The objective of this research is to develop methods for improving contrast in chest radiography and to evaluate the effectiveness of these improvements for the diagnosis and staging of cancer involving the lung. The methods for improving contrast will include highly efficient scatter rejection by improved antiscatter grids and scanning-slit techniques, and digital or optical image processing. Evaluation will be based on ROC nodule detection studies and statistically valid comparisons of the visibility of anatomic structures on clinical radiographs by experienced radiologist observers.
Specific aims i nclude: 1. Chest radiographs of 400-500 cancer patients will be recorded on very-wide-latitude film (DuPont WDR), using conventional and improved antiscatter grids. These films will be carefully documented for presence of nodules and other cancer-related disease and will provide a database of images for objective evaluation of techniques for improving contrast in clinical chest radiography. 2. Documented WDR-film chest radiographs will be used to evaluate the effects of a new optical image-processing technique and of highly efficient scatter rejection on nodule detection and visibility of anatomic structures. 3. Documented WDR-film chest radiographs will be use to evaluate the effects of digital image processing and image pixelation effects. Different image-processing algorithms combining unsharp masking and contrast enhancement, and different sampling aperture sizes and intervals will be evaluated for their effects on nodule detection and visibility of anatomic structure. 4. An optical image processor modeled after the principles of LogEtronics printers but with improved control of optical flare and more flexible adjustment of image-processing parameters will be designed, constructed, and evaluated for application to chest radiography. 5. Development and evaluation an electronic scanning-slit technique providing highly efficient rejection of scattered x-rays and veiling glare with large-field image intensifiers used for chest radiography will be continued. Advanced image-processing techniques, e.g., spatial filtration for contrast enhancement and dual-KV subtraction for separating bone vs. soft-tissue structures will be implemented with this device.
