Low concentration isotonic iodixanol on hepatic and portal veins imaging quality and renal function: A 320-detector CT dual-energy study

Abstract

Background:

The effects of different concentrations of isotonic contrast agents on the quality of hepatic and portal vein imaging and their impact on patients’ renal function remain unclear. This study aimed to assess whether lower concentrations of iodixanol yield high-quality imaging by comparing hepatic and portal vein imaging quality and evaluating renal function following dual-energy CT (DECT) scans.

Methods:

Patients were randomly assigned to receive either iodixanol at 270 mgI/mL (iodixanol 270 group, n = 50) or 320 mgI/mL (iodixanol 320 group, n = 50) and underwent DECT with energy-spectrum reconstruction. Quantitative and qualitative image quality parameters, as well as changes in renal function before and after scanning, were compared between the 2 groups.

Results:

Significant differences were observed in the contrast-to-noise ratio, signal-to-noise ratio of the main portal vein, and the computed tomography value of the hepatic vein between the 2 groups (P < .05). These parameters were superior in the iodixanol 320 group compared to the iodixanol 270 group. Qualitative evaluation scores for the portal and hepatic veins in the iodixanol 320 group were significantly higher than those in the iodixanol 270 group (P < .05). Postcontrast acute kidney injury incidence was higher in the iodixanol 320 group than in the iodixanol 270 group (20.5% vs 2.1%, P < .05), and this association remained significant after multivariable adjustment (adjusted odds ratio = 13.70, P = .002).

Conclusion:

With a fixed injection protocol, iodixanol 320 mgI/mL yielded higher venous enhancement, whereas iodixanol 270 mgI/mL provided diagnostically acceptable portal/hepatic venous image quality on DECT with a lower iodine load. After adjustment, 320 mgI/mL was associated with higher odds of postcontrast acute kidney injury; this safety signal requires confirmation in larger iodine-matched studies.

1. Introduction

Iodixanol is a widely used nonionic isotonic contrast agent.^[1] The application of a high dose of iodixanol plays a significant role in producing better quality venous phase images following arteriography and arterial phase images due to features of high hydrophilic, low chemotoxicity, and iso-osmolality.^[2,3] Even though iodixanol is widely used in clinical computed tomography (CT) scanning, postcontrast acute kidney injury (PC-AKI), which used to be called contrast-induced nephropathy, is not avoided. PC-AKI is an iatrogenic syndrome characterized by acute renal failure following the intravascular administration of iodinated contrast agents. This condition is typically defined by an absolute or relative increase in creatinine levels post-exposure to contrast agents, compared to baseline values, in the absence of other causes of renal impairment.^[4] Common risk factors include advanced age, diabetes mellitus, preexisting renal insufficiency, congestive heart failure, hemodynamic instability, and nephrotic syndrome. Additionally, the volume and type of contrast administered play crucial roles in the toxicity and are significant determinants of this condition.^[5] Previous studies have focused on pathology, risk factors, prognosis, and treatments for PC-AKI.^[6] It is rare to discuss the effects of different concentrations of isotonic contrast agents on the quality of integrated imaging of the hepatic and portal veins and their effects on patients’ renal function. However, in conventional single-energy CT, reducing contrast concentration under a fixed injection protocol inevitably leads to a lower total iodine delivery rate, which is the primary determinant of vascular enhancement.^[7] This fundamental limitation often results in compromised image quality, particularly in venous-phase imaging. Dual-energy CT (DECT) may mitigate this drawback: advanced post-processing techniques such as material decomposition and virtual monoenergetic imaging, can enhance iodine conspicuity – especially at low keV – thereby improving contrast-to-noise ratio (CNR).^[8,9] Accordingly, we hypothesize that for portal and hepatic venous imaging, DECT-based optimization can offset the lower iodine load delivered by 270 mgI/mL iodixanol (compared to 320 mgI/mL), yielding comparable diagnostic image quality while potentially offering renal safety advantages associated with a lower-concentration, isotonic agent. This hypothesis remains inadequately tested, because prior comparative studies have predominantly focused on arterial-phase imaging.^[10] Therefore, this study aims to investigate whether venous-phase DECT using lower-concentration iodixanol (270 mgI/mL) can achieve non-inferior image quality for portal and hepatic vein assessment compared to the standard concentration (320 mgI/mL), and evaluate its effects on renal function.

2. Material and methods

2.1. Patient selection

The study received approval from the Research Ethics Committee of our Hospital. All participants provided written informed consent prior to their inclusion in the study.

Recruitment occurred from June 2022 to August 2023, targeting participants aged 22 to 90 years undergoing upper abdomen CT scans at our hospital. A total of 100 patients were enrolled, comprising 53 males and 47 females.

Patients were eligible if they had met all the following conditions: The location and size of liver lesions or masses did not affect the observation of the portal vein; and Patients could control the automatic respiratory rhythm.

Patients were not eligible if they had any one of the following conditions: allergy to iodine contrast agents; hyperthyroidism; and severe heart, liver, or kidney insufficiency.

2.2. Randomization and blinding

Patients were randomly assigned to one of the 2 groups: intravascular administration of iodixanol at a concentration of 270 mgI/mL (iodixanol 270 group) or iodixanol at 320 mgI/mL (iodixanol 320 group). A statistician generated the randomization sequence using SAS software (version 9.4; SAS Institute Inc, Cary). Allocation concealment was ensured using sequentially numbered, opaque, sealed envelopes prepared by the same statistician, who was not involved in patient enrollment or subsequent data collection. Patients, radiologists, and data analysts performing quantitative image measurements were blinded to group assignment. Radiologists reviewed anonymized image sets without any clinical information in a random order, and qualitative scoring was performed primarily on standard grayscale-blended images resembling conventional CT to minimize bias.

2.3. Sample size consideration

A total of 100 patients (50 per group) was planned based on pragmatic considerations and the sample sizes used in previous similar imaging comparison studies. We conducted post hoc power analyses for the key outcomes. For the primary imaging outcome (portal vein CNR), the achieved power was 89% at α = 0.05, indicating adequate sample size. In contrast, for the renal safety outcome (incidence of PC-AKI), the post hoc power was only 3% given the observed incidence rates (2.1% vs 20.5%). This confirms that the study is underpowered to detect differences in renal safety outcomes, particularly for the relatively low incidence of PC-AKI.

2.4. Outcomes

The primary endpoint was the CNR of the main portal vein and the hepatic veins. The CNR for the portal vein was calculated as (CT_{main portal vein} − CT_{erector spinae})/noise, and for the hepatic vein as (CT_{hepatic vein} − CT_{liver parenchyma})/noise, where noise was defined as the standard deviation (SD) of the subcutaneous fat of the anterior abdominal wall.

Secondary endpoints included: the CT value (in Hounsfield Units) of the main portal vein and hepatic veins; the signal-to-noise ratio (SNR) of the main portal vein and hepatic veins; the qualitative evaluation score of portal and hepatic vein image quality based on a predefined 5-point scale; image noise; and renal safety outcomes, including changes in serum creatinine (SCr) and estimated glomerular filtration rate (eGFR) from baseline, and the incidence of post-contrast acute kidney injury (PC-AKI).

2.5. CT scan procedure

A CT scan of the upper abdomen was performed using Canon Aquilion ONE TSX-301C CT (Canon Medical Systems, Tokyo, Japan). After the routine scan, iodixanol was mass-injected through the elbow vein. When the CT threshold value of the abdominal aorta reached 240 HU, the arterial phase scan was delayed for 8 seconds, and then the portal vein phase and equilibrium phase scan were delayed for 30 and 120 seconds, respectively.

Iodixanol was used with a concentration of 270 mgI/mL in the iodixanol 270 group and 320 mgI/mL in the iodixanol 320 group. The total volume of iodixanol administered was 1.5 mL/kg, with a flow rate of 3.0 mL/s. This fixed-volume, fixed-flow-rate protocol reflects a common clinical practice where contrast volume is determined by patient weight, irrespective of agent concentration. Consequently, the total iodine load delivered was inherently higher in the 320 mgI/mL group. A dual-energy scan using a 320-detector was conducted during the portal vein phase, with the tube voltage alternating instantaneously between 80 and 135 kV. Automatic mAs matching technology was utilized. The scan parameters included a pitch of 0.9, a layer thickness of 5 mm, a field of view of 332 mm, and a rotation time of 0.5 seconds.

2.6. CT image reconstruction

The original data of the portal vein phase energy spectrum scan was analyzed using a dual source module on the Aquilion ONE workstation. CT scanners selected the enhanced mode in “DE Image View” list for data analysis, and the system automatically reconstructed contrast-enhanced images. Reconstructed contrast-enhanced images were then transferred to an ISP workstation, and volume rendering, maximum intensity projection, and multiplane reconstruction were used to complete future image reconstruction.

2.7. Image quality evaluation standards

Two abdominal radiologists (with 8 and 12 years of experience, respectively) independently evaluated image quality. Before studying the reading, both readers participated in a joint training/calibration session using sample cases not included in the study to standardize the application of the scoring criteria. They performed the readings independently in separate sessions, under standardized conditions in a dimly lit room using the same workstation and consistent abdominal window settings (width 350 HU, level 40 HU). They were blinded to clinical information and contrast concentration. The scores from the first reader were used for the primary analysis. Inter-reader agreement was assessed using the intraclass correlation coefficient (ICC) based on a two-way random-effects model with absolute agreement.

2.8. Quantitative evaluation parameters and qualitative evaluation score

One observer measured CT values of the main portal vein, liver parenchyma, hepatic vein, erector spinae, and subcutaneous fat of the anterior abdominal wall in the same layer in 2 groups. The SD of the hepatic vein CT value was the average of the left, middle, and right hepatic veins. SD of subcutaneous fat of the anterior abdominal wall was defined as image noise. The CNR and SNR were calculated using the following formulas. CNR_{main portal vein} = (CT_{main portal vein} − CT_{erector spinae})/SD_{abdominal wall fat}, SNR_{main portal vein} = CT_{main portal vein}/SD_{abdominal wall fat}. CNR_{hepatic vein} = (CT_{hepatic vein} − CT_{liver parenchyma})/SD_{abdominal wall fat}, SNR_{hepatic vein} = CT_{hepatic vein}/SD_{abdominal wall fat}.

Qualitative evaluation score of the portal and hepatic veins was based on the 5-grade system: The edge of the blood vessel was sharp, and fourth-grade and above branches could be seen, scored 5; the edge of the blood vessel was sharp, and third-grade branches could be seen, scored 4 points; the edge of the blood vessel was clear, and second-grade branches could be seen, scored 3; the edge of the blood vessel was not clear, and only first-grade branches were seen, scored 2; and only the main vessel was shown, scored 1.

2.9. Assessment of renal function

All patients were encouraged to maintain oral hydration before and after the CT. However, no standardized intravenous hydration protocol was implemented. Information on the use of nephrotoxic medications (e.g., nonsteroidal anti-inflammatory drugs and aminoglycosides) within 48 hours prior to the scan was not systematically recorded.

All patients fasted for 4 to 6 hours before the CT scan and drank 1000 mL warm water half an hour before the scan. SCr was measured 24 hours before (pre-scan stage) and 48 to 72 hours after iodixanol injection (post-scan stage), and the glomerular filtration rate (GFR) was calculated using CKD-EPI software. According to the GFR at the pre-scan stage, patients were divided into a normal renal function group (GFR > 90 m L•min⁻¹•1.73m⁻²) and an abnormal group (30–89 m L•min⁻¹•1.73m⁻²). PC-AKI refers to acute kidney injury occurring within 48 hours after contrast medium injection. The detailed definition is the increase of SCr of more than 26.5 μmol/L or more than 50% within 48 to 72 hours after the injection of an iodine contrast agent.

SCr, GFR, and PC-AKI ratio of the 2 groups were collected at the pre- and post-scan stages.

2.10. Statistical analysis

All analyses were performed using SPSS version 17.0 (SPSS Inc., Chicago). Continuous variables were reported as mean ± SD and categorical variables as counts (percentages). Between-group comparisons were performed using Student’s t tests for normally distributed continuous variables, nonparametric tests for non-normal distributions, and χ² tests for categorical variables. To control for multiplicity, we prespecified different adjustments for primary and secondary outcomes: a Bonferroni correction was applied to the 2 primary endpoints (CNR of the main portal vein and hepatic vein), with statistical significance set at P < .025; for multiple secondary quantitative comparisons (e.g., CT attenuation and SNR across vessels/organs), the false discovery rate (FDR) method was applied. The manuscript reports P values after these adjustments. To facilitate interpretation beyond P values, we additionally report 95% confidence intervals (95% CI) for between-group mean differences (MDs) for the primary endpoints and other key comparisons, and we provide standardized effect sizes (Cohen d) for the primary endpoints. A two-sided P < .05 was considered statistically significant unless otherwise specified by the multiplicity adjustment. Interobserver agreement for qualitative image-quality scores was assessed using the ICC, with ICC > 0.75 indicating high agreement. To account for potential confounding in the analysis of PC-AKI, we performed multivariable logistic regression with PC-AKI as the dependent variable and contrast concentration group (270 vs 320 mgI/mL) as the primary predictor, adjusted for prespecified covariates: age, sex, baseline eGFR (continuous), and diabetes mellitus.

3. Results

3.1. Baseline characteristics

Baseline characteristics were well balanced between the 2 groups, with no significant differences in age, weight, height, or body mass index (all P > .05; Table 1).

Table 1.

Baseline characteristics.

Group	Age (years)	Weight (kg)	Height (cm)	BMI
Iodixanol 270 n = 50	54.45 ± 15.23	59.27 ± 12.87	160.14 ± 8.63	23.24 ± 4.33
Iodixanol 320 n = 50	53.36 ± 12.05	59.44 ± 11.44	161.05 ± 7.18	22.70 ± 3.57
t value	0.360	−0.065	−0.520	0.617
P value	.719	.949	.604	.539

BMI = body mass index.

3.2. Quantitative evaluation parameters of image quality

After Bonferroni correction for the 2 primary endpoints, the main portal vein CNR differed significantly between groups (MD: 3.76, 95% CI: 2.08 to 5.44; P < .001; Cohen *d* = 0.97), whereas hepatic vein CNR did not (MD: 0.66, 95% CI: −0.58 to 1.90; P = .298; Cohen *d* = 0.24). For the secondary quantitative endpoints, FDR-adjusted analyses showed significant between-group differences in the main portal vein CT attenuation (MD: 42.63 HU, 95% CI: 27.48 to 57.78 HU) and SNR, and in hepatic vein CT attenuation (all P < .05). No significant differences were observed for hepatic vein SNR or image noise after FDR correction (all P > .05; Table 2). The distribution of this primary endpoint is visually summarized in Figure 1, which shows a marked upward shift in the 320 mgI/mL group.

Table 2.

Quantitative evaluation parameters of image quality.

Parameter	Group		Mean difference (95% CI)	Effect size (Cohen*d*)	t Value	P value
	Iodixanol 270 n = 50	Iodixanol 320 n = 50
Main portal vein
CT value (HU)	182.21 ± 30.85	224.84 ± 39.36	42.6 (27.5 to 57.8)	1.25	5.76	.000
CNR	8.82 ± 3.45	12.58 ± 4.26	3.8 (2.1 to 5.4)	0.97	−4.468	.000
SNR	13.85 ± 4.03	18.29 ± 5.17	4.4 (2.6 to 6.2)	0.95	−4.461	.000
Hepatic vein
CT value (HU)	176.15 ± 32.61	197.99 ± 25.63	21.8 (8.9 to 34.7)	0.74	−3.429	.001
CNR	6.04 ± 2.55	6.70 ± 3.02	0.7 (−0.6 to 1.9)	0.24	−1.047	.298
SNR	14.72 ± 3.73	16.21 ± 4.42	1.5 (−0.3 to 3.3)	0.37	−1.615	.110
Image noise (HU)	13.05 ± 3.39	13.79 ± 2.9	0.7 (−0.5 to 1.9)	0.23	−1.130	.261

The data are presented as mean ± standard deviation. The primary endpoint (after Bonferroni correction, with a significance level set at P < .025). Secondary endpoint, after correction for false discovery rate (FDR), P < .05. Interpretation of effect size: Cohen *d* values are typically interpreted as: ~0.2 (small effect), ~0.5 (medium effect), ≥0.8 (large effect).

CI = confidence intervals, CNR = contrast-to-noise ratio, CT = computed tomography, HU = Hounsfield unit, SNR = signal-to-noise ratio.

Figure 1.

Box plots comparing the primary quantitative endpoints between contrast agent groups. Distribution of (A) the contrast-to-noise ratio (CNR) of the main portal vein and (B) the CNR of the hepatic veins for the iodixanol 270 mgI/mL group and the iodixanol 320 mgI/mL group. The central line within each box represents the median; the box extends from the 25th to the 75th percentile. P < .001.

3.3. Qualitative evaluation score of image quality

The qualitative evaluation score of portal vein imaging derived from one observer versus another observer was 3.06 ± 0.73 versus 2.98 ± 0.76 in the iodixanol 270 group and 3.56 ± 0.86 versus 3.52 ± 0.76 in the iodixanol 320 group (P < .05). The qualitative evaluation score of hepatic vein imaging derived from one observer versus another observer was 2.76 ± 0.89 versus 2.72 ± 0.92 in the iodixanol 270 group and 3.16 ± 0.88 versus 3.08 ± 0.80 in the iodixanol 320 group (P < .05). The qualitative evaluation scores for both the portal and hepatic veins in the iodixanol 320 group were significantly higher than those in the iodixanol 270 group (P < .05). Representative cases (Figs. 2 and 3) illustrate improved vascular opacification and delineation with 320 mgI/mL on DECT reconstructions. Nevertheless, the image quality with 270 mgI/mL was still rated as diagnostically acceptable in most cases. The consistency of the qualitative evaluation scores between 2 different observers was assessed using the ICC. The ICC for the portal vein imaging evaluation was 0.883, and for the hepatic vein imaging evaluation, it was 0.952. Both values are higher than 0.75, indicating high consistency.

Figure 2.

Representative images for qualitative image-quality comparison from a 45-year-old man with a hepatic cyst (77.5 kg, 169 cm; BMI 27.1 kg/m²) receiving iodixanol 270 mgI/mL. (A) Conventional single-energy helical images of the portal vein: portal vein attenuation 131.2 HU, CNR 3.1, qualitative score 2. (B) Conventional single-energy helical images of the hepatic vein: hepatic vein attenuation 102.4 HU, CNR 1.97, qualitative score 2. (C) DECT with contrast-enhanced reconstruction of the portal vein: portal vein attenuation 187.5 HU, CNR 15.1, qualitative score 4. (D) DECT with contrast-enhanced reconstruction of the hepatic vein: hepatic vein attenuation 160.9 HU, CNR 9.8, qualitative score 4. Compared to conventional single-energy helical images, DECT contrast-enhanced reconstruction showed improved opacification and delineation of both the portal and hepatic veins. BMI = body mass index, CNR = contrast-to-noise ratio, DECT = dual-energy CT.

Figure 3.

Representative images for qualitative image-quality comparison from a 62-year-old woman with colorectal cancer and liver metastases (70.0 kg, 158 cm; BMI 28.0 kg/m²) receiving iodixanol 320 mgI/mL. (A) Conventional single-energy helical images of the portal vein: portal vein attenuation 178.5 HU, CNR 8.1, subjective score 3. (B) Conventional single-energy helical images of the hepatic vein: hepatic vein attenuation 166.0 HU, CNR 4.1, subjective score 3. (C) DECT with contrast-enhanced reconstruction of the portal vein: portal vein attenuation 247.7 HU, CNR 22.64, subjective score 5. (D) DECT with contrast-enhanced reconstruction of the hepatic vein: hepatic vein attenuation 244.0 HU, CNR 15.57, subjective score 5. Compared to conventional single-energy helical images, DECT contrast-enhanced reconstruction showed improved portal and hepatic venous opacification and delineation. BMI = body mass index, CNR = contrast-to-noise ratio, DECT = dual-energy CT.

3.4. Assessment of renal function

There were no significant differences in SCr and GFR in both groups, even at the pre-scan or post-scan stage (P > .05). However, the incidence of PC-AKI was higher in the iodixanol 320 group (20.5% [95% CI: 10.8% to 34.1%]) than in the iodixanol 270 group (2.1% [95% CI: 0.4% to 10.7%]). To control for potential confounding variables, a multivariable logistic regression analysis was conducted. After adjusting for age, sex, baseline eGFR, and the presence of diabetes mellitus, the use of iodixanol 320 mgI/mL remained independently associated with a significantly higher odds of developing PC-AKI compared to iodixanol 270 mgI/mL (adjusted odds ratio [aOR] = 13.70, 95% CI: 2.51 to 74.81, P = .002). In this model, lower baseline eGFR was also a significant independent predictor of PC-AKI (aOR = 0.96 per 1 mL/min/1.73m² increase, 95% CI: 0.93 to 1.00, P = .019), while hypertension was not a significant predictor (aOR = 0.93, 95% CI: 0.23 to 3.80; P = .923; Tables 3–4).

Table 3.

Assessment of renal function.

Group	Serum creatinine (μmol/L)			GFR (mL.min−1.1.73m−2)			PC-AKI (% [95% CI])
	Pre-scan	Post-scan	P value	Pre-scan	Post-scan	P value
Iodixanol 270 (270 mgI/mL)
Total	78.81 ± 23.18	77.08 ± 25.27	.728	81.16 ± 33.27	83.32 ± 32.36	.748	2.08% [0.4% to 10.7%]
Normal	66.43 ± 12.13	66.94 ± 13.74	.898	109.27 ± 25.01	108.84 ± 24.58	.954	0.00% [0.0% to 16.8%]
Abnormal	89.28 ± 25.26	85.66 ± 29.57	.638	57.38 ± 16.47	61.73 ± 20.08	.397	3.84% [0.7% to 18.9%]
Iodixanol 320 (320 mgI/mL)
Total	78.78 ± 26.08	75.95 ± 26.98	.639	81.14 ± 31.88	86.67 ± 37.36	.484	20.51% [10.8% to 34.1%]
Normal	68.91 ± 13.30	75.50 ± 22.08	.617	106.07 ± 31.21	102.85 ± 38.51	.791	17.64% [6.8% to 37.5%]
Abnormal	86.40 ± 30.92	76.84 ± 30.72	.309	61.87 ± 14.34	74.16 ± 31.93	.425	22.72% [10.1% to 42.8%]

The data are presented as mean ± standard deviation or percentage; GFR, PC-AKI, contrast agent-induced acute kidney injury. The incidence of PC-AKI between groups was compared using the chi-square test (270 vs 320 P = .007).

CI = confidence intervals, GFR = glomerular filtration rate, PC-AKI = postcontrast acute kidney injury.

Table 4.

Multivariable logistic regression analysis for predictors of PC-AKI.

Variables	Coefficient (β)	Standard error	Adjusted OR	95% CI	P
Contrast group (320 vs 270)	2.618	0.866	13.70	2.51–74.81	.002
Age (yr)	0.035	0.028	1.04	0.98–1.10	.209
Gender (Male vs Female)	−0.201	0.718	0.82	0.20–3.36	.779
Baseline eGFR	−0.038	0.016	0.96	0.93–1.00	.019
Diabetes (Yes vs No)	0.186	0.900	1.20	0.21–7.03	.837
Hypertension (Yes vs No)	−0.069	0.716	0.93	0.23–3.80	.923

CI = confidence intervals, PC-AKI = postcontrast acute kidney injury.

4. Discussion

CT vascular imaging always aims to have high-quality images using low-concentration contrast agents and low radiation doses.^[11–13] CT angiography (CTA) demonstrated that reducing contrast media doses could maintain both quantitative and qualitative image quality. However, qualitative image quality scores for CTA were poorer due to contrast media dose reduction, especially for the accurate assessment of small vessels.^[14,15] In clinical practice, physicians read abdominal CTA images scanned in the hepatic arterial phase (H-phase), portal vein phase (P-phase), and equilibrium phase (E-phase). Traditional CT scans and trigger methods make segmenting the portal vein on P-phase images complex even without considering other phase images.^[16] Lower-concentration contrast agents would make obtaining good-quality CT images of portal and hepatic veins more difficult. Previous studies predicted that applying energy spectrum CT and related reconstruction methods could improve the image quality of CTA significantly.^[8,17,18] Using a lower concentration isotonic contrast agent with a CT dual-energy scan and related reconstruction modes holds promise for acquiring high-quality images of the portal and hepatic veins during the late portal vein stage while minimizing damage to renal function and the incidence of contrast agent-induced nephropathy.

CTA often faces challenges and requires careful calibration of delay start times and contrast agent volumes to achieve optimal opacification of both large and small vessels. Pop compared contrast enhancement achieved by 4 different contrast agents, including sub-hypertonic and isotonic contrast agents, considering both contrast agent injection parameters and patient characteristics.^[19] Zhang et al evaluated the effect of different iodinated contrast agents on coronary contrast enhancement using a contrast agent injection protocol according to body surface area.^[20] However, CTA imaging quality derived from 2 different concentrations of isotonic contrast agents and the impacts on renal function of 2 different concentrations of isotonic contrast agents were rare and not well studied in detail.

In this trial, image quality evaluation of the portal and hepatic veins, including quantitative and qualitative evaluations, was obtained by using different concentrations of isotonic contrast agents of iodixanol, 270 and 320 mgI/mL, under the same total volume and flow rate. This fixed-volume, fixed-flow-rate protocol was an intentional design choice to reflect a common clinical scenario in which a standard volume-based protocol is used irrespective of concentration, resulting in a higher total iodine load for the 320 mgI/mL group. The original data from the portal vein phase dual-energy scan was reconstructed to optimize images. In quantitative evaluation, parameters of image quality, CNR, SNR value of the main portal vein, and CT value of the hepatic vein in the iodixanol 320 group were improved than the iodixanol 270 group. However, no significant differences were found in CNR, SNR value of the hepatic vein, and image noise between the 2 groups. The qualitative evaluation scores of the portal vein and hepatic vein in the iodixanol 320 group were significantly higher than the iodixanol 270 group. Some of the results mentioned above were inconsistent with previous studies.^[21,22] The causes of these results were supposed to be followed. At first, abdominal CTA images were usually obtained in the hepatic arterial phase, and trigger-tracking technology was used in previous studies of isotonic low-concentration contrast agents. The precision of trigger time points and scan phase detection improved imaging quality. However, this study concentrated on CTA images of the portal venous and hepatic venous phases. Traditional triggering methods struggled to accurately segment these phases without considering images from other phases, resulting in lower qualitative and quantitative evaluation scores for image quality. Another factor contributing to this was the flow rate used in this study, which was only 3 mL/s – lower than that used in previous studies. Morisaka et al evaluated the effect of decreased injection flow rates of contrast agents on aortic peak enhancement using a dynamic flow phantom and on aortic enhancement in clinical dynamic 80-kVp CT with a reduced contrast dose. They concluded that a decreased injection flow rate had a minor impact on aortic enhancement in vitro but did not significantly affect aortic enhancement in clinical dynamic 80-kVp CT.^[23]

While early studies suggested that DECT could improve image quality with lower contrast doses,^[8,10] more recent evidence has demonstrated that advanced DECT reconstructions can partly compensate for the lower iodine load of a 270 mgI/mL agent and achieve image quality comparable to 320 mgI/mL in arterial-phase imaging.^[24] In the present study, the 320 mgI/mL group demonstrated higher objective enhancement parameters; however, the clinical importance of these quantitative differences remains uncertain. Quality scores, which reflect overall diagnostic confidence, did not indicate a meaningful deficit with 270 mgI/mL, and most examinations in both groups were rated diagnostically acceptable or good. This suggests that, for the routine assessment of portal and hepatic venous anatomy in the studied phase, the lower enhancement with 270 mgI/mL did not meaningfully compromise interpretability. However, this study was not specifically designed to evaluate diagnostic accuracy for focal liver lesions. Future investigations should assess whether higher CNR or alternative reconstructions translate into measurable improvements in lesion detection rates, conspicuity, or inter-reader agreement for specific pathological findings.

Renal safety was a key secondary consideration. The observed incidence of PC-AKI in our study, particularly in the iodixanol 320 mgI/mL group, was higher than typically reported for iso-osmolar in contemporary literature. However, our study was not primarily designed or powered as a renal safety trial. Interpretation is limited by the absence of a controlled intravenous hydration protocol, the lack of systematic data on concomitant nephrotoxic medications, and incomplete adjustment for relevant comorbidities (e.g., diabetes mellitus and heart failure) known to influence renal risk. Therefore, the between-group difference in PC-AKI should be interpreted cautiously, given potential residual confounding and modest event counts. Nevertheless, in multivariable logistic regression analysis adjusting for age, sex, baseline renal function (eGFR), and diabetes, 320 mgI/mL remained independently associated with a significantly increased odds of PC-AKI, although the wide CI reflects substantial uncertainty.

Several limitations existed in this study. First, portal venous phase timing used a fixed 70-seconds delay rather than bolus tracking or individualized timing; inter-patient variation in hemodynamics may have influenced venous enhancement independent of contrast concentration. Second, both groups received the same volume and flow rate, resulting in different total iodine delivery and inherently favoring higher quantitative enhancement with 320 mgI/mL. Third, the phase identification/segmentation approach was not fully automated and may be less reliable in patients with altered circulation (e.g., cardiocirculatory disease) or severe hepatic dysfunction. Fourth, creatinine was measured only at 48 to 72 hours post-contrast, which may miss earlier transient changes or delayed rises beyond this window. Fifth, we did not perform dedicated phantom validation under our specific DECT settings. Finally, this single-center study used one scanner platform and a fixed weight-based injection protocol with a relatively small sample size, limiting generalizability, particularly for subgroup and safety analyses, and residual confounding cannot be excluded.

In conclusion, using a one-stop venous dual-energy CT protocol for portal and hepatic vein imaging, iodixanol 320 mgI/mL provided higher quantitative enhancement, in part because our volume- and flow-fixed injection strategy resulted in a higher total iodine load. Despite lower enhancement, iodixanol 270 mgI/mL achieved diagnostically acceptable image quality with DECT reconstruction. In adjusted analyses, iodixanol 270 mgI/mL was associated with lower odds of PC-AKI; however, renal safety findings should be interpreted cautiously given the study’s design and limited power for this endpoint. Overall, iodixanol 270 mgI/mL appears to be a feasible lower-iodine option for venous-phase portal and hepatic vein assessment on DECT, warranting confirmation in larger, iodine-matched prospective studies.

Author contributions

Data curation: Zhengyang Gao, Guihua Li, Taisong Tang, Mei Wang, Yi Peng.

Formal analysis: Taisong Tang.

Investigation: Zhengyang Gao, Guihua Li, Lunqing Pu.

Project administration: Jingjing Wu.

Resources: Lunqing Pu.

Supervision: Jingjing Wu.

Validation: Jingjing Wu.

Visualization: Jingjing Wu.

Writing – original draft: Jingjing Wu.

Writing – review & editing: Jingjing Wu.