Relationship between size‐specific dose estimates and image quality in computed tomography depending on patient size

Abstract This study investigates the relationship between contrast‐to‐noise ratio (CNR) and size‐specific dose estimate (SSDE) in computed tomography (CT) depending on patient size. In addition, the relationship to the auto exposure control (AEC) techniques is examined. A tissue‐equivalent material having human‐liver energy dependence is developed and used to evaluate these relationships. Three exposure dose levels (constant CT dose index, constant SSDE, and with AEC) are tested using four different phantom sizes (diameter: 15, 20, 25 and 30 cm) in two different CT scanners (SOMATOM Definition Flash, Siemens, and LightSpeed VCT, GE). The contrast‐to‐noise ratios (CNRs) are measured using the developed phantom. It is found that the CNR increases with decreasing phantom size at constant SSDE, although the increase ratio is smaller than that of the constant CT dose index. This result indicates that the image characteristics differ even when the patient dose received from the CT examination is equivalent for each patient size. In the case of AEC use, the CNR results of the Siemens scanner exhibit a similar trend to those obtained for constant SSDE, for each phantom size. This suggests that the AEC technique that maintains a constant image quality (CARE Dose 4D) for each patient size corresponds well to the image quality obtained for constant SSDE. These findings facilitate further understanding of the relationship between image quality and exposure CT dose depending on patient size.

CTDI vol does not reflect patient size and is a metric of radiation output obtained using a reference phantom rather than patient dose. To obtain the patient dose, the size-specific dose estimate (SSDE) has been proposed by American Association of Physicists in Medicine (AAPM) Task Groups 204 and 220. 11,12 In this approach, the patient dose can be estimated by multiplying CTDI vol by conversion coefficients determined with consideration of the patient size.
The image quality also varies in relation to patient size; however, many of the reports that have evaluated this relationship have considered image noise only as the image quality indicator, and utilized a uniform phantom. 5,7,13,14 This approach was adopted in those studies because the primary goal of CT examination is soft-tissue evaluation. 15 The energy dependence of soft tissue, that is, the contrast variation, is extremely small compared with those of contrast media and bone. In addition, no phantom having human-tissueequivalent energy dependence was available to those researchers.
Evaluations based on noise ignore the object contrast; thus, this approach is less relevant to image quality, because the contrast affects the lesion detectability. 10 The soft-tissue contrast and image noise change slightly with patient size. Therefore, considering its relation to exposure dose, the contrast-to-noise ratio (CNR) is an optimal index for image quality.
With regard to related research considering patient size, image quality, and CT dose, Boone et al. 1 have examined this topic in detail. In the literature, dose reduction protocols to achieve equivalent image quality based on patient size have been proposed. However, to the best of our knowledge, few studies have evaluated the image quality obtained under the equivalent patient dose received during CT examination. Further, as a CT technology related to patient dose and size, auto exposure control (AEC) is one of the most important techniques. AEC strategies are broadly divided into two categories regardless of attenuation level: "constant image noise (noise index-based technique)" and "constant image quality (reference mAs-based technique)." [2][3][4][5][6][7][8][9]16,17 Although these behaviors differ with respect to patient size, the relationship between the SSDE and image quality has not been adequately evaluated.
In this study, we develop a tissue-equivalent material having the same energy dependence as the human liver and investigate the relationship between the CNR and SSDE depending on patient size.
In addition, we examine the relationship between those parameters and the image quality given by AEC.

2.A | Tissue-equivalent material
A tissue-equivalent material having the same energy dependence as the human liver was jointly developed with Kyoto Kagaku Corporation (Kyoto, Japan). The inside of a 20-cm acrylic cylindrical case was filled with water and the phantom was set in place. Here, both the developed phantom and an existing phantom (SZ tissue-equivalent phantom, Kyoto Kagaku) were used for comparison. The CT numbers were measured at various energies to validate the accuracy using virtual monoenergetic images. Dual-energy CT (DECT) scans were performed using a dual source CT scanner (SOMATOM Definition Flash; Siemens Healthcare, Erlangen, Germany). The tube voltages were set to 100 and 140 kVp with a tin (Sn) filter. The DECT raw data were then transferred to a workstation (Syngo Multimodality Workplace, Siemens Healthcare), and the monoenergetic images were reconstructed at 10-keV intervals from 40 to 160 keV using a dedicated application (Monoenergetic; Siemens Healthcare). In addition, the liver CT numbers of 14 patients were measured retrospectively based on clinical data obtained through dual-energy scanning of the abdomen. The use of patient data was approved by the ethics committee of our institution.
Hence, the developed phantom was found to have almost equivalent CT numbers to the patient livers, although the CT numbers for the patient livers included individual differences (Fig. 1). In contrast, the CT numbers of the existing phantom decreased at lower energy; this phantom did not have sufficient energy dependence. Therefore, we decided to use the developed phantom as an object in this experiment.

2.B | Phantom and scan protocols
We employed four differently sized cylindrical cases having diame-

| 247
Healthcare) and a 64-section CT (LightSpeed VCT; GE Healthcare, Milwaukee, WI, USA). The image protocol was that used for standard abdominal examination in our hospital. The scan parameter details are listed in Table 1.
Three dose levels were tested for each scanner to assess mutual relations. First, the CTDI vol value displayed on the console was fixed at 20 mGy regardless of phantom size (constant CTDI). This CTDI vol value was chosen based on the Japanese DRLs 2015 for an abdominal/pelvic CT examination. 18 Second, the SSDE was fixed at a constant level. The SSDE is defined as follows: where f is the conversion factor obtained as a function of the patient's effective diameter. The f values for the SSDE are listed in

2.C | Evaluation approach
At each exposure dose level, the CNR was measured for each phan- Dose. [19][20][21][22][23][24] The FOM is an indicator used to evaluate image quality, which is normalized for the exposure dose and indicates the dose efficiency. On that basis, the dose ratio required to obtain an FOM comparable to that of the 30-cm-diameter phantom was calculated using the CNR as an image quality indicator. The SD was also used as an indicator, ignoring contrast, according to the relation FOM = (1/SD 2 )/Dose. In this study, to demonstrate the effect of this choice, the CNR and SD were both taken as indicators of image quality, and the dose ratio required to obtain the same image quality as that for the 30-cm phantom was examined.  (Fig. 3).

| DISCUSSION
The developed tissue-equivalent material exhibited energy-dependent CT numbers similar to those measured for the liver from patient data. Thus, sufficient accuracy was achieved for the phantom to be used as a tissue contrast object for CT examination. We used this material to evaluate the relationship between exposure dose and image quality.  Fig. 4(a)]. This depends on the effective energy of the x-ray beam in each scanner.
We measured the effective energy in a preliminary study, obtaining 55.3 and 59.0 keV for the GE and Siemens scanners, respectively.
Therefore, the contrast was slightly changed by the effect of beam hardening due to different phantom diameters. In particular, the GE scanner with low effective energy exhibited a large variation.
Regarding the relationship between the AEC and image quality, the image noise has been adopted as an indicator of image quality in many studies. 5,7,13,14 Although the image noise (i.e., SD) can be used to evaluate the variation of the tube current modulation, image quality evaluation  Our study has some limitations. First, cylindrical phantoms only were used in the experiments. The human body has an elliptical shape in many cases, and the body shape affects the AEC. Many studies have evaluated the relationship between AEC and patient size by assembling realistic human shapes from those of small to large patients. 1,7,19 The aim of the present study was to investigate several aspects of the relationship between SSDE and image quality; thus, the effect of body shape was not considered. Second, although the AEC incorporates various user-controlled parameters such as the noise index, reference mAs, and the strength of the CARE Dose 4D, we did not assess the effects of parameter adjustment. Finally, in clinical CT, low tube voltages, such as 80 and 100 kVp, are selected depending on the aim and patient size.
In future research, the effect of the tube voltage on the relationship between the SSDE and image quality should be investigated.

| CONCLUSION
We developed a tissue-equivalent material having human-liver energy dependence, which was used to evaluate the relationship between image quality and dose depending on patient size. The results of this study reveal that the image characteristics differ even when the patient dose received from the CT examination is equivalent for each patient size (constant SSDE). In addition, the AEC that retains constant image quality (CARE Dose 4D), rather than constant noise level (Auto mA), for each patient size corresponds well to the image quality obtained for constant SSDE. These findings facilitate further understanding of the relationship between image quality and exposure CT dose depending on patient size.

CONF LICT OF I NTEREST
The authors declare no conflict of interest.