Dual-source abdominopelvic computed tomography: Comparison of image quality and radiation dose of 80 kVp and 80/150 kVp with tin filter

Objective To compare the radiation dose and the objective and subjective image quality of 80 kVp and 80/150 kVp with tin filter (80/Sn150 kVp) computed tomography (CT) in oncology patients. Methods One-hundred-and-forty-five consecutive oncology patients who underwent third-generation dual-source dual-energy CT of the abdomen for evaluation of malignant visceral, peritoneal, extraperitoneal, and bone tumor were retrospectively recruited. Two radiologists independently reviewed each observation in 80 kVp CT and 80/Sn150 kVp CT. Modified line-density profile of the tumor and contrast-to-noise ratio (CNR) were measured. Diagnostic confidence, lesion conspicuity, and subjective image quality were calculated and compared between image sets. The effective dose and size-specific dose estimate (SSDE) were calculated in the image sets. Results Modified line-density profile analysis revealed higher attenuation differences between the tumor and normal tissue in 80 kVp CT than in 80/Sn150 kVp CT (127 vs. 107, P = 0.05). The 80 kVp CT showed increased CNR in the liver (8.0 vs. 7.6) and the aorta (18.9 vs. 16.3) than the 80/Sn150 kVp CT. The 80 kVp CT yielded higher enhancement of organs (4.9 ± 0.2 vs. 4.7 ± 0.4, P<0.001) and lesion conspicuity (4.9 ± 0.3 vs. 4.8 ± 0.5, P = 0.035) than the 80/Sn150 kVp CT; overall image quality and confidence index were comparable. The effective dose was reduced by 45.2% with 80 kVp CT (2.3 mSv ± 0.9) compared to 80/Sn150 kVp CT (4.1 mSv ± 1.5). The SSDE was 7.4 ± 3.8 mGy on 80/Sn150 kVp CT and 4.1 ± 2.2 mGy on 80 kVp CT. Conclusions The 80 kVp CT reduced the radiation dose by 45.2% in oncology patients while showing comparable or superior image quality to that of 80/Sn150 kVp CT for abdominal tumor evaluation.


Introduction
Optimization of the radiation dose delivered in abdominopelvic computed tomography (CT) imaging is important, especially in situations where repeated CT examinations are performed in patients. Patients with malignancy undergo repeated CT examinations for evaluation of treatment response and surveillance after treatment. Various strategies have been developed to further reduce the radiation dose, including low kVp, automated exposure control, and iterative reconstruction (IR) [1,2]. Automatic tube current modulation (ATCM) can allow a reduction in the radiation exposure during CT examination, which is also affected by the patient's size [3,4]. Currently, 80 kVp is feasible and increasingly used as the optimal radiation dose [5][6][7][8]; however, although using a low kVp reduces the radiation dose, it increases the image noise [9]. IR selectively reduces statistical noise in the images, thereby improving the quality of subtle image details and may facilitate dose reduction. In the past decade, the evolution from statistical-based to model-based iterative algorithms has improved the performance of IR algorithms [10]. Advanced modelled iterative reconstruction (ADMIRE; Siemens Healthcare, Forchheim, Germany) is a model-based IR that can reduce noise in raw data and may allow further dose reduction while generating images of acceptable quality. ADMIRE is based on the results of the pseudo raw data, in comparison to the measured data, which is subtracted and reinserted into the loop afterward [11].
Dual-source dual-energy CT (DECT) systems enable dual-energy data to be acquired using two X-ray sources at different energy levels, i.e., a variety of voltage and tube current combinations [12][13][14][15]. Dual-source DECT images use a blend of low-and high-kVp images to provide an image impression similar to a standard 120 kV image (i.e., 80/140, 80/Sn150, or 90/Sn150 kVp) to generate virtual noncontrast images [14]. In other words, low kVp/high kVp with tin filter (Sn) CT generates low kVp, virtual noncontrast, and blended images, which are widely used in oncology [14]. In addition, virtual monochromatic images can be used to magnify the iodine contrast-to-noise ratio (CNR) to enhance the differentiation between tumor and background using DECT in oncology [15,16]. Although initial DECT revealed a three times higher radiation dose than single-energy CT [17], recent studies have demonstrated that third-generation dual-source DECTs can be performed without a radiation dose penalty or impairment of image quality compared to single-energy CT with 100-120 kVp [18,19]. They suggested that DECT can be routinely used in patients as there is no increase in the radiation dose compared to single-energy CT. However, no comparison has been reported between images with low kVp from single-energy CT and blending images from DECT. If the performance of an 80 kVp CT scan is similar to that of a blending image (80/Sn150 kVp) of DECT while reducing the radiation dose, it may reduce the use of DECT in patients who receive repetitive scans. To our knowledge, this imaging scheme concept has not been explored previously.
Thus, the purpose of our study was to compare the image quality and radiation dose between 80 kVp and 80/Sn150 kVp CT scans performed by the third-generation dual-source DECT using the ADMIRE reconstruction algorithm.

Study design
This study was a retrospective analysis of CT images and was approved by our Institutional Review Board of Gil medical center with a waiver of the requirement for patient consent.

Study patients
We screened 191 consecutive patients who were referred for CT examinations of the chest and abdominopelvic region between August 2018 and March 2019 for assessment of response to chemotherapy, or surveillance of treated malignancy. Abdominopelvic CT was selected for this study. The inclusion criteria were as follows: (a) solid nodules of malignant liver tumor, other malignant visceral tumors (except liver tumors), metastatic lymph nodes (larger than 1.5 cm along the short axis), peritoneal tumors, extraperitoneal tumors (i.e., retroperitoneum, muscle, subcutaneous fat layer), and metastatic bone tumors, (b) nodules � 5 in number, and (c) nodules not typically hemangiomas or cysts. Six predetermined abdominopelvic lesions were modified from a previously described report [20]. We limited the number of nodules to minimize a cluster bias. We excluded 46 patients without a reference standard (MRI, PET/CT, or surgery), without any lesion on CT, or because of a change in protocol. Thus, 259 nodules in 145 consecutive patients were included (Fig 1). Nodules were selected and annotated by an experienced study coordinator (BLINDED-FOR-PEER-REVIEW), who was not involved in the image analysis.

CT examination protocol
CT examinations were performed using a 192-slice, third-generation dual-source CT scanner (SOMATOM Force, Siemens Healthcare, Forchheim, Germany) in the dual-energy mode with tube detectors A (80 kVp; reference tube current, 250 mAs) and B (Sn 150 kVp; reference tube current, 125 mAs), using tube current dose modulation (CARE dose 4D; Siemens Healthcare) ( Table 1).

Data reconstruction
Images from 80/Sn150 kVp CT (data from tube detectors A and B) and 80 kVp CT (data from tube detector A) scans were made with an axial reconstruction thickness of 3 mm using the ADMIRE algorithm at a strength level of 2 out of 5, each with an edge-enhancing convolution kernel (Br64). The blended images were reconstructed by applying a blending factor of 0.6 (M_0.6; 60% of the 80 kVp and 40% of Sn 150 kVp spectrum). The reconstruction was achieved on a multimodality workstation (syngo.via VB20, Siemens Healthcare).
image quality for diagnostic purposes was graded using a 5-point scale (1, nondiagnostic image quality, strong artefacts; 2, severe artefacts with uncertainty about the evaluation; 3, moderate artefacts with mild restricted assessment; 4, slight artefacts with unrestricted diagnostic image evaluation possible; and 5, excellent image quality with no artefacts).

PLOS ONE
Dual-source APCT: 80kVp and 80/Sn150 kVp Enhancement of organs was graded using a 5-point scale (1, very poor; 2, suboptimal; 3, acceptable; 4, above average; and 5, excellent). Image noise was graded using a 5-point scale (1, unacceptably high; 2, higher than average; 3, average; 4, less than average; and 5, minimum noise). A score was derived for axial images with a soft-tissue window setting (width, 400 HU; level, 50 HU) on a PACS system. Reviewers were allowed to change the window level and width as per their comfort level during analysis. For each presented acquisition, both readers had to independently assess the predetermined abdominopelvic lesions. Each item was graded using a five-point confidence index (1, very poor; 2, poor; 3, average; 4, high; and 5, excellent confidence) and an ordinal scale for lesion conspicuity (1, definite artefact [pseudolesion]; 2, probable artefact; 3, subtle lesion; 4, poorly visualized margins; and 5, well-visualized margins). A score of either 1 or 2 was considered unacceptable for diagnostic purposes.
Objective image analysis. Image noise was measured by drawing a circular region of interest (ROI; size between 1 and 3 cm 2 ), as standard deviations (SDs) in Hounsfield units (HU), on the axial images of two image sets in the following three anatomical regions by one blinded reader (BLINDED-FOR-PEER-REVIEW) [21][22][23]: the subcutaneous fat in the anterior abdominal wall, the right hepatic lobe parenchyma, and the lumen of abdominal aorta. Mean attenuation values were measured (in HU) for each ROI in all image sets. Signal-tonoise ratio (SNR = mean HU of interested area/fat noise) and CNR ([mean HU of interested area-mean HU of psoas muscle]/ fat noise) were calculated [9,20]. Fat noise was calculated as the mean of standard deviation of CT attenuation in ROIs placed within the mesenteric fat. ROIs were carefully placed at the same location between different image series.
The two image sets were used for modified line-density analysis [24]. The tumor was identified on 80/Sn150 images, and one reader, with 4 years of experience, positioned a line of 10 mm in length and 2 mm in width perpendicular to the tumor margins with one half of the line within the tumor and the other half within the normal tissue. Mean, minimum, and maximum HU values were measured within this 10-mm line using a multimodality workstation. Modified line-density analysis was performed on 128 nodules from 92 patients, as not available in patients with peritoneal seeding lesions.

Radiation dose
The dose-length product (DLP) and volume CT Dose Index (CTDI vol ) were recorded from the scanner dose page. The effective dose (in millisieverts, mSv) was calculated using the tissueweighting factor (men, k = 0.012; women, k = 0.017) based on the International Commission on Radiological Protection Publication 103 with modifications [25,26] as follows: Size-specific dose estimates (SSDEs) were calculated as follows: A conversion factor based on the effective diameter and a 32-cm-diameter polymethylmethacrylate phantom was used [27]. The effective diameter was calculated from the anteroposterior (AP) and lateral dimensions on the CT scan as follows: effective diameter ¼ ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi AP � lateral dimention p ½27�

Statistical analysis
Dose parameters and image analysis were compared between different dose CT scans using Student's t-test and chi-squared test. The interobserver agreement between the two readers for each of the evaluated subjective image quality parameter was estimated using the overall proportion of agreement [28], rather than κ statistics, as the latter is affected by the distribution of data across categories and was thus not a valid indicator of our data [29]. A P-value <0.05 was considered statistically significant. All statistical analyses were performed using SPSS Statistics for Windows, Version 21.0 (IBM Corp.).

Patient characteristics
Of the 145 patients in our study, 79 (54.5%) were men and 66 (45.5%) were women with a mean age ± standard deviation of 62.  Table 2).

Objective image quality
The 80 kVp CT had a significantly higher attenuation compared to 80/Sn150 kVp CT in the liver and aorta (125.9 HU vs. 107.2 HU, P<0.001; 208.2 HU vs. 160.5 HU, P<0.001; Table 3 and Fig 2). The CNR of the aorta in 80 kVp CT images was significantly higher than in 80/ Sn150 kVp CT images (P<0.001) whereas the CNR of the liver in 80 kVp CT images was mildly higher than that in 80/Sn150 kVp CT images (P = 0.197). The SNR of the liver and aorta tended to be higher in 80/Sn150 kVp CT images, although there was no significant difference between the image sets (P = 0.075 and 0.069, respectively). The mean differences between maximum and minimum attenuation within the tumor are shown in Table 4. The mean attenuation differences within the tumor in 80 kVp CT were significantly higher than those in 80/Sn150 kVp CT (127.2 HU vs. 107.0 HU, P = 0.05).

Radiation dose
Various dose parameters of the two image sets are shown in Table 6 and Fig 5. For the 80/ Sn150 kVp CT, the mean CTDI vol was 5.3 ± 1.4 mGy, and the SSDE was 7.4 ± 3.8 mGy, with an effective dose of 4.1 ± 1.5 mSv. For the 80 kVp CT, the mean CTDI vol was 2.9 ± 1.0 mGy, and the SSDE was 4.1 ± 2.2 mGy, with an effective dose of 2.3 ± 0.9 mSv. On average, the effective dose of 80 kVp CT scan was 45.2% ± 2.8% (range, 38.0-51.9%) less than that of 80/ Sn150 kVp CT.

Discussion
We investigated whether the performance of 80 kVp CT was comparable to that of the 80/ Sn150 kVp CT while reducing the radiation dose, to ascertain if it could be used in oncology patients who receive repetitive DECT scans. Our findings revealed that there was a 45.2% reduction in radiation dose by using the low tube voltage (80 kVp) CT imaging with ADMIRE while maintaining excellent image quality compared with DECT (80/Sn150 kVp). Contrastrelated features including enhancement of organs, CNR, and attenuation differences within the tumor border were higher in 80 kVp CT than in 80/Sn150 kVp CT, while SNR showed no significant differences. A potential explanation for these findings is that lowering the CT tube voltage results in a greater photoelectric effect of iodinated contrast media, with only a slight

PLOS ONE
Dual-source APCT: 80kVp and 80/Sn150 kVp increase in the noise because of improved tube current capabilities. Moreover, the advances in hardware equipped with Stellar detectors and software program in third-generation dualsource CT [30,31], which are supposed to be more sensitive to electron influx, are thus, doseefficient and generate high quality images while using 80 kVp. A few studies that investigated DECT, showed that the radiation dose was comparable to the single energy 100-120 kVp CT [18,32]. However, our study found that the 80 kVp CT showed superior image quality with a lower radiation dose than DECT in abdominal scans. Thus, DECT should not be routinely used in patients who may require repeated abdominopelvic CT scans. Several studies have reported that 80 kVp abdominopelvic CT can be used effectively while reducing radiation dose in patients with a normal BMI [33][34][35][36]. Studies have also reported that reducing the tube voltage from 120 kVp to 80 kVp resulted in a 48-65% dose reduction using either an identical tube current or automated tube current modulation [37][38][39]. However, extremely low tube voltage scans may result in noisier images that are inadequate for interpretation. An increase in tube current or use of IR may improve image quality and lesion

PLOS ONE
Dual-source APCT: 80kVp and 80/Sn150 kVp conspicuity [30]. Our study showed that the effective dose of 80 kVp CT (2.3 mSv ± 0.9) scan was 45.2% less than that of 80/Sn150 kVp CT (4.1 mSv ± 1.5) using IR. Low tube-voltage CT scan is a robust method for radiation dose reduction in abdominopelvic CT. Third-generation dual-source CT is more dose-efficient than second-generation dualsource CT because of technical advances and adjusted scan protocols while the image quality is consistently high with all assessed protocols [36]. The low kVp is more likely selected while using automated tube voltage selection (ATVS), thereby reducing the radiation dose in thirdgeneration dual-source CT. Park et al. reported that third generation dual-source CT demonstrated an increase in the number of 90-100 kVp CTs, whereas the majority of the patients underwent 100-120 kVp using ATVM on second generation dual-source CT [36]. Winklehner et al. reported that using third-generation dual-source CT, compared to second-generation, resulted in a tube voltage decrease of at least 10 kV in most patients (75%, 100-90 kVp) who received body CT angiography examinations [39]. These results support the use of low kVp

PLOS ONE
Dual-source APCT: 80kVp and 80/Sn150 kVp abdominopelvic CT in patients for reducing the radiation dose as well as for maintaining the diagnostic performance employed in our study. The low tube-voltage CT technique improves tumor conspicuity and tumor-to-tissue CNR [5,6]. In our study, mean differences between the maximum and minimum attenuation within the tumor was 127.2 HU in 80 kVp CT and 107.0 HU in 80/Sn150 kVp CT. Our results also demonstrated that the 80 kVp CT scan might be preferred over 80/Sn150 kVp CT for detecting hypervascular tumors because of better scores in all contrast-related features, increased enhancement of organs, CNR, and attenuation differences within the tumor border. The lesion conspicuity was slightly higher in 80 kVp CT compared to 80/Sn150 kVp CT, which means that low-density lesion detection is superior in 80 kVp CT compared to that in 80/Sn150 kVp CT. However, SNR of the liver and the aorta showed no significant difference between the

PLOS ONE
Dual-source APCT: 80kVp and 80/Sn150 kVp image sets in our study. Our results showed a balance between image noise and low-density lesion detectability while using low tube-voltage CT to evaluate solid organs. DECT of the abdomen has the advantage of analyzing CT images in various ways, including monoenergetic image analysis, liver fat and iron quantification, urinary calculi characterization, and gallstone imaging [15,37,38]. Another advantage of DECT is the faster acquisition time to discern the hemodynamic features from a moving organ (i.e., cardiac study) [40,41]. Unlike previous studies [18,32] that suggested the routine use of DECT, our study found that the performance alone was not sufficient to justify routine use of DECT for oncology patients who frequently undergo CT examinations, sometimes every 1-2 months, because the radiation dose is higher with 80/Sn150 kVp CT compared to 80 kVp CT. Moreover, we reported superior tissue contrast features with 80 kVp CT.
Our study had some limitations. First, it was a retrospective study involving oncology patients with a tumor lesion on abdominopelvic CT, which may have introduced a selection bias. Second, our study had a small number of patients with moderate to severe obesity; thus, we could not perform a robust comparison between 80/Sn150 kVp and 80 kVp CT in patients with moderate to severe obesity. Third, 80/Sn150 kVp scan combines 80 kVp and Sn 150 kVp to generate the image; therefore, 80/Sn150 kVp was set to have an unconditional higher radiation dose than 80 kVp scan. However, our aim was to focus on evaluating image quality between a low kVp image using single-energy scan and a blended image generated using DECT scan. Rather than increasing the radiation exposure of the patient through two scans, we considered that an intraindividual comparison would be ethical. Finally, we only evaluated the performance of a single low kVp image and a blended image in a single scan. We did not

Conclusions
In conclusion, 80 kVp CT showed a decrease in radiation dose exposure while providing comparable or superior image quality to that of 80/Sn150 kVp CT in oncology patients.