Background/Purpose:The phase 3 CHERISH trial demonstrated the efficacy of tocilizumab (TCZ), an interleukin‐6 receptor inhibitor, in patients with polyarticular‐course juvenile idiopathic arthritis (pcJIA). This analysis investigated the progression of radiographic joint damage in patients with pcJIA treated with TCZ for up to 104 weeks in CHERISH.Methods:Patients 2 to 17 years old with ≥6 months' active pcJIA for whom methotrexate failed received open‐label (OL) TCZ according to body weight (BW) (BW ≥30 kg, 8 mg/kg [n = 119]; BW <30 kg, randomly assigned 1:1 to 8 mg/kg [n = 34] or 10 mg/kg [n = 35]) every 4 weeks for 16 weeks. At 16 weeks, eligible patients (≥JIA ACR30 response) entered a 24‐week, randomized, double‐blind withdrawal period and were assigned 1:1 to placebo or continued TCZ. All patients entered an OL extension through week 104. Those receiving TCZ in all 3 parts were included in the continuous TCZ subgroup for which key radiographic end points were assessed. Radiographic progression was indicated as a positive change in adapted Sharp/van der Heijde score () (aSH) and/or a negative change in Poznanski score (), assessed on hand and wrist radiographs, from baseline to weeks 52 and 104.Results:Baseline and ≥1 postbaseline aSH and Poznanski scores were available for 45 and 35 patients, respectively, in the continuous TCZ subgroup and for 87 and 61 patients, respectively, in the total population. Reasons for missing data included early withdrawal, no consent, or unreadable radiographs (because of advanced damage resulting in unreliable measurements or growth plate fusion in postpubertal children), preventing assessment of Poznanski scores. Baseline demographics and disease characteristics were balanced for aSH and Poznanski populations and were similar to those for the full study population. At weeks 52 and 104, median changes from baseline aSH score of 0.5 (p = 0.70) and −1.0 (p = 0.11), respectively, and median changes from baseline Poznanski score of 0.3 (p = 0.07) and 0.6 (p = 0.004), respectively, indicating a lack of radiographic progression. With a cutoff of the smallest detectable difference, no radiographic progression was seen in 87.5% and 97.1% of patients based on aSH scoring and 93.5% and 96.0% of patients based on Poznanski scoring at weeks 52 and 104, respectively. Comparable proportions of the total radiographic population, including those exposed to placebo, remained radiographic progression free at week 104. Among the radiographic population, the annualized rates of progression to week 104 for aSH and Poznanski scores were 0.00 and 0.18, respectively.Conclusion:TCZ halted the progression of radiographic damage, leaving a large majority of pcJIA patients free of radiographic progression after 2 years of treatment.