Reproducibility of T_1ρ and T₂ quantification in a multi-vendor multi-site study

Objective: To evaluate the multi-vendor multi-site reproducibility of two-dimensional (2D) multi-echo spin-echo (MESE) T₂ mapping (product sequences); and to evaluate the longitudinal reproducibility of three-dimensional (3D) magnetization-prepared angle-modulated partitioned k-space spoiled gradient echo snapshots (MAPSS) T_1ρ and T₂ mapping (research sequences), and 2D MESE T₂ mapping, separated by 6 months, in a multi-vendor multi-site setting. Methods: Phantoms and volunteers (n = 5 from each site, n = 20 in total) were scanned on four 3 T magnetic resonance (MR) systems from four sites and three vendors (Siemens, General Electric, and Phillips). Two traveling volunteers (3 knees) scanned at all 4 sites at baseline and 6-month follow-up. Data was transferred to one site for centralized processing. Coefficients of variation (CVs) were calculated to evaluate reproducibility. Results: For baseline 2D MESE T₂ measures, average CV were 0.37–2.45% (intra-site) and 5.96% (inter-site) for phantoms, and 3.15–8.49% (intra-site) and 14.16% (inter-site) for volunteers. For longitudinal phantom data, intra-site CVs were 1.42–3.48% for 3D MAPSS T_1ρ, 1.77–3.56% for 3D MAPSS T₂, and 1.02–2.54% for 2D MESE T₂. For the longitudinal volunteer data, the intra-site CVs were 2.60–4.86% for 3D MAPSS T_1ρ, 3.33–7.25% for 3D MAPSS T₂, and 3.11–8.77% for 2D MESE T₂. Conclusion: This study demonstrated excellent intra-site reproducibility of 2D MESE T₂ imaging, while its inter-site variation was slightly higher than 3D MAPSS T₂ imaging (10.06% as previously reported). This study also showed excellent reproducibility of longitudinal T_1ρ and T₂ cartilage quantification, in a multi-vendor multi-site setting for both product 2D MESE T₂ and 3D MAPSS T_1p/T₂ research sequences.