Purpose: Digital tomosynthesis (DTS) has been applied to chest imaging to crudely separate objects at different depth levels. However, the spatial resolution in the posterior‐anterior (PA) direction is very poor. We implement a dual‐view DTS technique to improve the spatial resolution along the PA direction and compare the results with the single‐view DTS technique and the cone beam CT (CBCT) technique. Methods: Computer simulations and experiments were conducted with the CBCT, dual‐view DTS and single‐view DTS techniques using the same geometry. For simulations, clinical CT images obtained at 120 kVp were used to generate a digital chest phantom, with which the projection images were computed. For imaging experiments, an anthropomorphic chest phantom was scanned with a bench top experimental CBCT system. 300 projections were obtained for CBCT imaging using the FDK algorithm for reconstruction. 51 projections covering 60 degrees in the PA direction were used for single‐view DTS imaging using the iterative algorithm for reconstruction. 25 projection over 60 degrees in the PA direction and another 25 images over 60 degrees in the lateral direction were used for dual view DTS. Results: For sagittal slice images, the CBCT image closely resembled the original phantom image. The dual‐view DTS images were fairly accurate in depicting the shape and dimensions of the anatomy in the fulcrum, despite the presence of some artifacts. With the single‐view DTS images, the anatomy was poorly depicted in the PA direction. For coronal slice images, the single‐view DTS images appeared less blurred but were much more affected by artifacts from the off‐fulcrum objects, resulting in severe distortion. The dual‐view DTS images, while not free from artifacts, provides more accurate rendition of the anatomy. Conclusions: Our results show that the dual‐view DTS technique substantially improved the spatial resolution of the reconstructed images in the PA direction. This work was supported in part by grants CA104759, CA124585, and CA138502A1 from NIH‐NCI, a research grant CA00117 from NIH‐NIBIB, and a subcontract from NIST‐ATP.
ASJC Scopus subject areas
- Radiology Nuclear Medicine and imaging