Filter material selection for region of interest radiologic imaging.

Region of interest (ROI) fluoroscopic techniques have the potential for improving image quality and greatly reducing radiation dose. The utility of ROI imaging depends on the material used for the filter. Various materials for x-ray beam shaping ROI filters are evaluated using a computer simulation to determine their transmission and effect on radiographic image quality as a function of kVp. Selected measurements were also performed to verify the results of the simulation. Rare earth materials such as gadolinium (Gd) showed distinct advantages over conventional filter materials such as copper (Cu). The effects of Gd and Cu filters on transmission, contrast, and contrast to noise ratio using a CsI image receptor were investigated for beams of 40 to 100 kVp for a broad range of thicknesses of iodine, barium, bone, fat, and for a range of phantom thicknesses. The results were the following: (1) Gd generally increases image contrast while Cu substantially decreases contrast for all materials at all kVp's; (2) Gd has much less variation in transmission with kVp than Cu; and (3) for the same signal in the ROI and with thicknesses of filters which provides the same entrance phantom kerma, Gd provides higher contrast to noise ratio than Cu for lower kVp values while Cu provides higher contrast to noise ratio at higher kVp values. Analytical results of transmission and contrast improvement as a function of iodine concentration, filter thickness, and kVp compared favorably with experimental values. In conclusion, Gd appears to be the filter material of choice when it is important to maintain a higher level of image contrast in the periphery, such as during angiography, since image contrast is not degraded with Gd while it is degraded substantially with Cu. Although contrast is degraded with Cu, Cu may be the preferred filter material for higher kVp procedures such as GI exams since the detected signal in the periphery would be substantially higher permitting either greater reduction in patient dose or a decrease in image noise for the same dose reduction.