OBJECTIVES: To investigate if the OMERACT PsA MRI Scoring System (PsAMRIS), including a novel total inflammation score, shows sensitivity to change with an agent (abatacept) known to impact clinical outcomes in PsA. METHODS: We performed a post hoc analysis of a randomized phase IIb study of abatacept in patients with PsA and inadequate DMARD response. Participants received one of three abatacept dosing regimens [ABA3, ABA10 or ABA30/10 mg/kg (30 mg/kg switched to 10 mg/kg after two doses)] or placebo until day 169, then ABA10 through day 365. MRIs at baseline and days 85, 169 and 365 were centrally evaluated by two readers blinded to chronological order and treatment arm. Synovitis, osteitis, tenosynovitis, periarticular inflammation, bone erosions, joint space narrowing and bone proliferation were assessed using the PsAMRIS. A novel total inflammation score was tested. RESULTS: MRIs for 123 patients were included. On day 169, ABA10 and ABA30/10 significantly reduced MRI synovitis and tenosynovitis, respectively, vs placebo [differences -0.966 (P = 0.039) and -1.652 (P = 0.014), respectively]. Synovitis in the placebo group increased non-significantly from baseline to day 169, total inflammation and tenosynovitis decreased non-significantly and all measures improved significantly after a switch to ABA10 [-1.019, -0.940, -2.275 (P < 0.05), respectively, day 365 vs day 169]. Structural outcomes changed minimally across groups. CONCLUSION: Adults with PsA receiving ABA10 and ABA30/10 demonstrated significant resolution of inflammatory components of disease, confirmed by MRI, with synovitis and tenosynovitis improvements consistent with previously reported clinical responses for these doses. Results indicate that a reduction in OMERACT PsAMRIS inflammation scores may provide proof of tissue-level efficacy in PsA clinical trials. REGISTRATION: ClinicalTrials.gov (https://clinicaltrials.gov), NCT00534313.