Impact of software parameter settings on image quality of virtual grid processed radiography images : a contrast-detail phantom study

Objectives Mobile radiography systems are commonly used in the intensive care unit. The use of a physical antiscatter grid with these systems is uncommon because of drawbacks. In 2015, Virtual Grid (Fujifilm, Tokyo, Japan) became available for chest and abdomen examinations. In this study, we compared image quality (IQ) with a contrast-detail phantom (CDRAD 2.0; Artinis Medical Systems, Zetten, the Netherlands) of digital radiographs acquired without any grid (gridless) with those corrected for scatter by either software (SW)-based scatter correction (Virtual Grid) or a physical grid (grid). Furthermore, we determined the optimal Virtual Grid settings that lead to the best contrast-detail IQ score (inverse IQ figure). Materials and Methods Images were obtained with a cassette spot film device with an inserted portable flat-panel detector (Fujifilm, Tokyo, Japan). The CDRAD phantom was sandwiched between polymethylmethacrylate (PMMA) with total thicknesses of 12, 16, 21, and 26 cm to simulate patient attenuation and scatter. Tube voltages of 81, 90, 109, and 125 kVp were used to make the radiographs. In total, 12 different Virtual Grid settings (grid ratio, line pairs (lp)/cm, and type of interstitial material) were applied for every phantom thickness and tube voltage. Results An average increase of 32% in IQ was obtained when Virtual Grid images with a SW grid ratio 10:1 were compared with gridless images (P< 0.001). Increasing the SW grid ratio to 20:1 resulted in a further increased IQ. With a phantom thickness of 12 cm PMMA, Virtual Grid outperformed the removable physical grid presented in the cassette spot film device. The linear mixed-effects model showed that IQ is mainly affected by PMMA, tube voltage, and the SW grid ratio. Conclusions Virtual Grid improves images obtained without physical grid for a wide range of experimental conditions. Despite the different possible settings of the Virtual Grid SW, the most important parameter affecting IQ is the SW grid ratio