CARE Dose 4D combined with sinogram‐affirmed iterative reconstruction improved the image quality and reduced the radiation dose in low dose CT of the small intestine

Abstract Objective Multislice computed tomography (MSCT) has been used for diagnosis of small intestinal diseases. However, the radiation dose is a big problem. This study was to investigate whether CARE Dose 4D combined with sinogram‐affirmed iterative reconstruction (SAFIRE) can provide better image quality at a lower dose for imaging small intestinal diseases compared to MSCT. Methods The noise reduction ability of SAFIRE was assessed by scanning the plain water mold using SOMATOM Definition Flash double‐source spiral CT. CT images at each stage of radiography for 239 patients were obtained. The patients were divided into groups A and B were based on different tube voltage and current or the image recombination methods. The images were restructured using with filtered back projection (FBP) and SAFIRE (S1–S5). The contrast noise ratio (CNR), CT Dose index (CTDI), subjective scoring, and objective scoring were compared to obtain the best image and reformation parameters at different stages of CT. Results Twenty‐six restructuring patterns of tube voltage and current were obtained by FBP and SAFIRE. The average radiation dose using CARE Dose 4D combined with SAFIRE (S4–S5) reduced approximately 74.85% compared to conditions where the tube voltage of 100 kV and tube current of 131 mAs for patients with MSCT small intestinal CT enterography at plain CT scan, arterial stage, small intestine, and portal venous phase. The objective and subjective scoring were all significantly different among groups A and B at each stage. Conclusions Combination of CARE Dose 4D and SAFIRE is shown to decrease the radiation dose while maintaining image quality.

enterography with FBP and SAFIRE was more accurate in the diagnosis of advanced inflammatory ileal disease. However, the image quality and inferior visibility of small anatomic structures are still unsatisfactory. CARE Dose 4D is a four-dimensional automatic realtime dose adjustment technology which is put forward by Siemens.
The technology can determine the size of patients according to the positioning image, and automatically calculate the required tube current value and CT Dose index (CTDI) at different tube voltages, which will obtain the constant diagnostic image quality of all parts of the body with the lowest radiation Dose. Furthermore, CARE Dose 4D was successful to optimize radiation dose in previous studies. 22, 23 We planned to demonstrate CARE Dose 4D in the application of imaging small intestinal diseases.
In this study, we assessed the effect of SAFIRE on noise reduction with the ordinary water film and selected the adequate tube voltage and current. We also reported on the effects of the combination of CARE Dose 4D (Siemens Healthcare) and SAFIRE on the image quality and radiation dose reduction for small intestinal CT.     Fifty square centimeters of region of interest (ROI) was selected. The average CT number and standard deviation within the ROI were measured by one physician twice. The formulas were as follows: wherein, CT is the CT value, and SD is the standard deviation.

2.B.2 | Small intestine CT angiography
Small intestine filling method Each patient took 1200 ml 2.5% isotonic mannitol orally. Patients with poor tolerance and significant obstruction took iso-osmolar contrast medium orally with low intensity and high frequency. All patients with intestinal obstruction were treated with gastrointestinal decompression after examination.  Table 1.

Measurement parameters and methods
The scan parameters were the same as those of the control group  Table 2.  With respect to CNR and that of A7, A8, and A9 were obviously higher than the other groups via plain scan (P = 0.000). At the arterial period, the CNR of A1, A2, A5, and A6 was all higher than that of the other groups (P = 0.000-0.049). The CNR of A2 was the highest both at small intestine period and portal venous phase (P = 0.000; P = 0.000-0.003). However, the CNR of A4, A5, A7, A8, and A9 was higher at the small intestine period, while that of A2-A9 at the portal venous phase was lower than that of A1 (P = 0.177-0.826; P = 0.000-0.015).
The CTDI vol and SSDE of A7-A9 were all lower than those in other groups at the period of plain CT scan and arterial (P = 0.000).
There was no significant difference among three groups except in the plain CT scan (P = 0.297-0.461). At the small intestine period and the portal venous phase, the CTDI vol of A8 was both lower than the other groups (P = 0.000-0.005, P = 0.000) [ Fig. 4(a)]. The SSDE of A7, A9, and A1 was closely and lower than the other groups (P = 0.000-0.021) at portal venous phase.
In group B, both of CCR and CNR raised from B0 to B5 in turn.
Those of B5 was higher than the other four groups (0.000 < P < 0.05) [ Fig. 4(b)]. As the radiation dose did not exchange with the recombination mode in a single scan, the CTDI vol and SSDE were not analyzed.

3.B.2 | Subjective assessment of statistical differences between the groups based on group A and B
As respect to subjective assessment, significant difference was observed among group A, but there was no obvious difference among group B (Tables 4-7). The vascular score of A1 was higher than the other groups at each stage (P = 0.000). Meanwhile, there was no significant difference among group A3-A7 (P > 0.05), and they were both obviously higher than that of group A8 and A9 (P = 0.000). The subjective scoring of intestinal wall in group A1, A2, A4, A5 and A6 were not markedly different (P > 0.05), and higher than the other groups (P < 0.05) at plain CT scan and arterial period. Group A1, A4 and A6 at small intestine period and group A1, A2, A4 and A6 at portal venous phase were higher than the other groups (P < 0.05) and there was significant difference among these groups (P < 0.05). The presentation of lesions manifested that the score of group A1 was significantly higher than other groups at every stage (P = 0.000). Besides, that of group A2, A4, A5, A6 and A7 were all higher than that of group A3, A8 and A9 (P = 0.000) at plain CT scan while group A2-A7 was higher than group A8 and A9 at the other three stages (P = 0.000). However, there was no significant difference among group A2, A4, A5, A6 (P = 0.351-1.000) and A7 as well as A2-A7 at the other three stages (P > 0.05).
The subjective scoring of blood vessel were all higher in group B1-B5 than that in B0 (P < 0.05), and there no significant difference among group B1-B5 at each stage (P > 0.05). The intestinal wall and lesions scoring of group B1-B5 were both higher than that of B0 at plain CT scan and portal venous phase (P < 0.05). At arterial period, the intestinal wall of B2-B5 was higher than that of B0 and B1 while the lesions scoring of B1-B5 was higher than that of B0 (P < 0.05). However, the intestinal wall of B0-B3 was lower than that of B4 and B5 while the lesions scoring of B1-B5 was lower than that of B0 (P < 0.05).
All above, for the patients with standard weight (18.5 kg/m 2 ≤ BMI ≤ 23.9 kg/m 2 ), the better image quality and lower radiation dose could be obtained under the tube voltage of 100 kV and tube current of 131mAs using CAER Dose 4D and SAFIRE 4 or SAFIRE 5 compared to controls (Table 8).

| DISCUSSION
There are many factors that influence the radiation dose of CT including scan parameters, hardware and software. The study of reducing CT radiation dose is started from 1980s and the focus is fixed on optimizing scan parameters. The square of the tube voltage is proportional to the radiation dose of x ray and the decrease of it can influence the image noise and spatial resolution. 28   was considered as the best choice which could help obtain good image quality and A7 received the lowest radiation dose. Above all, A7 combined with SAFIRE 4 or SAFIRE 5 might be the best choice at this stage. When it was at the small intestine period, the CCR of A7 was highest, and the CNR of A6 was higher than that of A7 and A8, as well as the CTDI vol and SSDE of A8 were lower than other groups. The CCR and CNR of B5 were highest among the group B. The scoring of intestinal wall and the lesion of A8 were all lower than A6 and A7, while that of the blood vessel was slightly lower than that of A6.
Therefore, A7 combined with SAFIRE 4 or SAFIRE 5 was recommended as the method to obtain good image quality of small intestine.
The CCR of A7, A8, and A9 was all higher than the other groups at portal venous phase, the CNR of A2, A3, A6, and A7 was successively decreased. The CTDI vol and SSDE of A8 were the lowest. The CCR and CNR of B4 and B5 were higher than those in the other groups.
The subjective scoring using SARIRE were all higher than using FBP.
As a result, the combination of A7 and SAFIRE 4 or SAFIRE 5 was the best condition for the scan. The result was consistent with the study of Andrew et al. 18 However, the study of Andrew et al. has not compared the interaction of the scan parameter and iterative intensity.
Our study used lower scan conditions and higher iterative intensity to ensure the image quality. Besides, the CTDI vol had decreased by 74.85%. Nevertheless, this study still has some limitations. The subjective scoring replaced the objective measurement which might bring some deviation.

| CONCLUSION
In conclusion, the tube voltage of 100 kV, tube current of 131 mAs, and CARE Dose 4D were appropriate for people with a standard weight using a regular CT scan, arterial phase, small intestine period, and a portal venous phase when the MSCT was conducted. The

CONFLI CT OF INTEREST
The authors declare that they have no conflict of interest.