Elsevier

Academic Radiology

Volume 25, Issue 4, April 2018, Pages 519-528
Academic Radiology

Original Investigation
How Well Does Dual-energy CT with Fast Kilovoltage Switching Quantify CT Number and Iodine and Calcium Concentrations?

https://doi.org/10.1016/j.acra.2017.11.002Get rights and content

Rationale and Objectives

Because it is imperative for understanding the performance of dual-energy computed tomography scanner to determine clinical diagnosis, we aimed to assess the accuracy of quantitative measurements using dual-energy computed tomography with fast kilovoltage switching.

Materials and Methods

Quantitative measurements were performed for 16 reference materials (physical density, 0.965–1.550 g/cm3; diameter of rod, 2.0–28.5 mm; iodine concentration, 2–15 mg/mL; and calcium concentration, 50–300 mg/mL) with varying scanning settings, and the measured values were compared to their theoretical values.

Results

For high-density material, the maximum differences in Hounsfield unit values in the virtual monochromatic images at 50, 70, and 100 keV were −176.2, 61.0, and −35.2 HU, respectively, and the standard deviations over short- and long-term periods were 11.1, 6.1, and 3.5 HU at maximum. The accuracy of the Hounsfield unit measurement at 50 and 70 keV was significantly higher (P < 0.05) with higher radiation output and smaller phantom size. The difference in the iodine and calcium measurements in the large phantom were up to −2.6 and −60.4 mg/mL for iodine (5 mg/mL with 2-mm diameter) and calcium (300 mg/mL) materials, and the difference was improved with a small phantom. Metal artifact reduction software improved subjective image quality; however, the quantitative values were significantly underestimated (P < 0.05) (−49.5, −26.9, and −15.3 HU for 50, 70, and 100 keV, respectively; −1.0 and −17 mg/mL for iodine and calcium concentration, respectively) compared to that acquired without a metal material.

Conclusions

The accuracy of quantitative measurements can be affected by material density and the size of the object, radiation output, phantom size, and the presence of metal materials.

Introduction

Recent advances in scanner technology have led to increased clinical use of dual-energy computed tomography (DECT). DECT utilizes two different energy spectra to allow the analysis of energy-dependent changes in the attenuation of different materials. This use of high and low photon energies allows the differentiation of materials and makes possible the quantitative measurement of variables such as iodine and calcium concentrations, and thus, the DECT can tell us how much of these materials are present in the region of interest (ROI) (1). Although conventional computed tomography (CT) is routinely assessed in patterns of gray scale expressed as Hounsfield units, an iodine density map can provide physiological information, which can be helpful in diagnosis. Moreover, virtual monochromatic images (VMIs) at a specified photon energy level can be generated during the processing of material-density image data by calculating the linear attenuation coefficient. These images are less affected by the beam-hardening effect, and thus provide more accurate Hounsfield unit values than conventional CT scanners with a polychromatic energy beam (2). The beam-hardening effect has been recognized as one of the major concerns of inaccurate Hounsfield unit measurement because the low-energy photons within a polychromatic beam can reduce vascular enhancement in larger patients (3).

The advantages of DECT in clinical use are now an active area of research. Thieme et al. reported that DECT with pulmonary parenchymal iodine mapping identified pulmonary perfusion with a close correlation with perfusion single-photon emission CT and concluded that this technique might potentially enhance diagnostic accuracy (4). The advantages of the iodine mapping are its smaller spatial resolution compared to single-photon emission CT (in general, 10–20 mm) and a perfect spatial match of CT images with iodine mapping. Hu et al. performed unenhanced dual-energy head CT in the emergency department and analyzed the virtual noncalcium and calcium overlay images to distinguish calcification from hemorrhage. In the report, the accuracy of DECT for the detection of hemorrhage was 99%, whereas that of conventional CT was 87% (5). In a study by Cha et al., DECT with metal artifact reduction software (MARS) markedly reduced metallic dental artifacts and improved image quality in the buccal area and the tongue area (6). Patients with head and neck cancer are expected to benefit from the improvement of the image quality using MARS because accurate diagnostic staging considering the primary site, the size, and the existence of metastatic lymph nodes is imperative for proper treatment. The success of such investigations depends on how accurately DECT quantifies the concentration of materials such as iodine and calcium and Hounsfield unit values.

In clinical practice, scanning parameters, such as tube current, helical pitch, and the use of MARS, are changed to ensure consistent image quality in any patient. Selection of the appropriate scan field of view (SFOV) is also important to ensure that the scanner uses the correct calibration and beam-hardening correction data (7). Without the optimal setting, quantitative measurement of equal concentrations of a material may varies because of CT artifacts. Moreover, precision of CT scanner performance is also important in any clinical examination because scanners can vary from day to day or even at different times during the same day (1). These findings highlight the critical importance of understanding the performance of a conventional CT scanner, as well as DECT scanner before clinical introduction. Several authors have reported the compromise accuracy of the iodine quantification, and these studies utilized iodine-to-saline dilutions to simulate various iodine concentrations 8, 9, 10. However, with the use of the dilutions, it is difficult to assess the comprehensive accuracy of quantitative measurements, which fluctuates over periods of time ranging from hours to weeks. Moreover, to our knowledge, studies evaluating the accuracy of the quantitative measurement of the calcium concentration have not appeared. A newly developed multienergy CT phantom (Gammex RMI 1472; Gammex RMI, Middleton, WI) enables robust evaluation of DECT scanner performance, such as material discrimination, quantitative accuracy of multienergy scans, consistency, and stability. The phantom equips various reference materials with their theoretical values of Hounsfield unit and iodine and calcium concentrations.

Here, we represented the comprehensive accuracy of quantitative measurements of the CT number, and iodine and calcium concentrations with various scanning settings for reference materials of a multienergy CT phantom.

Section snippets

Phantoms

Because this was a phantom study, ethical approval was not required. Figure 1a shows a multienergy CT phantom, and the phantom is 16.5 cm deep and consists of inner (20 cm in diameter) and outer (40 cm in width and 30 cm in height) sections composed of water-equivalent material to simulate small and large patient sizes. The phantom contains cylindrical holes (2.85-cm diameter) for the placement of various rods of the reference materials. Table 1 shows the specification provided by the

Results

Figure 2 shows the mean difference between the measured (1) Hounsfield unit value, (2) iodine concentration, and (3) calcium concentration from theoretical values for each reference material. Underestimation of Hounsfield unit (>50 HU) was observed for high-density rods (iodine 15 mg/mL and calcium 300 mg/mL) and a small physical object (iodine 5 mg/mL with a 2-mm core) at the energy level of 50 keV (Fig 2a). Overall, differences from theoretical values at 50 and 70 keV were decreased in the

Discussion

Using a multienergy CT phantom, our study demonstrated the comprehensive performance of a single-source DECT scanner with a fast kilovoltage switching technique in quantitative measurements. The accuracy of Hounsfield unit measurements for various materials (Gammex RMI 467, Gammex RMI) using a DECT scanner (750 HD, GE Medical Systems) was reported by Goodsitt et al. (15), who found that Hounsfield unit values on monochromatic images could be very inaccurate, particularly for materials mimicking

References (24)

  • LiY. et al.

    Iodine quantification with dual-energy CT: phantom study and preliminary experience with VX2 residual tumour in rabbits after radiofrequency ablation

    Br J Radiol

    (2013)
  • ZhangD. et al.

    Objective characterization of GE Discovery CT750 HD scanner: gemstone spectral imaging mode

    Med Phys

    (2011)
  • Cited by (17)

    • Improvement of image quality for pancreatic cancer using deep learning-generated virtual monochromatic images: Comparison with single-energy computed tomography

      2021, Physica Medica
      Citation Excerpt :

      The relationship between the CT number (in Hounsfield unit, HU) of an anatomical structure in the VMI and the corresponding photon energy level (in electron-volts) follows various trends. For example, the HU value of bone increases and that of fat decreases as the energy level of the VMI decreases [15]. Considering these relationships, we hypothesized that VMIs can be generated using deep learning from conventional SECT images to improve quality and to support radiotherapy treatment planning.

    • A novel fast kilovoltage switching dual-energy CT with deep learning: Accuracy of CT number on virtual monochromatic imaging and iodine quantification

      2021, Physica Medica
      Citation Excerpt :

      VMI at the energy levels of 50, 70, and 100 keV and iodine density image were reconstructed on the basis of the low- and high-kVp projection data. With reference to the report by Ohira et al. [11], we performed a VMI analysis at 50, 70, and 100 keV because the usefulness of VMI at these energy levels has been reported in clinical practice: at 50 keV, the iodine dose could be reduced by 50% for CT aortography [14]; at 70 keV, the image is statistically equivalent to a 120 kVp single-energy CT [15]; and at 100 keV, pseudoenhancement never occurred in renal cysts [16]. Pseudoenhancement means an increase in the CT value on contrast-enhanced CT for renal cysts that generally cannot be contrast-enhanced.

    • Structure delineation in the presence of metal – A comparative phantom study using single and dual-energy computed tomography with and without metal artefact reduction

      2019, Physics and Imaging in Radiation Oncology
      Citation Excerpt :

      FKS-DECT in combination with MAR have been reported to improve image quality and delineation in patients [12] and tissue quantification in phantoms [13] by reducing artefacts from dental metal implants. However, additional artefacts can be introduced by the MAR algorithm itself [7,12,13]. A recent study by Kovacs et al. [14] reported that DECT with sequential scans and MAR improved the delineations accuracy in both phantoms and patients, compared to SECT.

    View all citing articles on Scopus

    Funding: This study was supported by Health and Labour Sciences Research Grants for Promotion of Cancer Control Programs (H26-Cancer Policy-General-014), JSPS KAKENHI Grant (Grant-in-Aid for Scientific Research [B] 15H04913 and Grant-in-Aid for Young Scientists [B] 17K15816), and Osaka Cancer Society.

    View full text