Date Published: April 25, 2019
Publisher: Public Library of Science
Author(s): Chia-Hao Chang, Yu-Ching Ni, Syuan-Ya Huang, Ho-Hui Hsieh, Sheng-Pin Tseng, Fan-Pin Tseng, Guillem Pratx.
The aim of this study was to develop a geometric calibration method capable of eliminating the reconstruction artifacts of geometric misalignments in a tomosynthesis prototype with dual-axis scanning geometry. The potential scenarios of geometric misalignments were demonstrated, and their effects on reconstructed images were also evaluated. This method was a phantom-based approach with iterative optimization, and the calibration phantom was designed for specific tomosynthesis scanning geometry. The phantom was used to calculate a set of geometric parameters from each projection, and these parameters were then used to evaluate the geometric misalignments of the dual-axis scanning-geometry prototype. The simulated results revealed that the extracted geometric parameters were similar to the input values and that the artifacts of reconstructed images were minimized due to geometric calibration. Additionally, experimental chest phantom imaging results also indicated that the artifacts of the reconstructed images were suppressed and that object structures were preserved through calibration. And the quantitative analysis result also indicated that the MTF can be further improved with the geometric calibration. All the simulated and experimental results demonstrated that this method is effective for tomosynthesis with dual-axis scanning geometry. Furthermore, this geometric calibration method can also be applied to other tomography imaging systems to reduce geometric misalignments and be used for different geometric calibration phantom configurations.
Traditional digital tomosynthesis (DTS) systems have been commonly used as clinical diagnostic tools in various medical imaging applications such as digital breast tomosynthesis and digital chest tomosynthesis [1–4]. During DTS imaging, a series of projection images of objects are acquired at a limited angle. These systems can provide images of computed tomography (CT)-like quality with in-depth information and at lower dose of radiation . However, the image quality of traditional chest DTS is limited because it only uses single-axis scanning geometry, as indicated in Fig 1a and 1b. Some structures parallel to the scanning direction are blurred (red arrows) because of ghost artifact distortion . Most chest DTS systems used in clinical settings have only the head–foot (HF)-axis scanning direction, resulting in the overlapping of the spinal and thoracic–aorta area (red arrows). Other studies have indicated that expanding the area of the scanning coverage can be more effective than increasing the sampling density for improving tomosynthesis image quality [6, 7]. Hence, as shown in Fig 1c, these limitations can be partially overcome by using dual-axis scanning geometry with a larger scanning coverage area (yellow arrows) [7, 8]. Nevertheless, the requirement for geometry alignment is higher for the system with dual-axis scanning geometry than for traditional tomosynthesis with single-axis scanning geometry. Simulation studies have demonstrated that image artifacts are induced by the geometric misalignments of the system, including axis shift and tilt. For such systems, the positions of the two scanning axes and image receptor are difficult to line up with the mechanical alignment method. Hence, a geometric calibration method must be used to obtain high-precision geometric parameters of the tomosynthesis system with dual-axis scanning geometry.
In this study, a geometric calibration method was developed based on an iterative projection matrix based algorithm for a tomosynthesis system with dual-axis scanning geometry. To improve the calibration accuracy, we used the same scanning trajectory as the clinical chest imaging with 92 projections for the geometric calibration. The calibration time for 92 projections is less than 90 seconds. Errors of the input and extracted geometric parameters were compared based on simulated results. For the same geometry setup with the actual system, the simulated results indicated that the values for the extracted geometric parameters were similar to the input values. Additionally, a dual-axis scanning geometry tomosynthesis system with axis shift and axis tilt was simulated to evaluate the effects of geometric misalignments on the reconstructed images. The reconstructed images with and without geometric calibration were quantitatively compared. The simulated results revealed that the tomosynthesis system demonstrated a higher tolerance for geometric misalignments than did the traditional tomography system because tomosynthesis has a small magnification factor and a rotation axis proximal to the image receptor. However, the requirement of geometric alignment is higher for tomosynthesis with dual-axis scanning than for traditional tomosynthesis with single-axis scanning geometry. The simulated results also indicated that the extracted geometric parameters had excellent accuracy for tomosynthesis image reconstruction and can eliminate artifacts caused by system misalignments. Finally, to validate the performance of the geometric calibration method in experimental applications, the method was applied to the TomoDR system. After the experiment, tomosynthesis images with and without calibration were compared in different chest areas. The experimental results indicated that the method can effectively minimize artifacts and preserve object structures caused by system geometric misalignments. And the quantitative analysis result also indicated that the MTF can be further improved with the geometric calibration. In summary, the geometric calibration method can be used not only to extract geometric parameters in a tomosynthesis system but also can be applied to other tomography imaging systems to reduce the occurrence of geometric misalignments. Furthermore, this method can also be used for other phantom configuration with known 3D marker coordinates.