Evaluation of BMI‐based tube voltage selection in CT colonography: A prospective comparison of low kV versus routine 120 kV protocol

Abstract Aim To explore the value of individualized kVp selection based on the patient's body mass index (BMI, kg/m2) in CT colonography (CTC). Materials and Methods Seventy‐eight patients underwent two CTC scans: conventional 120 kVp in supine position (Group A) with 30% Adaptive statistical iteration algorithm (ASIR–V) and BMI–based lower kV p in prone position (Group B): tube voltage was suggested by an experienced investigator according to the patient's body mass index (BMI; calculated as weight divided by height squared; kg/m (2)).70 kV for BMI < 23 kg/m2 (Group B1, n = 27), 80 kV for 23 ≤ BMI ≤ 25 kg/m2(Group B2, n = 21) and 100 kV for BMI > 25 kg/m2 (Group B3, n = 30). Group A, corresponding to the BMI value in Group B, was divided into A1, A2, and A3 subgroups for analysis. Groups B used ASIR‐V of different weights (30%–90% ASIR–V). The Hounsfield Unit (HU) and SD values of the muscles and the intestinal cavity air were measured, and the signal‐to‐noise ratio (SNR) and the contrast‐to‐noise ratio (CNR) of images were calculated. Imaging quality was evaluated by two reviewers and statistically compared. Results The 120 kV scans were preferred more than 50% of the time. All images had excellent quality with good consistency between reviewers (Kappa > 0.75, p < 0.05). The radiation dose was reduced in groups B1, B2 and B3 by 63.62%, 44.63%, and 32.14%, respectively, compared with group A (p < 0.05). The SNR and CNR values between group A1/A2/A3 and B1/B2/B3 + 60%ASIR‐V were not statistically significant (p < 0.05). There was no statistically significant difference between the subjective scores of group B combined with 60%ASIR‐V and group A (p > 0.05). Conclusion BMI‐based individualized kV CTC imaging significantly reduces overall radiation dose while providing an equal image quality with the conventional 120 kV.


INTRODUCTION
tolerate the procedure due to anxiety, fear, and discomfort. Although the overall complication rate is low, the potential risk of perforation and bleeding cannot be ignored during the colonoscopy examination. The literature showed that CT colonography (CTC) has progressively evolved into a validated assessment for colorectal diseases. 4 Its capacity of identifying CRC and large polyps in symptomatic and asymptomatic individuals is almost equivalent to a conventional colonoscopy. CTC is considered the best surgical positioning and navigation method because it is a safe and non-invasive examination that can detect lesions both inside and outside the intestine. [5][6][7] Moreover, CTC could display the whole colorectum from multiple directions and angles through two-dimensional (2D) and three-dimensional (3D) reconstructed images. 8,9 The main disadvantage of CTC is the potential damage of ionizing radiation, 10,11 CTC examination requires two scans in the supine and prone positions, with a more extensive scanning range from the apex of the diaphragm to the pubic joint. Therefore, keeping the dose as low as possible without significantly sacrificing image quality is strongly advised. There is a naturally high contrast between the air in the intestine and the soft tissues of the intestinal wall, allowing for low-dose CTC scanning. The effect of low tube voltage on CTC image quality has been evaluated. 12 In addition, Chang et al.'s research showed that 100 kV combined with FBP algorithm CTC imaging could reduce volume CT dose index (CTDIvol) by 20% and reduce dose-length product (DLP) by 16%. 13 Reduced tube voltage can increase image noise, leading to poor image quality that will affect the detection of small lesions. Therefore, this study adds the postweight ASIR-V algorithm to reduce the noise while reducing the radiation dose, thus addressing the image quality problem caused by the tube voltage reduction. Therefore, our study aimed to optimize the scanning voltages in CTC for all patients, and to evaluate the use of individualized kV (70 kV, 80 kV, 100 kV) combined with the optimal weight of ASIR-V protocol to reduce the patient's radiation dose, contributing to the establishment of BMI-based individualized CTC imaging protocols.

Participants
Patients who underwent CT Colonoscopy (CTC) examination from March 2021 to May 2022 were enrolled prospectively. The study protocol was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board. We also recommended the guideline to the China Clinical Trials Registration Platform, the WHO Clinical Trials Registration Institution, and the National Health Commission of the People's Republic of China. The exclusion criteria were set as follows: incomplete colonoscopy (n = 48), inadequate bowel preparation (n = 20), and inability to remove metal artifacts from the body trunk (n = 2). Each patient's demographic data, including age, gender, height, weight, and Body Mass Index (BMI), were collected at admission.

Preparation before scanning
All participants performed the standard bowel preparation, which comprised of three-meal low-residue diets before the day of CTC and the administration of laxatives which included magnesium sulfate, senna leaf, and ricinoleic acid before the examination. The smooth muscle of the abdomen was under relaxation with spasmolytics to compensate for the movements of the contents in the gastrointestinal tract. According to the ESGE guidelines, adequate bowel preparation should result in the absence of visible residual stool or fluid in the colon. Repeated colonoscopy on the following day (after further colon cleaning) was recommended in patients with insufficient bowel preparation. After completing all preparations, when patients were in the left lateral position with their hips and knees flexed to 90 degrees, a radiologist began insufflation of gas into the low rectum via a thin rectal catheter until the patient experienced mild discomfort. Considering comfort of patients and gas of completeness, the entire procedure was monitored via scout scan to ensure adequate colonic distension. The total volume of gas was approximately 1−2 L. 14 If the image quality was nonoptimal, additional gas was insufflated.

CT examination and data reconstruction
All patients were scanned twice in the prone and supine positions using a 256-row CT scanner (Revolution CT, GE Healthcare, USA). In the supine position, group A received a standard dose with 120 kV and 30% ASIR-V reconstruction. In the prone position, tube voltage was suggested by an experienced investigator according to the patient's body mass index (BMI; calculated as weight divided by height squared; kg/m (2)). In our study, participants were categorized in three groups: A1 and B1(BMI < 23), A2 and B2 (23 ≤ BMI ≤ 25), A3 and B3 (BMI > 25). In each subgroup, Groups A and B shared the same patient.70 kV was chosen for patients with a BMI under 23 kg/m 2 , 80 kV was chosen for those with a BMI between 23 and 25 kg/m 2 , and 100 kV was used for those with a BMI greater than 25 kg/m 2 , with reconstruction ranging from 30% to 90% ASIR-V. (10 percent interval). The remaining parameters remained constant ( done from the diaphragm to the pubic symphysis. CT volume dose index (CTDI vol ) and dose-length product (DLP) were recorded for each group. Effective radiation dose (in milliseiverts) was estimated by using the doselength product multiplied by a conversion coefficient for the abdomen (k = 0.015 mSv mGy −1 cm 1 ). 0.625 mm slice interval reconstructed images were sent to the AW 4.6 workstation. Radiologists reconstructed images using the CT Colon VCAR software, which included 2D axial images, multi-planar reconstruction (MPR) images, three-dimensional (3D) endoluminal images, and the colon's raysum. 15

Quantitative image analysis
On the axial image,the circular regions of interest (ROIs) were placed manually by the observer at three levels of uniform density on both the psoas major muscle and the air of the intestinal cavity in the same slicer, and HU values and SD values were measured and the averaged value of the three image levels were calculated keeping the ROI size around 100 mm 2 . The ROI was placed in the same position as the prone and supine datasets by using the "copy-paste"tool. The standard deviation of the air in the intestinal cavity was used as the background noise value. Finally, the signalto-noise ratio (SNR) and contrast-to-noise ratio (CNR) were calculated.

Qualitative image analysis
Two radiologists with over 10 years of diagnostic experience scored the CTC axial images and reconstructed images on a five-point scale using a double-blind method. The following criteria were demonstrated 16 : • 5 points: virtually no artifact • 4 points: low image noise and the presence of a few artifacts • 3 points: images with a moderate amount of noise and minor artifacts • 2 points: a high level of image noise and a moderate level of artifacts • 1 point: the highest level of noise plus the most obvious artifacts Different criteria were applied to 3D reconstruction images.
• 5 points: a very smooth bowel wall, visible mucosal folds, sharp edges, and a clear lesion morphology. • 4 points: the bowel wall is less smooth, the mucosal folds are less clear, the sharp edges are lost, and the lesion morphology is clear. • 3 points: a slightly rough bowel wall and imprecise mucosal folds with blurred edges. The morphology and structure of the lesion are visible, which does not affect diagnosis. • 2 points: the diagnosis is hampered by the rough intestinal wall, blurry mucosal folds and edges, and unclear lesion morphology. • 1 point: an extremely rough intestinal wall with imprecise mucosal folds and indistinguishable edges, as well as an indistinct lesion morphological structure.
In both evaluations, images with scores less than 3 were considered non-diagnostic.

Subjective analysis
Subsequently, paired images were presented to the radiologists using either 120 kV combined with 30% ACIR-V or with low kV and varying weights of ASIR-V. Each radiologist was asked to designate for each pair, viewed side by side: • the consistency of the lesions'location in CTC with the colonoscopy results,which were used as the standard; • which weight of ASIR-V (low-kV protocol) provided the identical image quality with the 120 kV; • the detection of lesions (polyps <5 mm, polyps ≥5 mm, adenoma, LST and, malignant tumor), if detected, on which protocol (120 kV, low-kV, or both) these lesions were better identified/analyzed. Only lesions detected and analyzed by both reviewers were ultimately considered.

Statistical analysis
The data were processed using SPSS 24.0 statistical software. X ± S was used to represent the one-way analysis of variance (ANOVA) for variables with a normal distribution. The Kruskal-Wallis H test was used to analyze variables with non-normal distribution. ANOVA with repeated measures was used to compare the SNR and CNR of ASIR-V images with varying weights. The Kappa test was utilized to estimate the consistency of two observers' subjective quality scores for the images. The Kruskal-Wallis H test was used to compare the radiation dose and the subjective quality scores. p < 0.05 was considered statistically significant.

Population
148 patients were approached, and 70 were excluded from the study. Seventy-eight patients were ultimately included ( Figure 1

Consistency of the lesions' location CTC/colonoscopy
All 78 patients had successful CTC examination and no repeated scans were needed. All the 78 patients underwent colonoscopy before or after the CTC examination. Colonoscopy showed that lesions in eight patients were in the ascending colon (10.26%), four in the right colic flexure of the colon (5.13%), one in the transverse colon (1.28%), one in the left colic flexure (1.28%), one in the descending colon (1.28%), 26 cases in the sigmoid colon (33.33%), five cases in the rectal sigmoid junction (6.41%), and 32 cases in the rectum (41.03%). Three patients' results of the 70 kV CTC scan protocol in the prone position with ultralow rectal cancer did not match with those of colonoscopy, and the other 24 patients were all consistent with colonoscopy. The results of CTC with 80 kV also showed that three patients with the rectal sigmoid junction did not agree with the results of colonoscopy. In 100 kV, the results showed that one patient with the ascending colon and two patients with rectum did not agree with the results of colonoscopy. Compared with colonoscopy, individualized low-kV CTC scan protocol had higher intrinsic consistency in confirming lesion' location.

Radiation dose
The

Qualitative image analysis
The subjective scores of the two observers on the image quality of each group were consistent (Kappa value = 0.684-0.907), and there was no statistically significant difference between the subjective scores of group B combined with 60% ASIR-V and group A (p > 0.05, Table 3). The SNR and CNR of group A1, A2, and A3 were higher than those of group B1, B2, and B3 below 60% ASIR-V, respectively, and the difference  was statistically significant (all p < 0.001), while the SNR between group A1/A2/A3 and B1/B2/B3 + 60%ASIR-V was not statistically significant (p > 0.05). There was no significant difference in the diagnostic confidence of 2Dimages and 3D-images between group A1/A2/A3 and B1/B2/B3 + 60% ASIR-V (all p ≥ 0.05). The 120 kV acquisition was preferred over the 80 kV for 100% of patients for both readers. Eighty-seven lesions in group A1 and B1 were detected and analyzed by both readers. Depending on the reader, Lesions were better seen

DISCUSSION
Colonoscopy is a preferred screening method for colorectal cancer and can be intervened during screening. However,colonoscopy is an invasive procedure requiring high tolerance, sedation, or analgesia. 17 On the contrary, CTC is a safe, noninvasive imaging technique that provides multidimensional images of the colorectal region in the preoperative evaluation of colorectal diseases. 18 Although several papers have been published, dose reduction and optimization in CT imaging still require consistent efforts. Over the past few years, various technological innovations have been introduced, such as bow-tie filter, 19 low-tube voltage scanning, 20 and iterative image reconstruction, 21 which steadily made efforts to reduce radiation dose. Questions such as how these techniques affect colorectal structural visibility and what is the best option for achieving a low radiation dose in CTC, remain unclear.
According to the principle of as low as reasonably achievable (ALARA), CT acquisition parameters should be optimized based on the patient's character to reduce radiation dose while meeting diagnostic requirements. Most studies have focused on middle-sized individuals rather than overweight or obese people when developing dose-reduction techniques. 22 To homogenize our findings in a population involving obese patients,we conducted three degrees of kV according to patients' body sizes. Unlike previous studies, this study uses BMI as the basis of individual acquisition parameters. In other words, our study selects the appropriate tube voltage according to the patient's BMI, followed by the optimal weight iterative reconstruction. Most previous studies reduced the radiation dose by decreasing mAs as the radiation dose has a linear relationship with mAs.
In contrast,the radiation dose has an exponential relationship with kV, so decreasing kV is more efficient in reducing the radiation dose. Low kV scanning can significantly increase image noise, while post-ASIR-V can reduce image noise even further. The square of tube voltage is approximately inversely proportional to the radiation dose, and lower tube voltage could reduce the radiation dose more effectively, which has been demonstrated that it could been successfully used in imaging techniques such as CTA. [23][24][25] Some studies showed that kVp may be selected based on body size (body mass index, body width on scout topogram, or patient weight), 26 for example, 100 kVp can be used for patients weighing less than 150 lbs. In our study, the mean weight was 67.68 ± 1.43 (SD) kg (150lbs = 68.0388555 kg). Our study selected suitable tube voltage and the optimal weight in iterative reconstruction based on BMI to carry out individualized low-dose CTC. Group A used conventional scanning with tube voltage 120 kVps, while group B according to the difference in BMI of patients, selected low tube voltage (70 kVp, 80 kVp, 100 kVp) for low-dose scanning. Post-ASIR-V has 11 weights ranging from 0% to 100%, so ASIR-V was included in this study. Adaptive Statistical Iterative Reconstruction (ASIR-V) is a rapid reconstruction technology at either the front-end or back-end. The former ASIR-V reduces radiation dose through an intelligently regulated tube current, and the post-ASIR-V had significant effects on noise reduction, contrast improvement, and artifacts elimination. There was no standard for the optimal weight of ASIR-V in optimized dose CTC images.
Furthermore, this study showed that individualized kV acquisition parameters based on the patient's BMI, combined with the optimal ASIR-V weight, had no significant impact on the 2D and 3D image quality and the diagnosis of colorectal diseases (Figure 3). 27 One advantage of ASIR-V reconstruction was low image noise, directly affecting image quality. The image with individualized low kV and high weight ASIR-V had the lowest noise, while the 2D axial image had the lowest score. In addition, the reduction of image noise was not necessarily accompanied by the improved 3D reconstructed image quality. In our study, subgroup analysis showed that low kV combined with the best weight ASIR-V protocol had an equal image quality with the conventional 120 kV combined with fixed weight ASIR-V protocol.
In a phantom study, Cohnen et al. assessed image quality and sensitivity at ultra-low radiation doses by using ultra-low tube currents to observe detectable colorectal polyps at 1 mSv doses. 28 The results demonstrated that the image noise decreased as ASIR-V weighting was increased; as a result, the SNR and the CNR were improved, and the CT value remained stable. Compared with the routine 120 kV with 30% ASIR-V images, low kV (70/80/100 kV) images showed no significant difference in the lesion morphology and could clearly show the intestinal tissue structure, which met the screening requirement of colon diseases. The DLP of the individualized CTC group decreased significantly (64.5%, 48.3%, and 33.4%, respectively), which fully proved the feasibility of CTC scanning. The lowest DLP of group B was 75.85 ± 9.75 mGy × cm, and the ED was 1.14 ± 0.15 mSv. Some studies showed that the iterative reconstruction algorithm could significantly reduce image noise and artifacts to improve detection accu-racy. The application of ASIR-V in CTC imaging was studied. It has been found that ASIR-V could enhance imaging quality and reduce radiation dose by 50%. The researchers found that it was possible to obtain accurate quantitative CTC images that were acceptable for image noise at 1 mSv dose levels. 29 In our study, increasing the iterative percentage of image reconstruction can significantly improve SNR and CNR. However, using a higher percentage of iterative reconstruction in the subgroup analysis of intestinal air did not significantly improve visual evaluation.
It was demonstrated that 60% ASIR-V was the optimal ASIR-V weight of the low kV CTC protocol. The results showed that 60% of the ASIR-V images could compensate for the noise caused by the reduction of kV, and the image quality is comparable to that of 120 kV. The subjective score of the axial images increased steadily in the range of 30%-60% ASIR-V but decreased after 60% ASIR-V. The subjective scores of 3D-reconstructed images of different ASIR-Vs are not significantly different,mainly because the mucosal and air interfaces in the 3D-reconstructed images are not sensitive to noise. The patient's prior intestinal preparation was the only factor that affected the image quality. 30 Therefore, we will explore individualized scanning solutions for patients with different body indexes, and find the optimal noise index and pre-ASIR-V weights for low kV scanning in the future.

CONCLUSIONS
In conclusion, the application of individualized tube voltage based on the patient's BMI combined with 60% ASIR-V technology for CTC imaging can retain good image quality, and at the same time, significantly reduce the radiation dose compared with the routine scan protocol of using a fixed 120 kV tube voltage.

AC K N OW L E D G M E N T S
Authors give heartfelt thanks to those who helped during the completion of this article. This research received no external funding.

C O N F L I C T O F I N T E R E S T S TAT E M E N T
The authors declare no conflict of interest.

I N S T I T U T I O N A L R E V I E W B OA R D S TAT E M E N T
The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Review Board (or Ethics Committee) of First Affiliated Hospital of Dalian Medical University (PJ-KS-KY-2019-49).

I N F O R M E D C O N S E N T S TAT E M E N T
Written informed consent has been obtained from the patient(s) to publish this paper.