The clinical utility of a one-shot energy subtraction method for thoracic spine radiography.

BACKGROUND: The interpretation of thoracic spine X-rays is difficult because these images cannot clearly visualize the thoracic spine because of the overlap with soft tissues, such as the heart and pulmonary blood vessels. Thus, to improve the clarity of thoracic spine radiographs using existing radiograph equipment, we have investigated a one-shot energy subtraction method to visualize thoracic spine radiographs. Our objective was to evaluate whether the thoracic spine radiographs generated using this method could visualize the spine more clearly than the corresponding original thoracic spine radiographs. METHODS: This study included 29 patients who underwent thoracic spine radiographs. We used a one-shot energy subtraction method to improve the clarity of thoracic spine radiographs. Image definition was evaluated using vertebrae sampled from each region of the thoracic spine. Specifically, these were: Th1, Th5, Th9, and Th12. Image definition was assessed using a three-point grading system. The conventional and processed computed radiographs (both frontal and lateral views) of all 29 study patients were evaluated by 5 spine surgeons. RESULTS: In all thoracic regions on both frontal and lateral views, the processed images showed statistically significantly better clarity than the corresponding conventional images, especially at all sampling sites on the frontal view and T5 and 9 on the lateral view. CONCLUSIONS: Thoracic spine radiographs generated using this method visualized the spine more clearly than the corresponding original thoracic spine radiographs. The greatest advantages of this image processing technique were its ability to clearly depict the whole thoracic spine on frontal views and the middle thoracic spine on lateral views.

New computed radiography processing condition for whole-spine radiography.

Computed radiography (CR) has many advantages compared with conventional radiographs, especially in image processing. Although CR is being used in chest radiography and mammography, it has not been applied to spine imaging. The purposes of this study were to formulate a set of new CR processing parameters and to test whether the resultant whole-spine radiographs visualized the spine more clearly than conventional images. This study included 29 patients who underwent whole-spine radiographs. We used 3 image processing methods to improve the clarity of whole-spine radiographs: gradation processing, dynamic range control processing, and multi-objective frequency processing. Radiograph definition was evaluated using vertebrae sampled from each region of the whole spine, specifically C4, C7, T8, T12, and L3; evaluation of the lateral view also included the sacral spine and femoral head. Image definition was assessed using a 3-point grading system. The conventional and processed CR images (both frontal and lateral views) were evaluated by 5 spine surgeons. In all spinal regions on both frontal and lateral views, the processed images showed statistically significantly better clarity than the corresponding conventional images, especially at T12, L3, the sacral spine, and the femoral head on the lateral view. Our set of new CR processing parameters can improve the clarity of whole-spine radiographs compared with conventional images. The greatest advantage of image processing was that it enabled clear depiction of the thoracolumbar junction, lumbar vertebrae, sacrum, and femoral head in the lateral view.