Image fusion in dual energy computed tomography for detection of various anatomic structures – Effect on contrast enhancement, contrast-to-noise ratio, signal-to-noise ratio and image quality

https://doi.org/10.1016/j.ejrad.2011.02.023Get rights and content

Abstract

Objective

The purpose of this study was to evaluate image fusion in dual energy computed tomography for detecting various anatomic structures based on the effect on contrast enhancement, contrast-to-noise ratio, signal-to-noise ratio and image quality.

Material and methods

Forty patients underwent a CT neck with dual energy mode (DECT under a Somatom Definition flash Dual Source CT scanner (Siemens, Forchheim, Germany)). Tube voltage: 80-kV and Sn140-kV; tube current: 110 and 290 mA s; collimation-2 × 32 × 0.6 mm. Raw data were reconstructed using a soft convolution kernel (D30f). Fused images were calculated using a spectrum of weighting factors (0.0, 0.3, 0.6 0.8 and 1.0) generating different ratios between the 80- and Sn140-kV images (e.g. factor 0.6 corresponds to 60% of their information from the 80-kV image, and 40% from the Sn140-kV image). CT values and SNRs measured in the ascending aorta, thyroid gland, fat, muscle, CSF, spinal cord, bone marrow and brain. In addition, CNR values calculated for aorta, thyroid, muscle and brain. Subjective image quality evaluated using a 5-point grading scale. Results compared using paired t-tests and nonparametric-paired Wilcoxon–Wilcox-test.

Results

Statistically significant increases in mean CT values noted in anatomic structures when increasing weighting factors used (all P  0.001). For example, mean CT values derived from the contrast enhanced aorta were 149.2 ± 12.8 Hounsfield Units (HU), 204.8 ± 14.4 HU, 267.5 ± 18.6 HU, 311.9 ± 22.3 HU, 347.3 ± 24.7 HU, when the weighting factors 0.0, 0.3, 0.6, 0.8 and 1.0 were used. The highest SNR and CNR values were found in materials when the weighting factor 0.6 used. The difference CNR between the weighting factors 0.6 and 0.3 was statistically significant in the contrast enhanced aorta and thyroid gland (P = 0.012 and P = 0.016, respectively). Visual image assessment for image quality showed the highest score for the data reconstructed using the weighting factor 0.6.

Conclusion

Different fusion factors used to create images in DECT cause statistically significant differences in CT value, SNR, CNR and image quality. Best results obtained using the weighting factor 0.6 for all anatomic structures used in this study.

Introduction

Compton scatter and the photoelectric effect are the two main mechanisms responsible for the absorption and scattering of photon energy range used in CT. Inside the energy range considered the total cross section of the Compton Effect is almost independent of photon energy, whereas the total cross section of the photoelectric effect is strongly energy-dependent. The differentiation of material in computed tomography (CT) based on their X-ray attenuation as quantified in Hounsfield Units and displayed in shades of gray at different window levels in normal CT scans [1]. A previous study [2] demonstrated that a low tube voltage (80-kV) scan can provide better contrast and conspicuity than a high voltage (140-kV) scan for the detection of hyper vascular liver tumors, as the low tube voltage scan takes advantage of the attenuation property of iodinated contrast material at 80-kV.

However, in a previous study despite the fact that 80-kV CT images showed a higher contrast-to-noise ratio (CNR) of simulated hypervascular liver lesions, the low tube voltage scan also showed increased noise compared with the high voltage tube scan. In addition, although increasing tube current is able to decrease noise of the low tube voltage scan, the CT system is not able to provide sufficient radiation dose as desired. Those factors thus limit the widespread use of an 80-kV scan in clinical practice. However, with dual-energy CT (DECT), the noise of the 80-kV data offset by the decreased noise of the 140-kV data, and therefore, the difficulty with routine use of low-kV CT because of increased noise could be minimize with this DECT technique. As the use of DECT has recently increased, image fusion techniques using 140-kV images and 80-kV images with DECT can provide a way to increase the CNR [3], [4], [5], [6]. Theoretically, if we adequately fuse both low and high voltage images, we can obtain better images that therefore balance the advantages and disadvantages of both low and high voltage images according to the attenuation difference between the lesion or structures of main interest and the background organ. For example, a 0.3 weighting factor means that 30% of the image information is derived from the 80 kV image and 70% from the 140 kV image. As the weighting factor increases, the image looks like more an 80-kV image. These routinely provided fused images are making anatomic structures or pathology differentiation without the benefit of DE processing. Characterized by low image noise, such images create the impression of a 120-kV image [7]. This is because of a dedicated DE convolution kernel that draws 70% of the fused image from the 140-kV image and 30% from the 80-kV image [7]. The aims of this study were to differentiate various body structures even without the presence of contrast media, to differentiate contrast enhanced structures and its (or lesion) vascularity from otherwise dense material in parenchymatous organs and differentiation of contrast-enhanced vessels. Thus, in effect to evaluate the effect of using different weighting factors on differentiation of body tissues/materials with and with out contrast enhancement under a dual energy CT.

Section snippets

Scanning machine

CT scans obtained by using a recently introduced Second-generation dual-source CT scanner (Somatom Definition Flash; Siemens Healthcare, Forchheim, Germany). This scanner has two X-ray tubes that simultaneously revolve around the patient's body and equipped with two 38.4 mm detectors that each acquire 128 slices of image data. The scanning speed of this scanner is 43 cm/s and a temporal resolution of 75 ms. The tubes can be operated independent of each other with respect to kilovolt (kV) and

Quantitative analysis

Mean CT values increased with increasing weighting factors. The highest CT values were detected for the factor of 1.0 (100% 80 kVp) (Table 2, Fig. 2), compared with data sets reconstructed using a weighting factor of 0.0. Compared to 0.0 weighted images the mean CT values generated by a weighting factor of 1.0 were almost twice as high. The SNR showed the highest values for the data sets reconstructed with a weighting factor of 0.6, followed by the factor 0.3 or 0.8 depending on the anatomic

Discussion

The intensity of contrast enhancement is almost double in most of the cases increasing the weighting factor from 0.0 to 1.0 (Table 2). The variability of iodine enhancement explains why it is possible to differentiate iodine from other materials or substances that do not show such attenuation behavior. Two mechanisms, Compton scattering and the photoelectric effect, explain this phenomenon. First, the Compton Effect has little influence on diagnostic CT images so it is not important in this

Conclusion

In summary, in a single-phase examination it is possible to differentiate various anatomic structures or lesions more specifically with dual-energy fusion weighted images with out any additional scanning (additional scan contribute extra dose). In this study demonstrates that using different weighting factors in DECT causes statistically significant changes in contrast enhancement and image quality of anatomic structures. The differentiation of iodine can be regarded as a most promising and

Conflict of interest statement

None declared.

References (18)

  • D.R. Holmes et al.

    Evaluation of non-linear blending in dual-energy computed tomography

    Eur J Radiol

    (2008)
  • J.K. Tajik et al.

    Subsecond multisection CT of regional pulmonary ventilation

    Acad Radiol

    (2002)
  • T.R.C. Johnson et al.

    Material differentiation by dual energy CT: initial experience

    Eur Radiol

    (2007)
  • S.T. Schindera et al.

    Hypervascular liver tumors: low tube voltage, high tube current multi-detector row CT for enhanced detection—phantom study

    Radiology

    (2008)
  • T.R. Johnson et al.

    Material differentiation by dual energy CT: initial experience

    Eur Radiol

    (2007)
  • A. Graser et al.

    Dual energy CT: preliminary observations and potential clinical applications in the abdomen

    Eur Radiol

    (2009)
  • F.F. Behrendt et al.

    Image fusion in dual energy computed tomography: effect on contrast enhancement, signal-to-noise ratio and image quality in computed tomography angiography

    Invest Radiol

    (2009)
  • K.S. Kim et al.

    Image fusion in dual energy computed tomography for detection of hypervascular liver hepatocellular carcinoma: phantom and preliminary studies

    Invest Radiol

    (2010)
  • J. Hausleiter et al.

    Radiation dose estimates from cardiac multislice computed tomography in daily practice: impact of different scanning protocols on effective dose estimates

    Circulation

    (2006)
There are more references available in the full text version of this article.

Cited by (40)

  • Clinical utility of quantitative dual-energy CT iodine maps and CT morphological features in distinguishing small-cell from non-small-cell lung cancer

    2019, Clinical Radiology
    Citation Excerpt :

    Therefore, qualitative and quantitative metrics derived from DECT would be of value in assisting clinical estimation of tumour subtypes to contribute to initial treatment selection and prognostic prediction, especially when invasive procedures are unavailable or pathological results indeterminate.21 In pursuit of high contrast-to-noise ratio and fine image quality in evaluating anatomical structures in DECT, fused images, taking advantage of both low- and high-voltage images, are frequently applied to assess morphological features.15 In the current study, 27 CT features were analysed on virtual 120 kV images and it revealed nine morphological features more likely to be indicative of SCLC rather than NSCLC, including central location, large size, absence of spiculation, lymphadenopathy, confluent mediastinal lymphadenopathy, encasement of mediastinal structures, vascular involvement, lung atelectasis and/or obstructive pneumonia, and homogeneous enhancement.

  • Microstructure and micropore formation in a centrifugally-cast duplex stainless steel via X-ray microtomography

    2019, Materials Characterization
    Citation Excerpt :

    In addition to some commercial lab-scale CT systems such as ZEISS Xradia series (Carl Zeiss Microscopy, München, Germany), some researchers have also reported on the design and development of lab-scale X-ray microtomography facilities, e.g. Mertens et al. [34] and Egan et al. [37]. In recent years, dual-energy computed tomography (DECT), which uses both the normal X-ray and a second less powerful X-ray to obtain an additional attenuation measurement and enhance the absorption contrast [38], has been implemented in both biomedical research fields [39–43] and security research fields [44–47]. This technique was also applied to materials science and engineering research such as metals [48,49] and composite materials [50,51] due to its advantage on separating the microstructure characteristics of the microconstituents.

  • Chapter 8 On the Horizon: Advanced Imaging Techniques to Improve Noninvasive Assessment of Cervical Lymph Nodes

    2017, Seminars in Ultrasound, CT and MRI
    Citation Excerpt :

    Early generations of DECT scanners generated higher radiation exposures of up to 3 times compared to conventional CT scanners.101 Subsequent improvements in image acquisition and postprocessing techniques significantly decreased radiation doses in DECT,102,103 and currently DECT can be performed without additional dose to the patient. Using DECT, Lam et al used different VMI energy levels ranging from 40-140 keV in 5-keV increments in 30 patients with HNSCCs.

  • Three-dimensional microstructure-based micromechanical modeling for TC6 titanium alloy

    2017, Materials Science and Engineering: A
    Citation Excerpt :

    Mechanical-property determination of constituent phases and development of the corresponding three-dimensional (3D) microstructural model are difficult, owing to this complex microstructure. To overcome these difficulties, dual-energy micro-computed tomography (DE-MicroCT) [13,14] was employed in this study to reconstruct the 3D microstructure model of the TC6 titanium alloy. This method has the unique ability to enhance the absorption contrast among constituent phases that have similar mass densities.

View all citing articles on Scopus
View full text