Background: For patients (pts) with muscle-invasive bladder cancer (MIBC) who are ineligible for cisplatin-based chemotherapy (cisplatin), the standard of care option is radical cystectomy (RC) alone. This prospective clinical trial investigated the safety and efficacy of escalating doses of atezolizumab (ATZ) as neoadjuvant therapy prior to RC in pts with non-metastatic urothelial cancer. Methods: This single-arm, single institution, phase II trial investigated the administration of one (n = 6), two (n = 5) or three (n = 9) cycles of ATZ (1200 mg IV given every 3 weeks) in pts with MIBC who are either ineligible for or refused cisplatin prior to RC. Key inclusion criteria were urothelial carcinoma of the bladder (T2-T4a,N0-1,M0) and inability to receive cisplatin-based treatment (eGFR < 60 mL/min, G≥2 neuropathy/hearing loss, pt decision). Pts with high-risk disease ( > pT2 or LN+) at the time of RC were eligible to receive adjuvant ATZ for up to 16 total cycles. Primary efficacy endpoint was pathologic complete response (pCR; pT0N0). Important secondary endpoints were safety of treatment, rates of pathologic downstaging and biomarker assessments in serial tissue samples. Pts were followed for up to 2 years following RC. Results: Among 20 pts with MIBC, median age was 69 (range 61-81) and 75% were male. Most commonly pts were cisplatin-ineligible due to low GFR (35%), hearing loss (25%) or neuropathy (10%); remainder refused cisplatin (30%). At trial enrollment, pT2, pT3, and pT4 was present in 80%, 15%, and 5% of pts and 10% had enlarged pelvic lymph nodes ( > 10 mm) on scans. All pts completed intended treatment cycles and all had RC within the defined timeframe ( > 3 weeks from last and < 12 weeks from first treatment). pCR at RC was 10% (2/20 pts), and was observed in pts receiving 1 and 2 cycles of ATZ. Pathologic downstaging (≤pT1N0) was achieved in 25% (5/20 pts) and observed across all three dose levels. Adjuvant ATZ was given to 8 pts. TRAEs of any grade during perioperative period occurred in 75% and G3 TRAEs in 10% (diarrhea, fecal incontinence). There were no G4 or G5 events. Median follow-up from the time of RC was 21.4 months at the time of data cutoff in 10/2020. Among evaluable pts, 1-year RFS and OS were 71% and 94% while 2-year RFS and OS were 64% and 75%. Conclusions: This prospective trial supports the safety and efficacy of ATZ as neoadjuvant therapy in MIBC. Although pCR and rates of downstaging were lower than what was previously reported in comparable neoadjuvant trials of checkpoint inhibitors in MIBC, pCRs in this trial were seen even in pts receiving only 1-2 doses of ATZ. Many pts had a durable recurrence-free period and all 4 evaluable pts who had pathologic downstaging were alive and disease free at 2 years post RC. Translational and biomarker work from this study is also being pursued. Clinical trial information: NCT02451423