OBJECTIVE: The use of musculoskeletal ultrasound is increasing among pediatric rheumatologists. Reliable scoring systems are needed for the objective assessment of synovitis. The aims of this study were: 1) to create a standardized and reproducible image acquisition protocol for B-Mode and Doppler ultrasound of the pediatric knee; and 2) to develop a standardized scoring system and determine its reliability for pediatric knee synovitis. METHODS: Six pediatric rheumatologists developed a set of standard views for knee assessment in children with juvenile arthritis. After a comprehensive literature review, practical exercises and a consensus process were performed. A scoring system for both B-Mode and Doppler was then developed and assessed for reliability. Inter-reader reliability or agreement among a total of 16 raters was determined using two-way single-score intraclass correlation coefficients analysis (ICC). RESULTS: Twenty-one views to assess knee arthritis were initially identified. Following completion of practical exercises and subsequent consensus processes, a total of 3 views in both B-mode and Doppler were selected: suprapatellar longitudinal and medial/lateral parapatellar transverse views. Several rounds of scoring and modifications resulted in a final ICC of: suprapatellar view B-mode 0.89 (95% CI 0.86-0.92), Doppler 0.55 (0.41-0.69); medial parapatellar view B-mode 0.76 (0.68-0.83), Doppler 0.75 (0.66-0.83); and lateral parapatellar view B-mode 0.82 (0.75-0.88), Doppler 0.76 (0.66-0.84). CONCLUSION: A novel B-Mode and Doppler image acquisition and scoring system for assessing synovitis in the pediatric knee was successfully developed through practical exercises and consensus process. Study results demonstrate overall good to excellent reliability. This article is protected by copyright. All rights reserved.