Introduction: This post hoc analysis of phase 2 trial data assessed the efficacy of tofacitinib on magnetic resonance imaging (MRI) outcomes with the detailed anatomy-based Canada–Denmark (CANDEN) MRI scoring system and evaluated tofacitinib suppression of spinal inflammation in patients with active ankylosing spondylitis (AS). Methods: Patients with active AS (per modified New York criteria) were randomized 1:1:1:1 to receive tofacitinib 2, 5, or 10 mg twice daily (BID), or placebo, in a 16-week, phase 2, double-blind clinical trial. Spine MRI assessments were performed at baseline and week 12. For post hoc analysis, MRI images from patients receiving tofacitinib 5 or 10 mg BID, or placebo, were re-evaluated by two readers blinded to time point/treatment and assessed by the CANDEN MRI scoring system. Least squares mean changes from baseline to week 12 were reported for CANDEN-specific MRI outcomes, with analysis of covariance used for comparisons of pooled tofacitinib and tofacitinib 5 or 10 mg BID versus placebo. p values without multiplicity adjustment were reported. Results: MRI data from 137 patients were analyzed. At week 12, CANDEN spine inflammation score and vertebral body, posterior elements, corner, non-corner, facet joint, and posterolateral inflammation subscores were significantly reduced with pooled tofacitinib versus placebo (p < 0.0001; except non-corner subscore, p < 0.05). Total spine fat score was numerically increased with pooled tofacitinib versus placebo. Conclusions: In patients with AS, tofacitinib treatment was associated with significant reductions in MRI scores of spinal inflammation versus placebo, as assessed by the CANDEN MRI scoring system. Tofacitinib reduced inflammation in posterolateral elements of the spine and facet joints, which has not been described previously. Trial Registration: ClinicalTrials.gov registry (NCT01786668).