Background: Patients (pts) with anal squamous cell carcinoma (ASCC) have poor outcomes and few treatment options. We report a pooled analysis of pembrolizumab (pembro) antitumor activity and safety in the ASCC cohorts of the multicohort studies KEYNOTE-028 (NCT02054806; phase 1b) and KEYNOTE-158 (NCT02628067; phase 2), providing a robust sample size and longer follow-up. Methods: Eligible pts were aged ≥18 y with histologically/cytologically confirmed metastatic/unresectable ASCC, had prior failure of/intolerance to standard therapy or no standard therapy options, measurable disease (RECIST v1.1), ECOG PS 0/1, and a tissue sample evaluable for PD-L1/biomarkers (KEYNOTE-028 required PD-L1 positivity). Baseline PD-L1 expression was assessed using a prototype IHC assay (QualTek) in KEYNOTE-028 and the PD-L1 IHC 22C3 pharmDx assay (Agilent Technologies) in KEYNOTE-158. Pts received pembro 10 mg/kg Q2W (KEYNOTE-028) or 200 mg Q3W (KEYNOTE-158) for 2 y or until PD/unacceptable AEs. The primary endpoint in both studies was ORR (per RECIST v1.1). Secondary endpoints were duration of response (DOR), PFS, OS, and safety. Results: 137 pts with ASCC were treated in KEYNOTE-028 (n = 25) or KEYNOTE-158 (n = 112) and were included in this analysis (median age, 61 y; 83.2% women; 73.0% had PD-L1–positive tumors). Median follow-up was 11.7 mo; 124 pts (90.5%) had discontinued treatment. ORR (95% CI) was 10.9% (6.3%–17.4%). 8 pts had CR and 7 had PR. ORR (95% CI) by PD-L1 status was 14.0% (7.9%–22.4%) in the PD-L1 positive group and 3.3% (0.1%–17.2%) in the PD-L1 negative group. Among all treated pts, median DOR was not reached (range, 6.0+ to 57.5+ mo). By Kaplan-Meier estimation, 84.6% of responders had a DOR ≥24 mo. Median PFS was 2.1 mo (95% CI, 2.0–2.1) and median OS was 11.7 mo (95% CI, 8.8–14.5). The 12-mo PFS and OS rates were 14.5% and 47.4%. 85 pts (62.0%) had +1 treatment-related AE, 24 pts (17.5%) with grade 3–4 events (no grade 5 events). 32 pts (23.4%) had immune-mediated AEs; 2 pts (1.5%) had infusion related reactions. Conclusions: In pts with previously treated advanced ASCC, pembro showed durable antitumor activity, particularly in pts with PD-L1–positive tumors, and manageable toxicity. Clinical trial information: NCT02054806 (KEYNOTE-028), NCT02628067 (KEYNOTE-158).