Development of image reconstruction algorithm with iterative method using maximum-likelihood expectation-maximization (MLEM) for chest digital tomosynthesis (CDT) X-ray system

Abstract Digital tomosynthesis system (DTS) has been increasingly used in chest X-ray imaging with a limited-projection angle image reconstruction method. Image reconstruction methods in chest digital tomosynthesis (CDT) have fundamental impacts on image characteristics. Among these methods, iterative algorithm provides the additional benefits of chest X-ray imaging. However, there is almost no research that maximum-likelihood expectation-maximization (MLEM) method in CDT system. Thus, the purpose of this study was (1) to develop MLEM reconstruction algorithm based on CDT system and (2) to quantitatively compare the image characteristics with respect to the total angular range and iteration number as applied this system. For that purpose, we performed both simulation using MATLAB program and experimental studies. Tomography was performed linearly with a maximum ±30° total angular range and then we acquired Shepp-Logan and PMMA-Aluminum (Al) phantom images using 120 kVp X-ray source for simulation and experiment, respectively. We evaluated and compared the MLEM reconstructed image characteristics using root-mean-square error (RMSE) and signal difference-to-noise ratio (SDNR) with various total angular ranges and iteration numbers. According to the results, we can acquire improved image characteristics as the number of iterations and the number of projections increased. Also, we found optimized iteration number for 20 iteration numbers in our CDT system geometry because there were few changes of image characteristics over 20 iteration numbers. In conclusion, we have presented usefulness and effectiveness of developed CDT image reconstruction algorithm.

[1]  James T Dobbins,et al.  Digital x-ray tomosynthesis: current state of the art and clinical potential. , 2003, Physics in medicine and biology.

[2]  L. Shepp,et al.  Maximum Likelihood Reconstruction for Emission Tomography , 1983, IEEE Transactions on Medical Imaging.

[3]  D. G. Grant.  Tomosynthesis: a three-dimensional radiographic imaging technique. , 1972, IEEE transactions on bio-medical engineering.

[4]  J. Dobbins.  Tomosynthesis imaging: at a translational crossroads. , 2009, Medical physics.

[5]  I. Sechopoulos.  A review of breast tomosynthesis. Part II. Image reconstruction, processing and analysis, and advanced applications. , 2013, Medical physics.

[6]  J. Boone,et al.  CT dose index and patient dose: they are not the same thing. , 2011, Radiology.

[7]  A. Celler,et al.  Comparison of quantitative accuracy of three edge-preserving image reconstruction strategies , 2011, 2011 IEEE Nuclear Science Symposium Conference Record.

[8]  Tianpeng Wang,et al.  A dual-view digital tomosynthesis imaging technique for improved chest imaging. , 2015, Medical physics.

[9]  Michael Grass,et al.  Filter calculation for x-ray tomosynthesis reconstruction. , 2012, Physics in medicine and biology.

[10]  I. Sechopoulos.  A review of breast tomosynthesis. Part I. The image acquisition process. , 2013, Medical physics.

[11]  Irène Buvat,et al.  Simulation-based evaluation and optimization of a new CdZnTe gamma-camera architecture (HiSens) , 2010, Physics in medicine and biology.

[12]  Y. Lee,et al.  Preliminary study on a chest digital tomosynthesis: development and evaluation , 2015 .