Effect of miscentering and low-dose protocols on contrast resolution in computed tomography head examination

https://doi.org/10.1016/j.compbiomed.2020.103840Get rights and content

Abstract

Background

Unoptimized protocols, including a miscentered position, might affect the outcome of diagnostic in CT examinations. In this study, we investigate the effects of miscentering position during CT head examination on the signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR).

Method

We simulate the CT head examination using a water phantom with a standard protocol (120 kVp/180 mAs) and a low dose protocol (100 kVp/142 mAs). The table height was adjusted to simulate miscentering by 5 cm from the isocenter, where the height was miscentered superiorly (MCS) at 109, 114, 119, and 124 cm, and miscentered inferiorly (MCI) at 99, 94, 89, and 84 cm. Seven circular regions of interest were used, with one drawn at the center, four at the peripheral area of the phantom, and two at the background area of the image.

Results

For the standard protocol, the mean CNR decreased uniformly as table height increased and significantly differed (p < 0.05) at +20 cm for MCS (435.70 ± 9.39) and −20 cm for MCI (438.91 ± 10.94) from the isocenter. Similarly, significant reductions (p < 0.05) were also noted for SNR for MCS (at +20 cm) and MCI (at −20 cm). For the low dose protocol, both CNR and SNR were significantly reduced (p < 0.05) at table heights of +20 and −20 cm from the isocenter.

Conclusion

Miscentering is proven to significantly affect the image quality in both low and standard dose protocols for head CT procedure. This study implies that accurate patient centering is one of the approaches that can improve CT optimization practice.

Introduction

In diagnostic radiology, image quality is one of the essential criteria for visualizing the abnormality. In any imaging modality, image quality is defined by the contrast, noise, signal, and sharpness of the images [1]. Generally, imaging performance relies on the contrast resolution and spatial resolution of the output images. Image quality can be quantified by both qualitative and quantitative measurement techniques. The pixelated signal of CT images is expressed in Hounsfield Units (HUs), a rescaled normalized function of the linear attenuation coefficient [2]. This HU value is derived from the CT Hounsfield scale, which sets a HU of 0 for water and a HU of 1000 for air. In clinical practice, human fat tissue has a HU of around −100 HU, soft tissues of 35–70 HU, and dense bone generally above 1000 HU.

CT image quality depends on image noise, image contrast, spatial resolution, and artifacts [1]. These factors determine the ability to perceive low-contrast structures and the visibility of details, which depend on the diagnostic task. Noise is a critical component in determining quality in medical imaging and is considered an excellent indicator for measuring CT image quality quantitatively [3,4]. Notably, the use of a digital imaging detector might cause electronic noise and Poisson noise in the system. Generally, the absorbed dose is inversely proportional to the noise [2,5]. This inverse relationship limit allows researchers to reduce the dose in clinical X-ray imaging. Hence, low dose but high-quality CT images are the main principles in CT optimizations. For example, adjusting the current tube line according to the patient's size might lead to an appropriate balance between image noise and radiation exposure [3]. Therefore, technologists play a vital role in the selection of scanning parameters, as well as positioning techniques [6].

Several researchers have reported that miscentering during the CT procedure is responsible for increasing the noise, which might affect diagnosis [[7], [8], [9], [10], [11], [12]]. Almohiy et al. found that mean image noise was increased significantly in a miscentered position as compared to optimized positioned patients who underwent intracranial CT examination [7]. Furthermore, a prior study on clinical CT chest and abdominal examination reported that an average miscentering of 2.1 cm below the isocenter results in a 6% increase in image noise [9]. Miscentering also magnifies the image as the patients are closer to the X-ray tube. This was demonstrated by Filev et al., where the image of the modified Livermore phantom showed 33% magnification from the isocenter image a table height of +6 cm [10]. However, this study was based on retrospective data and included only two categories, poorly centered and accurately centered. Thus, proper prospective experimental studies on image quality with precise miscentering, either superiorly or inferiorly, are needed.

Head injury imaging requires sufficient contrast as several internal structures lie within the head region. CT head is considered a gold standard for initial evaluation of the traumatic head injury; thus, it is essential to explore issues on head miscentering during CT. Therefore, we conducted a prospective experimental study to investigate the effects of miscentering CT head examination on image quality, especially towards image noise and contrast resolution.

Section snippets

Materials and method

In this study, we performed a CT head examination on a 16-cm diameter water phantom using a 64-slice multidetector CT scanner (MDCT) (CT Aquilion; Canon Medical Systems, Japan), a 3rd generation CT model. The scanner has an aperture of 72 cm, and the distance from the middle point to the inner surface of the gantry is 36 cm, as shown in Fig. 1. The scanner was composed of 896 elements, and one reference element arranged in a single row, with a maximum coverage for the helical scan of 1950 mm

Experimental setup

In this study, the effect of miscentering on image quality was evaluated using a standardized water phantom with a diameter of 16-cm. Since the size of the water phantom was smaller than body phantoms, the table height adjustment was made vertically according to the method described by Filev et al., with slight modifications [10]. The table height was set at 104 cm from the floor, which represents an isocenter position. Vertical adjustments from the isocenter were made, with a gap of 5 cm for

Image quality assessment

Three axial slices were selected from a single volume data and reconstructed into thicknesses of 15, 30, and 45 mm. On each selected image, there were five regions of interest (ROIs) inside the reconstructed image (ROI 1–5) and two ROIs in the background area (ROI 6–7) (Fig. 2). Mean CT numbers and SD were obtained directly from the ROIs and averaged before being recorded in the standardized form. Also, both values were used for the derivation of contrast-to-noise ratio (CNR) and

Statistical analysis

Data were statistically analyzed using the Statistical Package for Social Sciences (SPSS) software version 22.0 (IBM SPSS, Armonk, NewYork, USA). The Shapiro-Wilk test was used to determine the normality of the CNR and SNR data. The quantitative variables were presented descriptively and were expressed as a mean ± standard deviation (SD). Differences between the groups were determined using analysis of variance (ANOVA), while for data that were not normally distributed, the Kruskal Wallis test

Results

In this study, CT head examinations were miscentered by changing the table height in the gantry. Table 2, Table 3 indicate CT noise value, CNR, and SNR for standard and low dose protocols along the MCS and MCI, respectively. For the standard protocol, the lowest image noise was at the isocenter, with a value of 2.82 ± 0.12. Also, CNR (501.12 ± 1.50) and SNR (0.20 ± 0.04) were the highest at the isocenter. The highest mean noise values for MCS and MCI were 3.12 ± 0.09 and 3.07 ± 0.05,

Discussion

CT head examination can detect abnormality for a patient with a severe traumatic head injury, where its sensitivity ranges between 68% and 94%. These benefits have contributed to the extensive use of CT in the clinical environment, especially in an initial examination for trauma cases [13]. Thus, researchers are exploring issues related to CT radiation dose and image quality. Various clinical and technical settings might affect these two critical parameters, including patient centering. The

Conclusion

Here we demonstrate that extreme miscentering significantly affects image quality in both standard and low dose protocols for CT head procedures. Hence, a miscentered position should be avoidable as it can degrade the diagnostic value of the images, thereby leading to false-negative findings. To conclude, accurate patient centering can be considered as a valid CT optimization approach.

Declaration of competing interest

The authors declare that there is no conflict of interests regarding the publication of this paper.

Acknowledgment

The authors also wish to acknowledge support from Geran Putra of Universiti Putra Malaysia with the project number GP/IPM/UPM/9619800.

Muhammad Khalis Abdul Karim. received his B.Sc (Hons) in Diagnostic Imaging and Radiotherapy from Universiti Kebangsaan Malaysia (UKM) in 2008 and M.Sc in Physics from Universiti Teknologi Malaysia (2013). In December 2016, he completed his Ph.D in Physics at the Universiti Teknologi Malaysia (UTM). Khalis has expertise in medical physics particularly radiation dosimetry in medical imaging and has been appointed as an advisory member in several technical committee during his work. Aside from

References (18)

  • F. Zarb et al.

    Image quality assessment tools for optimization of CT images

    Radiography

    (2010)
  • L.W. Goldman

    Principles of CT: radiation dose and image quality Vol2

    J. Nucl. Med. Technol.

    (2007)
  • R.A. Brooks

    A quantitative theory of the Hounsfield unit and its application to dual energy scanning

    J. Comput. Assist. Tomogr.

    (1977)
  • S.T. Bache et al.

    Quantitation of clinical feedback on image quality differences between two CT scanner models

    J. Appl. Clin. Med. Phys.

    (2017)
  • A.H. Goenka et al.

    Image noise, CNR, and detectability of low-contrast, low-attenuation liver lesions in a phantom: effects of radiation exposure, phantom size, integrated circuit detector, and iterative reconstruction

    Radiology

    (2016)
  • T. Kaasalainen et al.

    Effect of patient centering on patient dose and image noise in chest CT

    Am. J. Roentgenol.

    (2014)
  • M.K.A. Karim et al.

    Assessment of knowledge and awareness among radiology personnel regarding current computed tomography technology and radiation dose

    J. Phys. Conf. Ser.

    (2016)
  • H. Almohiy et al.

    Correct patient centering increases image quality without concomitant increase of radiation dose during adult intracranial computed tomography

    J. Med. Imag. Radiat. Sci.

    (2016)
  • M.K.A. Karim et al.

    A comparative study of radiation doses between phantom and patients via CT angiography of the intra-/extra-cranial, pulmonary, and abdominal/pelvic arteries

    Radiol. Phys. Technol.

    (2019)
There are more references available in the full text version of this article.

Cited by (7)

View all citing articles on Scopus

Muhammad Khalis Abdul Karim. received his B.Sc (Hons) in Diagnostic Imaging and Radiotherapy from Universiti Kebangsaan Malaysia (UKM) in 2008 and M.Sc in Physics from Universiti Teknologi Malaysia (2013). In December 2016, he completed his Ph.D in Physics at the Universiti Teknologi Malaysia (UTM). Khalis has expertise in medical physics particularly radiation dosimetry in medical imaging and has been appointed as an advisory member in several technical committee during his work. Aside from working in clinical environment, Khalis also actively involved giving lecture in several workshops on radiation protection and nuclear security. To date, Khalis has authored/co-authored 35 international journals (with impact factor ranging from 0.2 to 5.1), 10 international proceedings and 3 national proceedings. He has reviewed 20 international journal papers from reputable and well-established publisher. His other research interests include medical imaging, radiation dosimeters and image processing.

View full text