Digital subtraction radiography has proved to be a powerful technique in several areas, ie. for analyzing images after injection of contrast material in angiography. Digital subtraction has not been applied to the analysis of bone radiographs because it has not been possible to maintaining proper registration of the object in serial radiographs. We have undertaken studies to determine the resolution that would be required to make digital subtraction useful in the analysis of bone radiographs, and have collaborated in studies developing hardware and software for digital subtraction of bone radiographs. Preliminary studies in cadaveric models of erosive bone diseases and in normal volunteers have demonstrated the ability to detect and quantitate localized lesions that represent bone loss of less than 3 mg, lesions that are below the limit of detectability on plain films. A protocol for clinical study of erosive bone lesions in patient volunteers with rheumatoid arthritis has been approved, and patients will soon be registered for this project.
