Performance and feasibility of single-grid dark-field radiography with a clinical x-ray source

Objective.X-ray dark-field imaging extends conventional radiography by capturing small-angle scattering from subresolution micro/nanostructures, offering sensitivity to pathological changes that remain invisible in standard attenuation images, such as early microstructural alterations associated with pulmonary, oncological, and musculoskeletal diseases. Building on recent progress in single-grid dark-field imaging, we demonstrate that scattering-sensitive contrast can be acquired using a clinical x-ray source with a standard, non-coherent cone beam, under conditions representative of routine diagnostic imaging. The purpose of this study is to investigate the capabilities of single-grid dark-field radiography in the aforementioned setting involving various phantom materials and varying acrylic thicknesses as a surrogate for patient size, assessing key performance metrics to inform clinical translation.Approach.Single-grid dark-field images of a phantom containing various material samples inducing x-ray dark-field signal and varying acrylic layers used as simplified soft-tissue surrogates were acquired under common clinical x-ray parameters to evaluate dark-field contrast, spatial resolution, processing time, and dose. In addition, we proposed and provided an automated image processing pipeline for single-grid dark-field radiography, enabling the retrieval of nearly artefact-free images without manual parameter tuning once calibrated.Main Results.Using the acrylic phantoms, single-grid and single-shot dark-field radiography was shown to be feasible for thicknesses up to 10 cm, producing images with detectable dark-field contrast differences while maintaining an estimated effective dose0.04 mSv. Our findings suggest that normalisation to a suitable reference sample, exhibiting a pronounced dark-field signal, represents a practical approach for enhancing measurement consistency in single-grid dark-field radiography when comparing data across different object positions or acquisition setups.Significance.Under clinically realistic imaging conditions, this study defines practical feasibility boundaries for single-grid dark-field radiography. These benchmarks provide guidance for pathological and specimen-based studies aimed at clinical translation of dark-field imaging as a complementary contrast channel in standard x-ray systems, with the potential to add diagnostically relevant subresolution information. In addition, pathways to further enhance dark-field contrast and address thickness-related limitations are highlighted, building on diagnostic benefits demonstrated in prior dark-field imaging studies.