Evaluation of a personalized total iodine dose (TID) combined with fixed injection duration (FID) protocol using high-concentration (400 mgI/mL) iodinated contrast medium in CT portal venography: a randomized trial Authors: Fanglu Zhou, Zhiwei Zhang, Hongwei Liang, Renqiang Yu, Wangjia Li, Xinyou Li Published: 2026 ABSTRACT BackgroundThe portal vasculature is closely associated with clinical outcomes in liver disease. Contrast-enhanced computed tomography (CT) serves as the first-line noninvasive tool for its evaluation. The contrast administration strategy critically influences CT image quality and diagnostic performance. This study aimed to investigate the effect of a personalized total iodine dose (TID) combined with a fixed injection duration (FID) protocol on CT imaging of the portal vein using a high-concentration iodine contrast medium (CM).MethodsA GE 64-slice spiral CT scanner and a 400 mgI/mL iodine CM were used. The total CM dose was calculated based on the patient’s total body weight (TBW) at doses of either 400 or 450 mgI/kg. The FID was set at 18 or 20 s, and the tube voltage was set at 100 or 120 kVp. These three parameters were varied to generate eight distinct injection and scanning protocols. Eligible patients were randomized into eight groups. The CT values of the main portal vein (CTmpv), left superior branch of the portal vein (CTlspv), and liver parenchyma (CTliver), image noise, contrast-to-noise ratio (CNR), volume CT dose index (CTDIvol), and dose-length product (DLP) were measured. Additionally, overall image quality, portal vein lesion detection, liver lesion detectability, maximum intensity projection (MIP), and volume rendering (VR) reconstructions of the portal vein were evaluated.ResultsLower tube voltage significantly increased CTmpv (P<0.001, partial η2=0.229), CTlspv (P<0.001, partial η2=0.127), CTliver (P<0.001, partial η2=0.089), and image noise (P<0.001, partial η2=0.118), while significantly reducing CTDIvol (P<0.001, partial η2=0.152) and DLP (P<0.001, partial η2=0.209). FID significantly affected CTmpv (P=0.039, partial η2=0.025) and CNR (P=0.009, partial η2=0.04) (both CTmpv and CNR were higher in the 20 s FID group than in the 18 s group). In the subjective evaluation, tube voltage significantly influenced portal vein lesion detection (P=0.004, partial η2=0.048); other scores (overall image quality, liver lesion detectability, VR, and MIP) showed no significant differences among the three factors, with no interactions. Meanwhile, the main effect of TID and FID, as well as all interaction effects, was non-significant (P>0.05).ConclusionsWhen using high-concentration iodine CM, all imaging protocols achieved satisfactory portal vein enhancement. The protocol combining a lower flow rate (20-s FID), a lower TID (400 mgI/kg), and a tube voltage of 100 kVp improved portal vein CT values and reduced radiation dose without compromising image quality, and thus represents a promising protocol for clinical CT portal venography. Introduction The portal venous system is the major source of hepatic blood supply. Its patency, anatomical variations, and the formation of collateral circulation are closely related to the diagnosis, management, and prognosis assessment of various liver diseases, including cirrhosis, portal hypertension, and liver cancer (1). Computed tomography angiography (CTA) has become the preferred noninvasive method for evaluating the portal vein (2,3); however, its imaging quality is highly dependent on the concentration of iodinated contrast medium (CM) in the vascular lumen (4). With continuous advancements in clinical practice, CM dosing regimens have evolved to optimize enhancement while maintaining patient safety. However, standardized recommendations or widely accepted guidelines for CM administration have yet to be established. To address the limitations of conventional protocols, modern abdominal computed tomography (CT) protocols are moving toward individualized iodine delivery rather than fixed dosing strategies (5-7). Studies have demonstrated that adjusting total iodine dose (TID) based on body weight can achieve more consistent target organ enhancement while reducing unnecessary iodine load (8-10). Among these, total body weight (TBW) has become a key determinant of contrast enhancement and is particularly suitable for clinical implementation due to its simplicity and ease of use (11). TID does not determine enhancement independently; rather, TID and the injection flow rate interact synergistically to determine the peak and time-to-peak of the vascular enhancement curve (12). Moreover, studies have shown that the fixed injection duration (FID) protocol provides superior vascular enhancement. This indicates that, compared to the fixed flow rate protocol, the FID protocol can better accommodate individual differences, particularly in patients with high body weights (13). The synergistic regulation of total dose and flow rate is crucial to obtain an enhancement curve morphology that meets diagnostic requirements under individualized low-dose conditions. Previous studies on body weight-based personalized dosing regimens for abdominal multiphase CT have primarily used medium-concentration CM, such as iopamidol at 370 mgI/mL or iohexol at 350 mgI/mL (14,15). The maximum iodine concentration of the non-ionic water-soluble iodinated CM iomeprol is 400 mgI/mL. High-concentration (400 mgI/mL) iomeprol offers unique advantages in multiphase abdominal contrast-enhanced CT and CTA examinations. High-concentration CM not only enhances vascular and parenchymal image quality but also improves lesion visibility (16,17). Compared to low-concentration CM, high-concentration CM allows for lower injection volumes and slower injection flow rates, thereby increasing patient comfort and tolerance (18). However, research on the use of 400 mgI/mL CM in abdominal CT remains limited, particularly regarding the application of individualized dosing strategies. This study aimed to investigate a TBW-based personalized TID combined with an FID protocol using high-concentration CM (400 mgI/mL) in abdominal multiphase contrast-enhanced CT, with a focus on differences in portal vein enhancement, overall image quality, and liver parenchymal enhancement. We present this article in accordance with the CONSORT reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0816/rc). Methods Patients This mixed-design study comprised an initial prospective pilot phase followed by a subsequent formal prospective randomized controlled trial (RCT) phase. No primary outcomes were altered after the formal RCT phase commenced. This study was not prospectively registered due to the low-risk nature of the imaging parameter optimization. The sample size was calculated based on the primary outcome [the CT value of the main portal vein (CTmpv)], assuming a mean difference of 16 Hounsfield units (HU) between extreme groups [standard deviation (SD) =20 HU] within a 2×2×2 factorial design (eight groups). To achieve 80% power at a significance level of α = 0.05, 26 patients per group were required (total n=208), after accounting for an anticipated 5% dropout rate. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board of The First Affiliated Hospital of Chongqing Medical University (ethics No. K2023-356), and informed consent was obtained from all patients. For a subset of patients enrolled in an initial pilot phase, a waiver of informed consent was granted as the study involved retrospective analysis of de-identified data without additional interventions. A total of 227 patients who underwent multi-slice spiral CT examination of the portal vein at the radiology department of The First Affiliated Hospital of Chongqing Medical University from July 2021 to June 2025 were enrolled in the study. The inclusion criteria were as follows: consciousness and normal cognitive function; the ability to actively cooperate with the examination; no contraindications to CT scanning; the capacity to provide written informed consent; and an age of 18 years or older. The exclusion criteria were as follows: (I) allergy to iodinated CM or renal insufficiency precluding the examination; and/or (II) incomplete baseline data, inadequate breath-holding, or severe image artifacts (e.g., artifacts caused by internal metal objects) compromising image quality or parameter measurements. Eligible patients were randomized into eight groups. The randomization sequence was generated by a statistician using a computerized block randomization method with variable block sizes of 8, 16, and 24. Allocation concealment was ensured by sequentially numbered, opaque, sealed envelopes. Both the research coordinators who enrolled the patients and the patients were blinded to group allocation. The CT radiographers were not blinded due to the need to adjust injection parameters. Outcome assessors (radiologists) and data analysts accessed only de-identified and coded image data, and were thus blinded to group assignment. Patients were not involved in the design, conduct, reporting, or dissemination of this study. Scanning protocol A GE multi-slice spiral CT scanner (Discovery CT 750 HD, GE Medical Systems, LLC, Waukesha, WI, USA) was used for portal vein CTA. Scanning was performed in the cranio-caudal direction, covering the area from 2 cm above the diaphragm to the lower edge of both kidneys. The following scanning parameters were employed: pitch, 0.984:1; rotation speed, 0.5 s/r; and adaptive tube current modulation technology (based on a fixed noise index of 8). Patients were divided into the following two groups based on tube voltage: group A1 (100 kVp) and group A2 (120 kVp). Contrast enhancement was performed using a threshold-triggered multiphase scanning method. The region of interest (ROI) was set at the abdominal aorta at the level of the porta hepatis, with a threshold of 120 HU. All patients underwent a multiphasic scan, including an unenhanced phase, arterial phase, portal venous phase, and venous phase. When the threshold was reached, diagnostic scans were acquired after an additional delay of 8 s on average for the arterial phase, 20–25 s for the portal venous phase, and 45–50 s for the venous phase. CM injection protocol All patients received a CM injection via the median cubital vein of the right hand. The CM used was iomeprol (400 mgI/mL). All patients received the injection using an identical high-pressure syringe (Stellant D Injector, Bayer Medical Care Inc., Indianola, PA, USA), the same indwelling needle (Introcan Safety 3 IV Catheter, B. Braun Melsungen AG, Melsungen, Germany; catheter gauge: 20-G), and matching disposable high-pressure syringe accessories (CT-200 High-Pressure Angiographic Syringes and Accessories, AnderMed, Zibo, China). The maximum injection pressure was uniformly set at 300 psi. Harms were defined as any adverse event (AE) related to the contrast injection procedure (e.g., extravasation and allergic-like reactions). Each patient was directly observed during and for 30 minutes after the injection, and any event was recorded on a predefined AE form. The total CM volume was calculated based on the patient’s TBW. The CM volume was calculated using the following formula: Patients were divided into two groups based on the body weight coefficient: group B1 (400 mgI/kg) and group B2 (450 mgI/kg). The flow rate was determined using the following formula: Patients were divided into two groups based on FID: group C1 (18 s) and group C2 (20 s). Following the injection of CM, a 10-s saline flush was administered [saline volume (mL) = flow rate (mL/s) × 10 (s)]. The flow rate for this saline injection was maintained at the same rate as that of the CM. Tube voltage, TBW-adjusted TID, and FID were arranged into a 2×2×2 factorial design (group 1: 100 kVp-400 mgI/kg-18 s, group 2: 100 kVp-400 mgI/kg-20 s, group 3: 100 kVp-450 mgI/kg-18 s, group 4: 100 kVp-450 mgI/kg-20 s, group 5: 120 kVp-400 mgI/kg-18 s, group 6: 120 kVp-400 mgI/kg-20 s, group 7: 120 kVp-450 mgI/kg-18 s, and group 8: 120 kVp-450 mgI/kg-20 s). Image analysis The primary outcome was the CTmpv. The secondary outcomes included objective parameters [CT values of the left superior branch of the portal vein (CTlspv) and liver parenchyma (CTliver), image noise, contrast-to-noise ratio (CNR), volume CT dose index (CTDIvol), and dose-length product (DLP)], and subjective qualitative image quality assessments. Axial slices were reconstructed with a thickness of 1.25 mm and an interval of 1.25 mm, along with a thickness of 5 mm and an interval of 5 mm. Both the 1.25- and 5-mm images were reconstructed using identical parameters: filtered back projection with a standard algorithm, a standard-plus kernel, and 50% Adaptive Statistical Iterative Reconstruction-V (ASIR-V; GE Healthcare, Waukesha, WI, USA). The 5-mm portal venous phase images were imported into a GE ADW4.11 workstation. Display and window settings were standardized across all groups, with a window width of 300–360 HU and a window level of 30–40 HU for soft tissue. Two radiologists, each with ≥5 years of experience, measured the relevant parameters using a double-blind method. They were completely blinded to the specific scanning and injection parameters, patient group assignments, and general patient characteristics, including height, weight, age, and sex. The average of three measurements was used as the final value. The following parameters were assessed: (I) CTmpv and CTlspv: CTmpv was measured using a circular ROI with an area of approximately 0.5 cm2 at the level of the porta hepatis. The SD of CT values for the erector spinae muscle (SDmuscle) at the same slice was also recorded as image noise. Additionally, the CT value of the intrahepatic portal vein branch, specifically the more prominent left superior branch of the portal vein, was measured (ROI area: 0.1±0.02 cm2). The measurement ROI was located at the center of the lumen and occupied two-thirds of the cross-sectional area of the lumen (Figure 1). (II) CTliver: the CTliver was measured while avoiding blood vessels and lesions (ROI area: 0.5±0.05 cm2) (Figure 1). (III) CNR: the CNR of the portal vein was calculated as follows: CNR = (CTmpv − CTliver)/SDmuscle. A high CNR between the portal vein and surrounding liver parenchyma improves the detection of lesions, including anatomical variations, stenosis, and embolism of the portal vein. Improving CNR is crucial for hepatic vascular imaging because it directly enhances the visualization of fine vascular branches (3).(IV) CTDIvol and DLP values: the CTDIvol and DLP values of the portal venous phase were recorded from the radiation dose report. Quantitative parameter measurements. (A) Measurement of the CTmpv at the level of the porta hepatis. (B) Measurement of the CTlspv. (C) Measurement of the SDmuscle at the same slice level. (D) Measurement of the CTliver while avoiding blood vessels and lesions. CT, computed tomography; CTliver, CT value of the liver parenchyma; CTlspv, CT value of the left superior branch of the portal vein; CTmpv, CT value of the main portal vein; SD, standard deviation; SDmuscle, SD of CT value of the erector spinae muscle. The portal venous phase 1.25-mm images were imported into a GE ADW4.11 workstation and underwent post-processing for maximum intensity projection (MIP) and volume rendering (VR). Two radiologists, each with ≥5 years of experience in abdominal CT diagnosis, scored the images from each group in a blinded manner across five aspects: overall image quality, focal lesion conspicuity in the liver, portal vein lesion detection, portal vein VR image quality, and MIP image quality. The scoring details are presented in Table 1. Table 1Scoring criteria for qualitative analysisScoreCriteriaOverall image qualityFocal lesion conspicuityPortal vein lesion detection3D-VR/3D-MIP images0Marked image noise, poorly visualized vascular contrast and branches, non-diagnosticLesion completely imperceptible; impossible to characterizeVascular structure completely unrecognizable; no visible lumen contour or courseIndistinct vascular borders; only the main portal vein is visible with poor contrast1Marked image noise, relatively poorly visualized vascular contrast and branches, moderately impacting diagnosisLesion faintly visible, but difficult to distinguish from artifacts or normal structuresOnly localized main trunk visible; branches and suspected abnormal areas cannot be assessedRelatively indistinct vascular borders; only first-level branches are visible with poor contrast2Moderate image noise, averagely visualized vascular contrast and branches, mildly impacting diagnosisLesion identifiable, edges poorly defined; approximate location and extent can be determinedMain trunk and major branches visible, but luminal details (e.g., mild stenosis and mural thrombus) are unclearRelatively clear vascular borders; second-level branches are visible with moderate contrast3Mild image noise, well-visualized vascular contrast and branches, not affecting diagnosisLesion clearly visible, margins relatively distinct; preliminary assessment of nature (e.g., cystic/solid) possibleGood visualization of entire vessel course; abnormal signs (e.g., caliber change and filling defect) can be identifiedWell-defined vascular borders; third-level branches are clearly visible with excellent contrast4Excellent image quality, excellently visualized vascular contrast and branchesLesion highly conspicuous, sharp margins; internal features (e.g., enhancement pattern and necrosis) can be accurately evaluatedExcellent depiction of vascular tree; thrombosis, stenosis, and subtle pathology can be precisely assessedSharp vascular borders; fourth-level and higher branches are visible with uniform enhancement and excellent contrast3D, three-dimensional; MIP, maximum intensity projection; VR, volume rendering. Scoring criteria for qualitative analysis 3D, three-dimensional; MIP, maximum intensity projection; VR, volume rendering. Statistical analysis The statistical analysis was performed using SPSS version 26.0. Measurement data are expressed as mean ± SD. A P value <0.05 was considered statistically significant. One-way analysis of variance (ANOVA) was used to compare baseline characteristics, including height, weight, and age, among multiple patient groups. Additionally, three-way factorial ANOVA was applied to compare quantitative parameters, including portal vein CT values, SD, CNR, DLP, and CTDIvol among groups and to test the main effects of tube voltage, TID, and FID, as well as all two-way and three-way interactions. Post hoc comparisons for significant interactions were conducted using Tukey’s honestly significant difference test to control the familywise error rate. Comparisons of subjective scores among multiple groups of images were performed using the Aligned Rank Transform (ART) for three-way ANOVA. The intraclass correlation coefficient (ICC) was used to assess consistency between the two observers, with ICC values ≥0.75 indicating good consistency, values of 0.4–0.75 indicating moderate consistency, and values ≤0.4 indicating poor consistency. The consistency of scores between the two physicians was evaluated using the kappa test, with kappa values ≥0.75 indicating good consistency, values of 0.4–0.75 indicating moderate consistency, and values ≤0.4 indicating poor consistency. Results Patient characteristics Among the initial cohort of 227 patients, five were excluded due to significant respiratory motion artifacts compromising measurement accuracy, resulting in a final study population of 222 patients. The participant enrollment process is summarized in Figure 2. All injections were performed as per protocol by the same team of trained radiographers. No deviations were noted, and all patients adhered to the protocol. No injection-related AEs were observed in any of the included patients. The enrolled patients were categorized into the following disease categories: hepatobiliary primary space-occupying lesions (n=76; benign: 23, malignant: 42, unspecified: 11); suspected hepatic metastases from extrahepatic primary malignancies (n=94; confirmed cases: 24, unconfirmed cases: 13, no metastasis: 57); non-space-occupying hepatic pathologies, including hepatitis and hepatic steatosis (n=24); and other abdominal conditions (e.g., cholecystitis, pancreatitis, and abdominal pain) (n=28). Among these, cirrhosis (n=32), portal hypertension (n=28), and ascites (n=9) were identified. The study population comprised 105 male and 117 female participants, with a mean age of 58.89±12.84 years, a height of 162.12±7.84 cm, a body weight of 67.07±9.03 kg, and a body mass index (BMI) of 25.52±2.97 kg/m2. No statistically significant differences were observed in age, height, or weight among the eight patient groups, as summarized in Table 2. Flowchart of participant inclusion and exclusion. Tube voltage, TBW-adjusted TID, and FID were arranged into a 2×2×2 factorial design (group 1: 100 kVp-400 mgI/kg-18 s, group 2: 100 kVp-400 mgI/kg-20 s, group 3: 100 kVp-450 mgI/kg-18 s, group 4: 100 kVp-450 mgI/kg-20 s, group 5: 120 kVp-400 mgI/kg-18 s, group 6: 120 kVp-400 mgI/kg-20 s, group 7: 120 kVp-450 mgI/kg-18 s, and group 8: 120 kVp-450 mgI/kg-20 s). FID, fixed injection duration; TBW, total body weight; TID, total iodine dose. Table 2General information of patientsParametersGroup 1Group 2Group 3Group 4Group 5Group 6Group 7Group 8FPNumber2929322525262729Height (cm)160.79±9.49160.48±5.79161.38±6.93161.88±7.04164.00±6.06168.53±5.91164.16±11.83161.153±8.821.9470.065Weight (kg)69.32±8.4569.48±6.7164.16±7.9766.72±9.1269.06±8.7669.00±8.4862.50±9.8466.69±11.651.8330.084Age (years)60.62±11.3259.03±12.3658.57±13.0560.02±14.954.73±8.3058.46±11.0556.33±18.4161.30±12.770.4560.865Data are presented as number or mean ± SD. Tube voltage, TBW-adjusted TID, and FID were arranged into a 2×2×2 factorial design (group 1: 100 kVp-400 mgI/kg-18 s, group 2: 100 kVp-400 mgI/kg-20 s, group 3: 100 kVp-450 mgI/kg-18 s, group 4: 100 kVp-450 mgI/kg-20 s, group 5: 120 kVp-400 mgI/kg-18 s, group 6: 120 kVp-400 mgI/kg-20 s, group 7: 120 kVp-450 mgI/kg-18 s, and group 8: 120 kVp-450 mgI/kg-20 s). FID, fixed injection duration; SD, standard deviation; TBW, total body weight; TID, total iodine dose. General information of patients Data are presented as number or mean ± SD. Tube voltage, TBW-adjusted TID, and FID were arranged into a 2×2×2 factorial design (group 1: 100 kVp-400 mgI/kg-18 s, group 2: 100 kVp-400 mgI/kg-20 s, group 3: 100 kVp-450 mgI/kg-18 s, group 4: 100 kVp-450 mgI/kg-20 s, group 5: 120 kVp-400 mgI/kg-18 s, group 6: 120 kVp-400 mgI/kg-20 s, group 7: 120 kVp-450 mgI/kg-18 s, and group 8: 120 kVp-450 mgI/kg-20 s). FID, fixed injection duration; SD, standard deviation; TBW, total body weight; TID, total iodine dose. Quantitative results The measurements of CTmpv, CTlspv, CTliver, and SDmuscle showed good consistency between the two measurers, with ICC values of 0.820, 0.806, 0.845, and 0.793, respectively. Summary statistics for the quantitative data are presented in Table 3 and Figure 3, while Table 4 presents the three-factor ANOVA results, including the main effects and interactions among the three factors. Table 3Results of the quantitative analysisParametersGroup 1Group 2Group 3Group 4Group 5Group 6Group 7Group 8CTmpv (HU)224.48±16.94235.48±27.19224.85±36.21236.84±34.23186.53±34.88194.92±22.79191.98±28.84193.12±25.80CTlspv (HU)194.68±33.25214.17±33.32199.23±33.65212.96±29.46177.00±17.60178.76±24.52183.33±31.83182.41±19.45CTliver (HU)112±18.34113.89±16.04115.66±18.32113.42±16.31106.00±10.9797.71±16.46105.05±12.65101.99±14.25SDmuscle7.51±1.207.70±1.106.77±1.496.89±1.596.34±1.116.16±1.156.36±0.875.83±1.57CNR15.94±3.9016.17±5.1517.12±7.1518.85±6.7313.12±4.1816.27±6.8413.72±3.5516.85±6.41CTDIvol (mGy)12.70±2.2711.94±1.9411.88±1.7712.62±2.4312.89±2.4713.04±2.8413.49±3.1213.61±3.34DLP (mGy·cm)296.74±32.51274.82±29.01288.49±30.87274.73±30.17327.61±43.19320.91±40.28321.11±32.21312.90±39.20Data are presented as mean ± SD. Tube voltage, TBW-adjusted TID, and FID were arranged into a 2×2×2 factorial design (group 1: 100 kVp-400 mgI/kg-18 s, group 2: 100 kVp-400 mgI/kg-20 s, group 3: 100 kVp-450 mgI/kg-18 s, group 4: 100 kVp-450 mgI/kg-20 s, group 5: 120 kVp-400 mgI/kg-18 s, group 6: 120 kVp-400 mgI/kg-20 s, group 7: 120 kVp-450 mgI/kg-18 s, and group 8: 120 kVp-450 mgI/kg-20 s). CNR, contrast-to-noise ratio (of the main portal vein); CT, computed tomography; CTDIvol, volume CT dose index; CTliver, CT value of the liver parenchyma (avoiding blood vessels and lesions); CTlspv, CT value of the left superior branch of the portal vein; CTmpv, CT value of main portal vein; DLP, dose-length product; FID, fixed injection duration; HU, Hounsfield units; SD, standard deviation; SDmuscle, SD of the CT value of the erector spinae muscle; TBW, total body weight; TID, total iodine dose. Results of the quantitative analysis Data are presented as mean ± SD. Tube voltage, TBW-adjusted TID, and FID were arranged into a 2×2×2 factorial design (group 1: 100 kVp-400 mgI/kg-18 s, group 2: 100 kVp-400 mgI/kg-20 s, group 3: 100 kVp-450 mgI/kg-18 s, group 4: 100 kVp-450 mgI/kg-20 s, group 5: 120 kVp-400 mgI/kg-18 s, group 6: 120 kVp-400 mgI/kg-20 s, group 7: 120 kVp-450 mgI/kg-18 s, and group 8: 120 kVp-450 mgI/kg-20 s). CNR, contrast-to-noise ratio (of the main portal vein); CT, computed tomography; CTDIvol, volume CT dose index; CTliver, CT value of the liver parenchyma (avoiding blood vessels and lesions); CTlspv, CT value of the left superior branch of the portal vein; CTmpv, CT value of main portal vein; DLP, dose-length product; FID, fixed injection duration; HU, Hounsfield units; SD, standard deviation; SDmuscle, SD of the CT value of the erector spinae muscle; TBW, total body weight; TID, total iodine dose. Box plots of quantitative image quality parameters and radiation dose. (A) CTmpv. The black line indicates the reference threshold of 220 HU. The CT values were significantly higher in groups 1–4 (100 kVp) than in groups 5–8 (120 kVp). (B) CTlspv. The red line indicates the reference threshold of 200 HU. The CT values were significantly higher in groups 1–4 (100 kVp) than in groups 5–8 (120 kVp). (C) CTliver. (D) DLP. DLP values were significantly lower in groups 1–4 (100 kVp) than in groups 5–8 (120 kVp). (E) Image noise, represented by the SDmuscle. (F) CNR. CNR, contrast-to-noise ratio; CT, computed tomography; CTliver, CT value of the liver parenchyma; CTlspv, CT value of the left superior branch of the portal vein; CTmpv, CT value of the main portal vein; DLP, dose-length product; HU, Hounsfield units; SD, standard deviation; SDmuscle, SD of CT value of the erector spinae muscle. Table 4Results of a three-way factorial ANOVA of quantitative analysisParameterskVpTIDFIDkVp × TIDkVp × FIDTID × FIDkVp × TID × FIDFPFPFPFPFPFPFPCTmpv50.38<0.001*0.5660.4534.3250.039*0.1560.6931.3160.2530.6330.4270.0070.931CTlspv24.634<0.001*0.4410.5082.8830.0910.110.7412.6020.1090.1770.6740.0230.878CTliver16.575<0.001*0.3610.5491.1610.283<0.010.9891.0260.3130.010.9210.7430.39SDmuscle22.811<0.001*3.870.0540.2140.6452.0080.1581.3270.2510.2470.620.1130.737CNR2.5060.1153.3160.077.0450.009*1.4750.2260.3530.5530.0190.8910.0150.903DLP44.842<0.001*1.0710.3021.6300.2030.0780.780.8870.3480.0910.7630.1920.662CTDIvol30.38<0.001*2.7510.0990.0110.9171.4560.2293.5190.0620.0250.8750.9580.329*, P<0.05. kVp × TID represents the two-way interaction between kVp and TID; kVp × FID represents the two-way interaction between kVp and FID; TID × FID represents the second-order interaction between FID and TID; kVp × TID × FID represents the three-way interaction among kVp, FID, and TID. ANOVA, analysis of variance; CNR, contrast-to-noise ratio (of the main portal vein); CT, computed tomography; CTDIvol, volume CT dose index; CTliver, CT value of the liver parenchyma (avoiding blood vessels and lesions); CTlspv, CT value of the left superior branch of the portal vein; CTmpv, CT value of main portal vein; DLP, dose-length product; FID, fixed injection duration; kVp, tube voltage; SD, standard deviation; SDmuscle, SD of CT value of the erector spinae muscle; TID, total iodine dose. Results of a three-way factorial ANOVA of quantitative analysis *, P<0.05. kVp × TID represents the two-way interaction between kVp and TID; kVp × FID represents the two-way interaction between kVp and FID; TID × FID represents the second-order interaction between FID and TID; kVp × TID × FID represents the three-way interaction among kVp, FID, and TID. ANOVA, analysis of variance; CNR, contrast-to-noise ratio (of the main portal vein); CT, computed tomography; CTDIvol, volume CT dose index; CTliver, CT value of the liver parenchyma (avoiding blood vessels and lesions); CTlspv, CT value of the left superior branch of the portal vein; CTmpv, CT value of main portal vein; DLP, dose-length product; FID, fixed injection duration; kVp, tube voltage; SD, standard deviation; SDmuscle, SD of CT value of the erector spinae muscle; TID, total iodine dose. For CTmpv, significant main effects were observed for tube voltage (A1 vs. A2: 229.33±34.71 vs. 192.58±21.19 HU; partial η2=0.229) and FID (C1 vs. C2: 211.89±32.90 vs. 222.43±37.39 HU; partial η2=0.025), whereas the main effect of TID and all interactions were non-significant (P>0.05). CTmpv was higher in the 100 kVp group than in the 120 kVp group, and CTmpv was also higher in the 20 s FID group than in the 18 s group. For CTlspv, the main effect of tube voltage was significant (A1 vs. A2: 204.39±33.34 vs. 180.19±23.04 HU; partial η2=0.127). CTlspv was higher in the 100 kVp group than in the 120 kVp group. However, the main effects of FID and TID, as well as all interactions, were non-significant (P>0.05). For CTliver, the main effect of tube voltage was significant (A1 vs. A2: 113.96±17.28 vs. 103.46±13.32 HU; partial η2=0.089), while the main effects of FID and TID, as well as all interactions, were non-significant (P>0.05). CTliver was higher in the 100 kVp group than in the 120 kVp group. For SDmuscle, the main effect of tube voltage was significant (A1 vs. A2: 7.18±1.41 vs. 6.18±1.19; partial η2=0.118). SDmuscle was higher in the 100 kVp group than in the 120 kVp group. The main effects of FID and TID, as well as all interactions, were non-significant (P>0.05). For CNR, the main effect of FID was significant (C1 vs. C2: 15.15±5.74 vs. 17.14±6.16; partial η2=0.04). CNR was higher in the 20 s FID group than in the 18 s group. However, the main effects of tube voltage and TID, as well as all interactions, were non-significant (P>0.05). Significant main effects of tube voltage were observed for both DLP (A1 vs. A2: 284.48±31.65 vs. 320.89±38.46, partial η2=0.209) and CTDIvol (A1 vs. A2: 12.27±1.89 vs. 13.15±2.43, partial η2=0.152). Specifically, both DLP and CTDIvol were significantly lower in the 100 kVp group than in the 120 kVp group. All interactions were non-significant (P>0.05). Subjective evaluation results The scores of overall image quality, focal lesion conspicuity in the liver, portal vein lesion detection, portal vein VR image quality, and MIP image quality between the two observers showed good consistency, with kappa values of 0.928, 0.892, 0.806, 0.870, and 0.845, respectively. Subjective image quality scores are summarized in Table 5. For overall image quality (tube voltage P=0.816, TID P=0.86, FID P=0.743), liver lesion detectability (tube voltage P=0.072, TID P=0.556, FID P=0.464), VR image quality (tube voltage P=0.455, TID P=0.303, FID P=0.163), and MIP image quality (tube voltage P=0.362, TID P=0.829, FID P=0.173), the main effects of tube voltage, TID, and FID were not significant, and none of the interactions were statistically significant. The image quality observed in the 100 kVp + 20 s + 400 mgI/kg group was also satisfactory (Figure 4A-4F). For portal vein lesion detectability, the main effect of tube voltage was significant (A1 vs. A2: 3.52±0.70 vs. 3.19±0.81, P=0.004, partial η2=0.048), while the main effects of TID (P=0.783) and FID (P=0.418) were not significant, and none of the interactions were statistically significant. Table 5Results of image quality scoresParametersGroup 1Group 2Group 3Group 4Group 5Group 6Group 7Group 8Overall image quality3.93±0.263.90±0.313.86±0.353.88±0.333.93±0.263.85±0.383.83±0.393.89±0.31Focal lesion conspicuity in the liver3.51±0.693.72±0.593.5±0.743.44±0.583.40±0.513.31±0.853.58±0.673.23±0.59Portal vein lesion detection3.51±0.693.68±0.603.47±0.773.44±0.713.26±0.703.08±0.863.33±0.983.08±0.763D-VR images3.62±0.563.59±0.683.36±0.623.56±0.773.40±0.743.62±0.513.25±0.973.54±0.523D-MIP images3.69±0.543.62±0.733.55±0.633.64±0.703.40±0.743.69±0.483.42±0.793.69±0.48Data are presented as mean ± SD. Tube voltage, TBW-adjusted TID, and FID were arranged into a 2×2×2 factorial design (group 1: 100 kVp-400 mgI/kg-18 s, group 2: 100 kVp-400 mgI/kg-20 s, group 3: 100 kVp-450 mgI/kg-18 s, group 4: 100 kVp-450 mgI/kg-20 s, group 5: 120 kVp-400 mgI/kg-18 s, group 6: 120 kVp-400 mgI/kg-20 s, group 7: 120 kVp-450 mgI/kg-18 s, and group 8: 120 kVp-450 mgI/kg-20 s). 3D, three-dimensional; FID, fixed injection duration; MIP, maximum intensity projection; SD, standard deviation; TBW, total body weight; TID, total iodine dose; VR, volume rendering. Results of image quality scores Data are presented as mean ± SD. Tube voltage, TBW-adjusted TID, and FID were arranged into a 2×2×2 factorial design (group 1: 100 kVp-400 mgI/kg-18 s, group 2: 100 kVp-400 mgI/kg-20 s, group 3: 100 kVp-450 mgI/kg-18 s, group 4: 100 kVp-450 mgI/kg-20 s, group 5: 120 kVp-400 mgI/kg-18 s, group 6: 120 kVp-400 mgI/kg-20 s, group 7: 120 kVp-450 mgI/kg-18 s, and group 8: 120 kVp-450 mgI/kg-20 s). 3D, three-dimensional; FID, fixed injection duration; MIP, maximum intensity projection; SD, standard deviation; TBW, total body weight; TID, total iodine dose; VR, volume rendering. Comparison of portal venous imaging under different scan and injection parameters. (A-C) Images of a 63-year-old male with a hepatic mass obtained using a 100 kVp, 400 mgI/kg, and 20-s FID protocol: (A) axial slab MIP image showing a hepatic mass with clear visibility; (B) 3D MIP image depicting homogeneous portal venous enhancement with fourth-order branch visualization; (C) 3D CR highlighting invaded branches. (D-F) Images of a 47-year-old female, obtained using a 120 kVp, 450 mgI/kg, and 20-s FID protocol: (D) slab MIP image demonstrating hepatic anatomy; (E) 3D MIP image showing adequate portal venous enhancement; (F) 3D VR image distinctly visualizing the portal vein. (G-I) Images of a patient with hepatocellular carcinoma and cirrhosis obtained post-treatment using a 100 kVp, 450 mgI/mL, and 18-s FID protocol: (G) axial image showing adequate portal vein enhancement with clear visualization of the thrombus (arrow) in the main portal trunk; (H) MPR image illustrating the longitudinal extent of the thrombus (arrow); (I) slab MIP image showing preserved distal portal vein branches (arrow). (J-L) Images from the same clinical scenario and injection protocol obtained at 120 kVp: (J) axial image showing suboptimal enhancement with poor thrombus demarcation (arrow); (K) MPR image providing limited visualization (white arrow pointing to the thrombus); (L) slab MIP image indicating inadequate enhancement of distal branches. 3D, three-dimensional; CR, cinematic rendering; FID, fixed injection duration; MIP, maximum intensity projection; MPR, multiplanar reconstruction. Discussion In current clinical practice, some patients undergoing routine contrast-enhanced abdominal CT scans still receive insufficient CM doses, resulting in low iodine concentration in target vessels and inadequate lesion enhancement, which may affect diagnosis and clinical evaluation. Conversely, excessive CM doses remain an issue in some patients (19). TBW-based CM dosing protocols have shown promise in improving abdominal enhancement and vascular image quality (20,21). In abdominal CT vascular enhancement imaging, the CM injection flow rate is also a key parameter affecting enhancement. Studies have found that FID protocols demonstrate superior vascular enhancement (13). Compared to fixed-rate protocols, FID protocols can more effectively adapt to individual differences, especially in patients with higher body weights. The present study systematically evaluated and compared portal vein and liver enhancement and image quality using a high-concentration CM (400 mgI/mL) at conventional tube voltages (100–120 kVp) with a personalized TID based on TBW combined with an FID protocol. The results showed that under the premise of using high-concentration iodine CM, this synergistic protocol achieved satisfactory enhancement: even at a tube voltage of 120 kVp, CTmpv exceeded 170 HU, with a satisfaction rate as high as 82%. Further, the 400 mgI/kg protocol at 100 kVp reduced the CM dose without affecting portal vein enhancement, outperforming the corresponding 120 kVp protocol and achieving comparable performance to the 450 mgI/kg protocol at 100 kVp. Compared to the recommended 520 mgI/kg (22), it reduced the dose by approximately 23.1%. This study found that under the same body weight coefficient and FID, reducing the tube voltage significantly increased the CTmpv and its branches while reducing the radiation dose, without compromising subjective image quality scores and CNR values. This phenomenon can be explained by the physical mechanism of CT imaging: lower tube voltage brings the X-ray photon energy closer to the K-edge absorption peak of iodine (33.2 keV), enhancing the photoelectric effect of the CM, thereby increasing the attenuation coefficient of the intravascular iodine CM and significantly enhancing the CT values of blood vessels (23,24). Additionally, lowering the tube voltage reduces the proportion of high-energy photons, thereby decreasing radiation scatter and noise and improving the signal-to-noise ratio of the image, maintaining better image quality (25). Although lower tube voltage increases image noise (9), which may reduce soft tissue contrast, its impact on vascular assessment is limited due to the high intrinsic attenuation of contrast-filled blood vessels (26,27). Consequently, the CNR of the portal vein did not decrease. Moreover, we found that in the assessment of portal vein lesion detectability, the 100 kVp group performed better than the 120 kVp group. This finding likely stems from the fact that reducing the tube voltage elevated the CT values of the portal vein, thereby enhancing the contrast between the low-density emboli and the surrounding lumen (Figure 4G-4L). Importantly, using 100 kVp substantially reduced the patient radiation dose, with DLP decreasing from 320 mGy·cm in the 120 kVp protocol to 280 mGy·cm in the 100 kVp protocol. However, excessively low tube voltage may lead to insufficient photons, increasing quantum noise and affecting image clarity, especially in patients with high body weights (28). This study selected 100 kVp as a balance point, considering both the enhancement and noise (29), aligning with the current trend of low-dose CT imaging technology. Moreover, clinically reducing the radiation dose while maintaining image quality may help minimize potential radiation risks to patients. This strategy is consistent with previous research findings on the use of low tube voltage to improve liver vascular imaging (30). Notably, the proportion of patients with high body weight in this cohort was relatively low (23% of patients had a body weight ≥75 kg). Previous studies have shown that for patients with a body weight exceeding 80–90 kg, low tube voltage (e.g., 80–100 kVp) may result in insufficient X-ray penetration, leading to increased image noise and reduced signal-to-noise ratio, which could potentially attenuate their benefits compared with conventional 120 kVp protocols (31). Therefore, the benefits of low-kVp imaging strategies observed in this study may be limited to patients with normal body weight and should not be extrapolated to those with high body weight (especially those with a BMI >30 kg/m2). Under consistent tube voltage and FID settings, a reduction in the body weight coefficient—corresponding to a lower TID—did not result in decreased portal vein CT values, liver enhancement amplitude, or contrast enhancement of the portal vein and liver. This finding is consistent with the results reported by Li et al., who similarly noted no statistically significant difference in portal vein enhancement between groups receiving 400 and 450 mgI/mL at a tube voltage of 100 kVp (6). The three-dimensional (3D) reconstructed images of the portal vein were of good quality, indicating that based on TBW-based calculations, a total iodine amount of 400 mgI/kg is sufficient for portal vein imaging and liver enhancement assessment. The enhancement of portal vein branches at all levels was uniform, and the anatomical structure of the portal vein was clearly displayed on 3D-VR and MIP images. This result may be related to the saturation effect of intravascular iodine concentration and hemodynamic regulatory mechanisms. Specifically, the enhancement effect of iodine CM depends not only on its absolute concentration but also on the iodine injection rate, blood flow velocity, and vascular volume (5). When the total iodine amount reaches a certain threshold, the iodine concentration in the portal vein tends to saturate, and further increases in iodine dose result in diminishing improvements in CT values. Further, individualized body weight-adjusted iodine dosage can more accurately match patients’ blood volume and circulatory status, avoiding the renal burden and potential toxicity risks associated with excessive iodine. In this study, stratified body weight-based iodine dosing showed the feasibility of reducing the iodine dose to 400 mgI/kg while maintaining imaging quality (Figure 5). The CTliver and CTmpv obtained under the 100 kVp and 450 mgI/kg protocol were also consistent with those reported by Du et al. (32). Images of a 59-year-old male with a lesion in S6, which was pathologically diagnosed as cholangiocarcinoma. Images were acquired using scanning and injection parameters of 100 kVp, 400 mgI/kg, and 20 s. (A) Images showing the space-occupying lesion (arrow) in liver segment 6. (B) Measurement of the CTmpv, which was 211 HU. (C) Slab MIP image of the portal vein showing homogeneous enhancement of the portal vein branches (arrow). (D) Volume-rendered reconstruction clearly showing the portal venous anatomy. (E) Holographic visualization of the portal vein and mass, clearly revealing that the mass was adjacent to the right branch of the portal vein. (F) Magnified holographic view highlighting the precise spatial relationship and adjacency of the tumor to the portal vein branches. CT, computed tomography; CTmpv, CT value of the main portal vein; HU, Hounsfield units; MIP, maximum intensity projection. Under the same tube voltage and TID conditions, increasing the FID did not reduce the CT values of the distal portal vein branches or the degree of liver enhancement during the portal venous phase. However, increasing the FID was associated with increased CTmpv. Theoretically, the key factor determining the degree of portal vein enhancement is the amount of iodine flowing through the portal vein per unit time and per unit volume (17). This finding may be attributed to the fact that a longer injection duration extends the time window for contrast agent inflow, facilitating more complete opacification of the main portal vein. This process reduces mixing heterogeneity between contrast-enhanced and unenhanced blood, thereby minimizing laminar flow artifacts (pseudo-filling defects) that typically decrease CT values. This is consistent with previous studies showing that a high injection rate results in rapid transit of contrast agent through the portal vein system, while a lower injection rate allows prolonged contrast residence in the portal vein (33). Additionally, due to the small diameter and low blood volume of distal portal vein branches, increasing the FID to 20 s did not significantly alter the iodine flux rate in these branches, and their CT values were less sensitive to changes in iodine concentration within the vein compared with those of the main portal vein. Therefore, increasing the FID to 20 s not only improves the CTmpv and CNR of the main portal vein without compromising the enhancement of distal branches or the liver, but also enables a lower injection rate, thereby reducing the risk of injection-related AEs and patient discomfort. For example, the average CM injection rate was 3.42 mL/s in group 1 and 2.99 mL/s in group 2, representing a reduction of approximately 12.5%. Sun et al. showed that in patients with limited venous access, the use of high-concentration CM (e.g., 400 mgI/mL) while maintaining the same iodine delivery rate (1.12 gI/s) significantly alleviated pain and discomfort during injection while further improving image quality (34), which is consistent with our findings. Therefore, individualized adjustment of the injection rate to balance enhancement performance and patient tolerance is a key step in achieving high-quality portal vein imaging. This study had several limitations. First, it was a single-center, single-machine study; thus, larger-scale studies are needed to confirm whether the protocol combining a tube voltage of 100 kVp, a body weight coefficient of 400 mgI/kg, and a 20-s FID is suitable for portal vein examinations on different CT scanner models. Second, the proportion of patients with high body weight in this cohort was relatively low; thus, future studies should include a larger proportion of patients with high body weight to validate the efficacy and safety of low‑kVp imaging in this subgroup. Further, this study included patients with different diseases; however, the proportion of patients with cirrhosis or significantly altered portal system hemodynamics was limited. Further research needs to be conducted to determine the optimal CM protocol for these specific populations. Conclusions The use of high-concentration CM (400 mgI/mL), a personalized TID protocol based on TBW, and the FID protocol all achieved good portal vein visualization. In particular, employing 100 kVp instead of 120 kVp increased portal vein CT values, potentially improving the detectability of portal vein lesions, and a 20-s FID enhanced the CNR of the main portal vein. Thus, the combined protocol of 100 kVp, a TID of 400 mgI/kg, and a 20-s FID maintained liver image quality while enhancing portal vein imaging, and simultaneously reducing radiation dose and contrast agent usage. This protocol represents a promising approach for clinical CT portal venography. Supplementary The article’s supplementary files as