PURPOSE: To analyze reproducibility of ADC measurements on different MRI scanners. METHODS AND MATERIALS: Diffusion-weighted imaging (DWI) of a home-made phantom (gelatine of different concentration was enclosed in case of circle pillar shape) was performed with the following protocol; spin-echo type echo planar imaging; TR=7000 msec, TE=minimal, matrix=128 x 128, FOV=220 mm, slice thickness=5mm, NEX=1, b-factor=0, 500, and 1000 sec/mm(2) on four different 1.5T MR scanners [GE Signa Horizon (two of them), GE Signa Horizon Lx, and Siemens Magnetom Symphony]. Image analysis were performed on a workstation (GE Medical Systems) or on an operators console of Symphony (for data of Symphony). We used two-point method. Regions of interest (ROIs) were set on every part of ADC map of home-made phantom and ADCs were measured. A linear approximation graphs were also made with every MRI scanner and b-factor. RESULTS: A linear approximation was performed, formulas were y=-0.08x+2.32, r(2)=0.84(Lx), y=-0.09x+2.58, r(2)=0.89(Horizon-1) y=-0.11x+2.67, r(2)=0.95(Horizon-2), y=-0.08x+2.42, r(2)=0.97(Symphony) at b-factor of 500sec/mm(2). y=-0.08x+2.33, r(2)=0.91(Lx), y=-0.09x+2.43, r(2)=0.93(Horizon-1), y=-0.08x +2.48, r(2)=0.92(Horizon-2), y=-0.07x+2.39, r(2)=0.96(Symphony) at b-factor of 1000 sec/mm(2). CONCLUSIONS: ADC values measured at two MRI scanners (same model and older than the other scanners) are higher than other scanners on b-factor=500 sec/mm(2), but at b-factor=1000 sec/mm(2), a good relationship between every scanners was obtained. This result seems to depend on generation of the MRI scanners; newly scanners have good relations b-factor 500 and 1000 sec/mm(2). Pulse sequence design may be improved and optimized in newly scanners. When ADCs were used as quantitative value, a proper b-factor must be used.
PURPOSE: To analyze reproducibility of ADC measurements on different MRI scanners. METHODS AND MATERIALS: Diffusion-weighted imaging (DWI) of a home-made phantom (gelatine of different concentration was enclosed in case of circle pillar shape) was performed with the following protocol; spin-echo type echo planar imaging; TR=7000 msec, TE=minimal, matrix=128 x 128, FOV=220 mm, slice thickness=5mm, NEX=1, b-factor=0, 500, and 1000 sec/mm(2) on four different 1.5T MR scanners [GE Signa Horizon (two of them), GE Signa Horizon Lx, and Siemens Magnetom Symphony]. Image analysis were performed on a workstation (GE Medical Systems) or on an operators console of Symphony (for data of Symphony). We used two-point method. Regions of interest (ROIs) were set on every part of ADC map of home-made phantom and ADCs were measured. A linear approximation graphs were also made with every MRI scanner and b-factor. RESULTS: A linear approximation was performed, formulas were y=-0.08x+2.32, r(2)=0.84(Lx), y=-0.09x+2.58, r(2)=0.89(Horizon-1) y=-0.11x+2.67, r(2)=0.95(Horizon-2), y=-0.08x+2.42, r(2)=0.97(Symphony) at b-factor of 500sec/mm(2). y=-0.08x+2.33, r(2)=0.91(Lx), y=-0.09x+2.43, r(2)=0.93(Horizon-1), y=-0.08x +2.48, r(2)=0.92(Horizon-2), y=-0.07x+2.39, r(2)=0.96(Symphony) at b-factor of 1000 sec/mm(2). CONCLUSIONS: ADC values measured at two MRI scanners (same model and older than the other scanners) are higher than other scanners on b-factor=500 sec/mm(2), but at b-factor=1000 sec/mm(2), a good relationship between every scanners was obtained. This result seems to depend on generation of the MRI scanners; newly scanners have good relations b-factor 500 and 1000 sec/mm(2). Pulse sequence design may be improved and optimized in newly scanners. When ADCs were used as quantitative value, a proper b-factor must be used.