Background: Antigen escape relapse has emerged as a major challenge for long-term disease control post CD19-directed therapies, to which dual-targeting of CD19 and CD22 has been proposed as a potential solution. Methods: Between Mar 2016 and Jan 2018, we conducted a pilot study (ChiCTR-OPN-16008526) in 38 patients (pts), who had refractory/relapsed B-cell non-Hodgkin lymphoma (B-NHL), to evaluate the efficacy and safety of sequential infusion of anti-CD19 and anti-CD22, two single-specific, third-generation CAR19/22 T-cell “cocktail”. The cutoff date for data collection was Apr 30, 2018. Results: At a minimum follow-up of 3 months (mos), 26 of the 36 evaluable pts achieved an overall response (ORR), including 18 with a complete response (CR) and 8 with a partial response (PR). The ORR at mo 3 was consistent in different subgroups, irrespective of pathologic subtypes, cell of origin, cytogenetic or genomic aberrations. At the data cutoff, 15 of the 18 pts who had a CR at mo 3 maintained their responses, 2 of 8 pts who had a PR within 3 mos continued to have a CR without additional therapies. Collectively, the best ORR was 83.3%, with a best CR rate of 55.5% and a best PR rate of 27.8%. With a median follow-up of 5.3 mos (range, 0.4 to 16.2), the median PFS was 5.8 mos, and the median OS was not reached (NR). Pts received therapy at first relapse had better PFS than those who received therapy at the time with primary refractory diseases or at multiple relapses. Notably, pts who achieved an overall response at mo 3 (R3m) had significantly extended PFS and OS when compared with pts who did not. Repeated biopsy and IHC was conducted in 3 of the 13 pts. However, loss of CD19 or CD22 was not detected. Of the 9 pts with IgH/MYC translocation, with a median follow-up of 10.1 mos, the median PFS and median OS were NR. At data cutoff, 7 pts who had achieved R3m maintained their responses, including all the 4 pts with double-hit lymphoma. However, of the 10 pts with del(17p) or TP53 mutation, with a median follow-up of 5.3 mos (range, 2.7 to 14.5), the median PFS was 3.6 mos and the median OS was 9.9 mos. All pts experienced reversible CRS, with 21.1% were of high-grade. Neurotoxicity developed in 13.2% pts and were all low-grade. Conclusions: Our results indicated that sequential infusion of CAR19/22 T-cell is efficient and safe for pts with B-NHL. Dual antigen targeting is a promising approach to circumvent antigen loss relapse after CAR T-cell therapy. The impact of genetic subtypes and clinical parameters further underscores the critical importance of personalized immunotherapies. Clinical trial information: ChiCTR-OPN-16008526.