completed all 3 chemo cycles (median 3). Median pembrolizumab treatments was 5 (range 1-6) in both cohorts. Median follow-up was 54.8 mos (CE) and 29.2 mos (CI). Median age was 64 (CE) and 73 (CI) and stage > T2 at diagnosis was 42.9% (CE) and 60% (CI). PaIR rate for evaluable CE pts was 54% (95% CI 38-69) with 41% pts (95% CI 27-57) down staged to ypT0. 18 mo RFS was 82% (95% CI 66-91) and 36 mo OS was 78.9% (95% CI 65-90). PaIR rate for evaluable CI pts was 53% (95% CI 37-69) with 41% pts (95% CI 26-58) downstaged to ypT0. 18 mo RFS was 65.1% (95% CI 48-78) and 36 mo OS was 65.7% (95% CI 47-79). Most common ≥ grade 3 toxicities were anemia (28.3%), hypertension (28.3%), and neutropenia (22.2%), with cytopenias more common in CE than CI pts. One death from post-operative sepsis occurred in cohort CE. Immune related adverse events (irAEs) ≥ grade 3 include elevated liver enzymes (3.7%), rash (2.5%), pneumonitis (2.5%), and colitis (2.5%). Conclusions: Neoadjuvant chemo-immunotherapy demonstrated significant down staging in CE and CI MIUC pts prior to definitive surgery, meeting the primary endpoint. Survival correlated with pathologic response. Further phase III studies will be necessary to confirm the efficacy and safety of these regimens. Ongoing analysis of biomarkers of response is underway. Clinical trial information: NCT02365766.