Monte Carlo study of grid performance in diagnostic radiology: task dependent optimization for screen-film imaging.

An optimization of anti-scatter grid design using Monte Carlo techniques in diagnostic radiology is presented. The criterion for optimization was to find the combinations of the grid parameters (lead strip width, grid ratio and strip density) and tube potential which result in the lowest mean absorbed dose in the patient at fixed image contrast. The optimization was performed in three irradiation geometries, representing different scattering conditions (paediatric examinations, and two adult lumbar spine examinations) and was restricted to grids using fibre materials in covers and interspaces. Grid designs currently available were studied, as were designs which use thinner strips (< 30 microns) and higher grid ratios (> 18). It was found that grids with widely different strip densities (strips cm-1) and grid ratios can have good performance provided that they are used with appropriate strip width and tube potential. With increasing amounts of scatter, the optimal grid requires thicker strips and higher grid ratios. Increasing the strip density and using thinner strips and higher grid ratios are generally required. Grids with low strip density (25 strips cm-1) were found to be less sensitive to alterations in strip width. Optimal grids for paediatric radiology require thinner strips (10-20 microns) than those in currently available grids. Grids on the market are best suited for examinations of the adult body in anteroposterior (AP) view. In the adult lateral view, representing the largest scattering volume, higher grid ratios (> 18) than those in existing grids would be optimal. Examples of good grid designs are given for each examination.