Image quality dependence on in-plane positions and directions for MDCT images

https://doi.org/10.1016/j.ejrad.2009.03.060Get rights and content

Abstract

Objective

The present study was performed to examine the dependence of image quality on in-plane position and direction in computed tomography (CT) imaging using the modulation transfer function (MTF), noise power spectrum (NPS) and analysis of signal-to-noise ratio (SNR). For detailed analysis of SNR, the low-contrast detectability was compared using simulated small low-contrast objects.

Materials and methods

Three models of multidetector-row CT (MDCT) were employed. The measurement positions for MTF were set to the isocentre and several peripheral areas, and NPS and SNR were calculated for the isocentre and 128 mm off-centre. To evaluate directional dependence, the one-dimensional physical properties were measured separately in the radial and azimuthal directions. Seven radiological technologists also performed a perceptual detection study at the different in-plane positions using computer-simulated low-contrast images.

Results

The results of MTF and SNR differed between the isocentre and the peripheral area. The MTF values also tended to decrease with distance from the isocentre, and the SNR values in the low frequency range for the peripheral area were superior to those for the isocentre. In the detection study, the low-contrast detectability in the peripheral area was 13–40% higher than the value in the isocentre.

Conclusion

The results of the present study indicated that clinical CT images have remarkable non-uniformity of image quality. Therefore, the detailed analysis performed in this study will provide useful information for the development of advanced image processing applications, such as computer-aided diagnosis (CAD) and de-noising of CT images.

Introduction

The data acquisition procedure in CT systems currently typified by multidetector-row CT (MDCT) [1], [2], [3] differs from general radiographic systems, such as flat-panel detectors (FPDs) [4], [5]. In general radiographic systems, a two-dimensional detector carries out data acquisition for the two-dimensional distribution of X-ray projection directly. On the other hand, in CT systems based on the Radon theorem, an image is reconstructed from many one-dimensional projection data measured at various projection angles (0–360°) around the object. Therefore, the CT images are thought to have characteristic physical properties that differ from those obtained using the general digital radiography system. For example, there are some remarkable image quality differences between the central and peripheral areas in axial CT images. For example, the depiction of edges in the peripheral area is often inferior to that close to the centre, and also a lower level of image noise can be seen in the peripheral area.

A number of previous analytical reports on CT image evaluation have described the image quality results measured in the vicinity of the isocentre of scanning area. However, there have been few detailed reports of image quality analysis including the peripheral region. Methods to obtain such detailed image quality may allow us to obtain more useful information and better scan parameters than current recommendations. Furthermore, such detailed physical data may provide useful information for the advanced development of data processing procedures using CT images. Thus, it may be meaningful to clarify image quality of CT systems over the entire scanning area. In the present study, we analysed image quality dependence on the in-plane positions and directions using the modulation transfer function (MTF) [6], [7], [8], [9], [10], [11], noise power spectrum (NPS) [12], [13] and signal-to-noise ratio (SNR) [13], [14], [15] in three MDCT systems. In addition, the perceptual low-contrast detectability was examined using computer-simulated low-contrast objects implanted into uniform noise images.

Section snippets

Multidetector-row CT (MDCT) systems

The three MDCT systems evaluated in this study were a SOMATOM Volume Zoom (VZ; Siemens Medical Systems, Forchheim, Germany), a Siemens SOMATOM Emotion6 (Emotion6; Siemens Medical Systems) and an Aquilion16 (Toshiba Medical Systems, Tokyo, Japan). The images obtained using these CT systems were transferred to a computer for analysis using the DICOM (Digital Image Communication of Medicine) image transfer protocol.

Measurement of MTF

The MTF as the resolution property index was measured using a thin tungsten wire

Results

Fig. 4 shows results of MTF dependence on the in-plane position for the three MDCTs. The graphs are shown separately for the x- and y-directions for each MDCT. All the MTF decreased with distance from the isocentre. In addition, the degrees of decrease in x- and y-directions were different significantly, and the degree in each direction showed similar trends among the three MDCTs.

The rates of decrease in spatial frequency at 50% MTF in the most distant position of 192 mm off-centre were 35% (x

Discussion

CT systems can provide various image properties by selecting various built-in image reconstruction kernels. Generally, a standard reconstruction kernel is used for soft tissue parts, such as the abdomen and mediastinum. On the other hand, a high-resolution reconstruction kernel is used for organs with fine structures, such as the inner ear, lungs and bones. These two reconstruction kernels differ significantly in terms of their spatial frequency responses and edge-enhancement properties.

References (22)

  • S. Stefan et al.

    Spiral interpolation algorithm for multislice spiral CT. I. Theory

    IEEE Trans Med Imaging

    (2000)
  • T. Fuchs et al.

    Spiral interpolation algorithm for multislice spiral CT II measurement and evaluation of slice sensitivity profiles and noise at a clinical multislice system

    IEEE Trans Med Imaging

    (2000)
  • K. Taguchi et al.

    Algorithm for image reconstruction in multi-slice helical CT

    Med Phys

    (1998)
  • M.J. Yaffee et al.

    X-ray detectors for digital radiography

    Phys Med Biol

    (1997)
  • J.H. Siewerdesn et al.

    Empirical and theoretical investigation of the noise performance of indirect detection, active matrix flat-panel imagers (AMFPIs) for diagnostic radiography

    Med Phys

    (1997)
  • C.J. Bischof et al.

    Modulation transfer factor of the EMI CT head scanner

    Med Phys

    (1977)
  • P.M. Joseph et al.

    The influence of modulation transfer shape on computed tomographic image quality

    Radiology

    (1982)
  • E.L. Nickoloff

    Measurement of the PSF for a CT scanner: appropriate wire diameter and pixel size

    Phys Med Biol

    (1988)
  • American Association of Physicists in Medicine. Specification and acceptance testing of computed tomographic scanners....
  • J.M. Boone

    Determination of the presampled MTF in computed tomography

    Med Phys

    (2001)
  • M.L. Giger et al.

    Investigation of basic properties in digital radiography. 1. Modulation transfer function

    Med Phys

    (1984)
  • Cited by (15)

    • Optimisation of an MDCT abdominal protocol: Image quality assessment of standard vs. iterative reconstructions

      2014, Physica Medica
      Citation Excerpt :

      The MTF analysis results are depicted in Fig. 2 for certain coupled-smooth and medium-smooth filters (no extended scale). The data for the FBP filters are comparable to Boedeker [18,22], who constructed the extended scale acquisitions, because few differences can be attributed to the scale or the off-axis position of the tungsten wire (see, e.g., [21]). A comparison showed significant differences (paired t-test, p < 0.05) for certain coupled filters even if the differences were small and less than 0.007.

    • Location dependency of the spatial resolution of cone beam computed tomography for dental use

      2013, Oral Surgery, Oral Medicine, Oral Pathology and Oral Radiology
    View all citing articles on Scopus
    1

    Tel.: +81 76 265 2528.

    2

    Tel.: +81 76 265 2524.

    3

    Tel.: +81 562 93 2904.

    View full text