The Psoriatic Arthritis Magnetic Resonance Imaging Scoring System (PsAMRIS) and MRI Whole-Body Scoring System for Inflammation in Peripheral Joints and Entheses in Inflammatory Arthritis (MRI-WIPE) have not been used together to assess treatment of psoriatic arthritis in a clinical trial. We aimed to assess the effect of apremilast treatment on inflammation, with outcomes measured by PsAMRIS and MRI-WIPE. MOSAIC was a phase 4, multicentre, single-arm, open-label study conducted at 29 sites across ten countries (Belgium, Canada, Denmark, Germany, Italy, Russia, Spain, Switzerland, the UK, and the USA). Adults aged 18 years or older with a documented diagnosis of psoriatic arthritis for a duration of 3 months to 5 years self-enrolled and were included if they met the classification criteria for active psoriatic arthritis at screening. Patients were required to have at least three swollen and three tender joints with hand involvement and at least one active enthesitis site according to the Spondyloarthritis Research Consortium of Canada enthesitis index or the Leeds enthesitis index. Patients were excluded if they had previous treatment with a biological disease-modifying antirheumatic drug or previous treatment with more than two conventional synthetic disease-modifying antirheumatic drugs. After a 5-day titration period, patients received apremilast 30 mg orally twice per day. Concomitant stable methotrexate up to 25 mg per week was permitted. The primary endpoint was change from baseline to week 24 in a composite inflammation score of bone marrow oedema, synovitis, and tenosynovitis in the hand as assessed by PsAMRIS. The full analysis set and safety population included all enrolled patients who received at least one dose of apremilast. This completed study is registered with ClinicalTrials.gov (NCT03783026). Between Feb 6, 2019, and May 11, 2022, 123 patients were enrolled in the MOSAIC study. Of these 123 patients, 122 (99%) were treated with apremilast and included in the full analysis set and safety population. 67 (55%) of 122 patients were female, 55 (45%) were male, and 116 (95%) were White. 80 (66%) of 122 patients completed 48 weeks of treatment. The least squares mean change from baseline to week 24 in the composite inflammation score of bone marrow oedema, synovitis, and tenosynovitis as assessed by PsAMRIS was -2·32 (95% CI -4·73 to 0·09). 95 (78%) of 122 patients had at least one treatment-emergent adverse event. Six (5%) patients had a severe treatment-emergent adverse event and six (5%) patients had a serious treatment-emergent adverse event. No serious treatment-emergent adverse events were considered to be related to apremilast. Apremilast improved inflammation in joints and entheses on assessment of MRI measures in the hand and the whole body. Our findings encourage the use of MRI, including whole-body MRI, as an objective outcome measure in trials in patients with psoriatic arthritis. Amgen.