Data correction for gantry-tilted local CT

https://doi.org/10.1016/j.compmedimag.2008.01.001Get rights and content

Abstract

The gantry-tilted helical cone-beam computed tomography (CT) has an inherent problem that the relative shift of the region of interest (ROI) blurs the reconstructed image. This problem becomes more serious in the gantry-tilted local CT imaging due to the nature of local scanning. This paper proposes a new method to improve the gantry-tilted local imaging by correcting the local scanning data. Computer simulations show that the proposed method can enhance the local imaging performance to a certain extent in terms of the image sharpening and artifacts reduction.

Introduction

In the past decade, the helical multi-slice CT has been investigated and developed for rapid, volumetric and high-resolution scanning in medical diagnosis [1], [2], [3], [4], [5]. The helical scanning mode provides a continuous scanning instead of the old step-and-shoot mode, while a cone-beam projection with a multi-slice detector can enhance the scanning speed and imaging resolution to a great extent.

In many circumstances, it is necessary to tilt the CT scanning gantry with respect to the patient and couch. This is to reduce the metal artifacts caused by tooth fillings, avoid unnecessary projection to eyeballs during a cerebrum scanning, or simply to deal with some special image plane which may not lie in the normal transversal plane. For example, the gantry is tilted in order to focus the projection on the curvature of the spine in the cervical spine tomography [6]. Currently, many medical CT scanners are equipped with tiltable gantries such as SIEMENS SOMATOM DRH Scanner which can conduct ±25° tilted scanning.

Due to the tilted gantry, the helical scanning geometry becomes more complex. The image quality may be degraded since the conventional imaging algorithms are not compatible with the gantry-tilted CT image reconstruction. Efforts have been made to improve the gantry-tilted helical imaging and the research on this problem has been active. In 2000, Hsieh proposed an error correction scheme for the 2D rebinning approximate reconstruction of the gantry-tilted helical cone-beam CT [6]. In 2003, Hein et al. proposed an FDK-type reconstruction algorithm for gantry-tilted cone-beam scanning [7], [8], which is based on the gantry-normal FDK reconstruction and without significant additional computation. In 2004, Noo et al. proposed a general reconstruction scheme for the gantry-tilted cone-beam CT [9]. It suggests to transform the gantry-tilted projection data into a virtually normal projection data which can be adopted by some 3D reconstruction algorithms.

The gantry-tilted CT scanning is often for some special diagnosis where the safety issue is more important. Compared with the normal scanning, it is more necessary to limit the radiation dose in the gantry-tilted scanning than in the normal scanning. The local tomography is an effective technique for limiting the radiation dose, which is to the reconstruct the local ROI images using only local projection data. Its main purpose is to limit the X-ray exposure to the patient [11], [12]. In this paper the local tomography is applied to the gantry-tilted cone-beam projection for safer diagnosis. A data compensation scheme is presented to deal with the deviation problem occurring in the local gantry-tilted reconstruction, so to improve the image reconstruction accuracy.

The outline of this paper is as follows. In Section 2 the geometry of the gantry-tilted helical multi-slice CT scanning is described and the proposed error correction method for dealing with the image degradation is presented. The local reconstruction procedure is presented in Section 3, followed by Section 4 which presents the simulation results of the proposed method in comparison with that of the existing methods.

Section snippets

Gantry-tilted helical multi-slice scanning

The geometry of gantry-tilted helical multi-slice scanning with a gantry-tilted angle θ is shown in Fig. 1, where the couch and patient are translated in the z direction while the X-ray source is rotating around the gantry circle which has axis zt. Two coordinates systems “ xyzt” and “ tszt” are defined in the gantry-tilted plane. The axis x and y are fixed on the gantry plane while t and s rotate with the X-ray source. The rotation angle between the axis s and y is denoted by β.

Referring

Reconstruction

Most of CT image reconstruction algorithms are based on the well known Fourier Slice Theorem, as followsf(x,y)=[Pβ(p)R(p)](p=xcosβ+ysinβ)dβ,where f(x,y) is the reconstructed attenuation function, Pβ(p) is the transform of the sinogram data, “ ” denotes the convolution, and R(p) is the ramp filter. The convolution with R(p) is as followsR(p)Pβ(p)=HddpPβ(p),where H() denotes Hilbert transform. Unlike the derivative operator d/dp, Hilbert transform is not a local operation. As a result, the

Simulation and evaluation

In this section, the computer simulation is implemented to evaluate the proposed data correction method and to compare with the existing correction methods [8]. The parameters of the simulate gantry-tilted scanner are listed in Table 1.

First, a low-contrast Shepp–Logan phantom is adopted for the simulation [1]. Its grey values in the main area ranges between [1.0, 1.04] with the grey value 0 in the outside dark area. Therefore, it is highly sensitive to imaging noise. The local ROI is scanned

Conclusion

The imaging performance of gantry-tilted system can be degraded by its geometric nature yielding deviation between reconstruction isocenter and the gantry isocenter. To compensate this deviation, this paper proposes a linear interpolation based data correction scheme to form a corrected data set for local reconstruction. In comparison with the existing methods, the new approach offers a better reconstruction performance which is demonstrated in the computer simulation results. In addition to

Hongzhu Liang: Received B.Eng. degree in Xi’an Jiaotong University in 2000, and M.Eng. degree in Huazhong University of Science and Technology in 2003, both in China. Since 2003 he has been with School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore, as a research student pursuing Ph.D. degree. His research is mainly on the algorithm for the helical computed tomographic imaging.

References (13)

  • Ge Wang et al.

    A general cone-beam reconstruction algorithm

    IEEE Trans Med Imag

    (1993)
  • Hu. Hui

    Multi-slice helical CT: Scan and reconstruction

    Med Phys

    (1999)
  • K. Taguchi et al.

    Algorithm for image reconstruction in multi-slice helical CT

    Med Phys

    (1998)
  • S. Schaller et al.

    Spiral interpolation algorithms for multi-slice spiral CT I: Theory

    IEEE Trans Med Imag

    (2000)
  • S. Schaller et al.

    A new approximate algorithm for image reconstruction in cone-beam spiral CT at small cone-angles

    (1997)
  • I. Hein et al.

    Feldkamp-based cone-beam reconstruction for gantry-tilted helical multi-slice CT

    Med Phys

    (2003)
There are more references available in the full text version of this article.

Cited by (1)

Hongzhu Liang: Received B.Eng. degree in Xi’an Jiaotong University in 2000, and M.Eng. degree in Huazhong University of Science and Technology in 2003, both in China. Since 2003 he has been with School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore, as a research student pursuing Ph.D. degree. His research is mainly on the algorithm for the helical computed tomographic imaging.

Cishen Zhang: Received the B.Eng. degree from Tsinghua University, China, in 1982 and Ph.D. degree in electrical engineering from Newcastle University, Australia, in 1990. Between 1971 and 1978, he was an electrician with Changxindian (February Seven) Locomotive Manufactory, Beijing, China. He carried out research work on control systems at Delft University of Technology, The Netherlands, from 1983 to 1985. After his Ph.D. study from 1986 to 1989 at Newcastle University, he was with the Department of Electrical and Electronic Engineering at the University of Melbourne, Australia as a lecturer, senior lecturer and associate professor and reader till October 2002. He is currently with the School Electrical and Electronic Engineering and School of Chemical and Biomedical Engineering at Nanyang Technological University, Singapore. His research interests include signal processing, medical imaging and control.

Ming Yan: Received the B.Eng. degree and M.Sc. degree in automation, both from Tsinghua University, China, in 2000 and 2003, respectively. Since 2003, he has been pursuing the Ph.D. degree at the Nanyang Technological University, Singapore. His research interests include computed tomography reconstruction algorithms and medical imaging.

View full text