Diagnostic performance and dose reduction of virtual noncontrast imaging in pediatric computed tomography urography

Introduction

Computed tomography (CT) plays a crucial role in the diagnosis of pediatric urinary tract diseases, including congenital anomalies, tumors, hydronephrosis, and calculi (1,2). However, because children undergo rapid growth and active cellular proliferation, they are highly sensitive to ionizing radiation. Consequently, adhering to the as-low-as-reasonably-achievable (ALARA) principle without compromising diagnostic image quality remains a key challenge in pediatric imaging. Conventional CT urography (CTU) includes a four-phase protocol (unenhanced, corticomedullary, nephrographic, and excretory phases). Although this approach provides comprehensive structural and functional evaluation of the entire urinary tract—including the renal calyces, pelvis, ureters, and bladder—it requires a substantial cumulative radiation dose.

Dual-source dual-energy CT (DECT) represents an advancement in conventional single-energy CT, as it involves the simultaneous acquisition of two distinct X-ray energy spectra. Its primary strength is advanced material decomposition, allowing it to conduct iodine subtraction from contrast-enhanced images to generate virtual noncontrast (VNC) images. This capability may facilitate the omission of true noncontract (TNC) acquisitions, thereby reducing radiation exposure (3). Additionally, DECT provides unique diagnostic value in tissue characterization, calculus composition analysis, and vascular imaging (4-6).

Although the applications of VNC are well-validated in abdominal imaging, studies on this modality have predominantly focused on isolated lesion types or adult populations. Comprehensive research addressing the full spectrum of pediatric urinary diseases (e.g., tumors, hydronephrosis, malformations, and trauma) remains limited (7,8). We therefore conducted a retrospective study with the aim of comparing the diagnostic utility of VNC images derived from the corticomedullary and nephrographic phases with that of TNC images across various tissues and pathologies. By evaluating CT attenuation differences in normal renal parenchyma and conditions including tumors, hydronephrosis, cysts, calculi, trauma, and malformations, we assessed the clinical feasibility of substituting TNC with VNC in pediatric CTU. Ultimately, the findings of this study may provide an evidence base for minimizing cumulative radiation exposure in pediatric patients. We present this article in accordance with the GRRAS reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0695/rc).

Methods

Study population

This retrospective, single-center, intraindividual paired comparative study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. It was approved by the Institutional Review Board of Wuhan Children’s Hospital (Wuhan Maternal and Child Healthcare Hospital), Tongji Medical College, Huazhong University of Science & Technology (approval No. 2026R011-E01), and written informed consent was obtained from the legal guardians of all pediatric patients. We reviewed 217 consecutive pediatric patients who underwent four-phase dual-source CTU at our institution between June 2022 and December 2024, with each patient serving as his or her own control. Indications for the examination included structural urinary tract abnormalities, traumatic renal injuries, and renal masses. The cohort comprised 161 males and 56 females, with a mean age of 7.07±4.50 years (range, 2 days–17 years) and a mean weight of 26.42±14.93 kg (range, 2.10–70.20 kg). The clinical diagnoses included 91 cases of hydronephrosis, 39 renal cysts, 21 renal malignancies, 9 duplex collecting systems, 8 renal lacerations/contusions, 7 ectopic kidneys, and 6 renal calculi. The 21 renal malignancies included 19 Wilms tumors, 1 clear cell sarcoma, and 1 rhabdomyosarcoma, and patients with these malignancies had a mean age of 2.86±2.18 years (range, 4 months–7 years). Additionally, 36 patients had structurally normal kidneys but presented with other conditions, including inguinal hernia (n=4), cystitis (n=3), and urachal cyst (n=1).

All patients fasted prior to the examination. Sedation was administered to children who were unable to cooperate during the scan. The exclusion criteria were as follows: (I) severe motion artifacts compromising diagnostic image quality; (II) a solitary kidney, either due to congenital unilateral agenesis or prior nephrectomy; and (III) severe hepatic or renal dysfunction.

CT acquisition protocol

All examinations were performed with a third-generation dual-source CT scanner (SOMATOM Force, Siemens Healthineers, Erlangen, Germany). The standard CTU protocol comprised four phases: unenhanced, corticomedullary, nephrographic, and excretory. The unenhanced scan ranged from the diaphragm to the inferior margin of the pubic symphysis. The contrast-enhanced phases typically ranged from the upper renal poles to the pubic symphysis, with field-of-view adjustments being tailored to the specific lesion extent. To mitigate motion artifacts and minimize radiation exposure, the unenhanced and excretory phases were acquired with a high-pitch spiral mode (Flash mode). The scan parameters were as follows: automatic tube voltage selection (CARE kV) with a reference voltage of 100 kV, automatic tube current modulation (CARE Dose4D) with a reference quality tube current-time product of 280 mAs, a detector collimation of 2 mm × 192 mm × 0.6 mm, and a gantry rotation time of 0.25 seconds. All images were reconstructed at a slice thickness of 1.0 mm and an increment of 0.7 mm via advanced modeled iterative reconstruction (ADMIRE) at a strength level of 3. The corticomedullary and nephrographic phases were acquired in dual-energy mode. The tube voltages were set at 80 kV and Sn150 kV (with tin filtration). The tube current was automatically modulated (CARE Dose4D) through use of dual-tube reference quality settings of 260 mAs for tube A (80 kVp) and 130 mAs for tube B (Sn150 kVp), and the gantry rotation time was set to 0.5 seconds. A nonionic contrast medium (iohexol; 300 mgI/mL) was administered via a peripheral vein with a dual-syringe power injector. The dose was weight-based at 1.5–2.0 mL/kg (maximum limit 80 mL) and injected at a rate of 0.3–3.5 mL/s. This was immediately followed by an injection of 6–30 mL of saline chaser at the identical rate. A bolus tracking technique was employed to time the corticomedullary phase. A region of interest (ROI) was placed within the abdominal aorta at the level of the renal hilum, with a predefined trigger threshold of 120 Hounsfield units (HU). The corticomedullary scan commenced with an 8-second delay after the threshold was reached. The nephrographic phase was acquired 30 seconds after the completion of the corticomedullary phase. The excretory phase was performed in routine fashion 5 minutes after the nephrographic phase; however, a 15-minute delay was applied for patients with impaired excretion, such as those with severe hydronephrosis or ureteral obstruction.

Image postprocessing and reconstruction

All thin-slice images from the corticomedullary and nephrographic phases were transferred to a dedicated postprocessing workstation (syngo.via version VB60A; Siemens Healthineers). Dual-energy analysis was performed via the “CT Dual Energy” application, and VNC images were generated via the standard “Liver VNC” profile. This profile was selected because it represents the vendor-recommended abdominal standard default optimized for multimaterial decomposition (iodine, fat, and soft tissue) across the entire abdomen and retroperitoneum. Although renal structures and cysts are primarily nonfatty, the kidneys are anatomically enveloped by perirenal fat, and the renal sinus contains significant adipose tissue. Compared with the standard “Virtual Unenhanced” models based on basic iodine-water-air material matrices, the three-material matrix of the “Liver VNC” profile explicitly accounts for fat, thereby effectively minimizing mathematical miscalculations and edge-subtraction artifacts at tissue-fat boundaries to ensure optimal data stability. This approach aligns with previously conducted studies on abdominal and urological dual-energy protocols in which the abdominal liver VNC profile was successfully employed for renal evaluations and CTU protocols (9,10). Subsequently, the generated VNC images and the original TNC images were concurrently loaded into the multi-modality reading workspace to facilitate the precise placement of ROIs (Figure 1).

Figure 1 A 3-year-old boy with a Wilms tumor on the left side. Images from the (A) corticomedullary phase, (B) nephrographic phase, and (C) TNC scan, along with VNC images reconstructed from the (D) corticomedullary and (E) nephrographic phases. AP, arterial phase; TNC, true noncontrast; VP, venous phase; VNC-AP, virtual noncontrast image reconstructed from the arterial phase; VNC-VP, virtual noncontrast image reconstructed from the venous phase.

Image quality assessment

Objective image quality evaluation: ROIs were manually drawn on the TNC images and subsequently copied and pasted onto the corresponding VNC image reconstructed from the arterial phase (VNC-AP) and VNC image reconstructed from the venous phase (VNC-VP) images to ensure identical size and anatomical positioning. Each ROI was measured three times, and the mean CT attenuation (HU) and standard deviation (SD) were recorded (11). In instances of automatic misregistration, the software’s synchronization function was disabled, and the ROIs were manually adjusted to the exact corresponding anatomical slice. Measurements were obtained from various targets, including renal masses, hydronephrosis, duplex renal parenchyma, renal lacerations/contusions, ectopic renal parenchyma, renal calculi, and normal renal parenchyma. Additionally, background reference tissues at the same slice level, specifically, the abdominal aorta, erector spinae muscle, and subcutaneous fat, were measured.

Criteria for data selection: to maintain statistical independence, only one primary lesion or target region per patient was included for analysis. The selection protocol was as follows: for unilateral abnormalities, the lesion was measured, and the contralateral unaffected kidney was assigned to a normal renal parenchyma group. For bilateral abnormalities, only the most clinically significant lesion was measured, and neither kidney was included in the normal group. For bilaterally normal kidneys, the left renal parenchyma was selected as the standard measurement. After these stringent criteria were applied, 183 of the 217 patients contributed data to the normal renal parenchyma group (Figure 2).

Figure 2 Flowchart illustrating the patient grouping protocol and the derivation of data from normal renal parenchyma and lesions. CTU, computed tomography urography.

ROIs were drawn as large as anatomically feasible, typically ranging from 0.5 to 1.0 cm² (areas <0.5 cm² were permitted for small lesions or thin layers of subcutaneous fat in infants). Although standard guidelines for adults recommend an ROI diameter >1 cm for absolute quantification (12), pediatric structures strictly preclude this. To guarantee high measurement reliability despite constraints on ROI size, each ROI was measured three times and averaged. Spatial identity across phases was strictly maintained with a synchronized copy-and-paste function. Careful attention was paid to strictly avoid areas of necrosis, tissue margins, calcifications, macroscopic blood vessels, and prominent image artifacts. For space-occupying lesions, the ROI was precisely positioned on the axial slice with the maximum cross-sectional area of the tumor. Background image noise was defined as the SD of the attenuation measured within the erector spinae muscle. The signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) were calculated according to the following equations:

SNR=|HUROI|SDmuscle[1]

CNR=|HUROI−HUmuscle|SDmuscle[2]

where HU_ROI is the mean CT attenuation of the target region, HU_muscle is the mean CT attenuation of the erector spinae muscle, and SD_muscle is the background image noise.

Subjective image quality evaluation: Subjective image quality was independently evaluated by two radiologists, each with over 5 years of diagnostic experience. The readers were blinded to the specific imaging protocols (VNC vs. TNC) and patient clinical data. Image quality was graded on a 5-point Likert scale according to the following scheme (13,14): score 5, excellent (no artifacts, minimal noise, with excellent delineation of anatomical structures and lesion details); score 4, good (mild artifacts or noise, but clear visualization of the anatomical structures and lesion details); score 3, fair (moderate artifacts or noise slightly affecting detail visibility, but with structures remaining identifiable and the image being adequate for diagnosis); score 2, poor (substantial artifacts or noise blurring anatomical structures and obscuring details, compromising diagnostic confidence); and score 1, unacceptable (severe artifacts or noise rendering anatomical structures unidentifiable and nondiagnostic). Images receiving a score of ≥3 were considered diagnostically acceptable. Any interreader discrepancies in the initial scoring were resolved through joint review until a consensus was reached.

Radiation dose estimation

The dose-length product (DLP) for the TNC phase and the cumulative DLP across all acquisition phases were obtained directly from the scanner’s automated dose report. The effective dose (ED) was calculated according to the following equation: ED=DLP×k, where k represents the age- and anatomy-specific (abdomen and pelvis) dose conversion coefficient [mSv/(mGy·cm)]. In accordance with established pediatric dosimetry guidelines (15,16), the k values were applied based on the patient’s age as follows: 0.049 for 0< age <1 year; 0.030 for 1≤ age <5 years; 0.020 for 5≤ age <10 years; and 0.015 for age ≥10 years.

Statistical analysis

Statistical analyses were performed with SPSS version 29.0 (IBM Corp., Armonk, NY, USA) and Python version 3.11 (Python Software Foundation, Wilmington, DE, USA). The normality of the paired differences for CT attenuation, SD, SNR, and CNR between the TNC and VNC images was first assessed. For normally distributed continuous variables, a paired Student t-test was used. If the assumption of normality was violated, the Wilcoxon signed-rank test was applied. To comprehensively evaluate the imaging methods, Bland-Altman analysis was employed to define the systematic bias and 95% limits of agreement (LoA) between VNC and TNC. Additionally, to overcome the LoA reflecting standalone method reliability, intraclass correlation coefficients (ICCs) were computed to quantify the absolute agreement and measurement accuracy of each reconstruction method independently. Interreader agreement for the subjective image quality scores was evaluated via the Cohen kappa statistic.

Results

Objective image quality evaluation

Tables 1,2 systematically present the CT attenuation values, SNR, CNR, and SD for major anatomical and pathological entities across conventional TNC, DECT-derived VNC-AP, and VNC-VP image sets within the study cohort.

Compared to images from TNC, those from both VNC phases demonstrated a slight, systematic underestimation of CT attenuation across all evaluated tissues, with the exceptions of subcutaneous fat and renal calculi (mean difference: <8.62 HU). The primary tissue metrics are summarized in Table 1, while the quantitative parameters for the renal calculi and small sample pathology subgroups are detailed separately in Table 2. This tight margin demonstrated good agreement and validated the quantitative accuracy of the VNC algorithm. Furthermore, VNC images yielded a higher mean CNR and lower image noise (8.49 and 8.44 HU for VNC-AP and VNC-VP, respectively) than did TNC images (mean SD: 9.09 HU). No significant differences were observed between the two VNC phases in terms of SNR, CNR, or SD (P>0.017), indicating comparable overall image quality.

For subcutaneous fat, CT attenuation was slightly higher with VNC than with TNC (mean bias: <2.82 HU). However, VNC significantly underestimated the attenuation of renal calculi (mean difference: ~330 HU). Nevertheless, both the SNR and CNR for renal calculi remained >50 on VNC images. This finding indicates that despite the limited quantitative accuracy in HU, the high inherent contrast with surrounding structures preserved the excellent visibility and detectability of the calculi (Figure 3).

Figure 3 A 4-year-old boy with renal calculus on the right side. Images correspond to the (A) TNC scan, (B) VNC-AP, and (C) VNC-VP. HU, Hounsfield units; SD, standard deviation; TNC, true noncontrast; VNC-AP, virtual noncontrast image reconstructed from the arterial phase; VNC-VP, virtual noncontrast image reconstructed from the venous phase.

Bland-Altman analysis was conducted separately for the two independent pairs (TNC vs. VNC-AP and TNC vs. VNC-VP) across subcohorts with n>10 being used to evaluate their individual agreement with the reference standard. The analysis revealed that VNC produced a consistent, minor systematic underestimation for nonadipose tissues (<8.62 HU) and a minor overestimation for adipose tissue (<2.82 HU) across both reconstruction phases. The multipanel presentation in Figure 4 provides a visual summary of how each VNC phase individually agrees with the TNC baseline within the same tissue context; however, it does not imply a direct statistical comparison between VNC-AP and VNC-VP or a superiority analysis of these two modes.

Figure 4 Bland-Altman plots evaluating the agreement of CT attenuation between TNC and VNC images across various anatomical structures and lesions. Subgroups of (A) renal malignancies (n=21), (B) renal cysts (n=39), (C) hydronephrosis (n=91), (D) the abdominal aorta (n=217), (E) subcutaneous fat (n=217), (F) normal renal parenchyma (n=183), and (G) the erector spinae muscle (n=217). The X-axis represents the mean CT attenuation (in HU) of the TNC and VNC images, and the Y-axis represents their difference. Each panel displays two independent comparisons side by side (left: TNC vs. VNC-AP; right: TNC vs. VNC-VP) to illustrate that both VNC phases independently satisfy clinical agreement criteria against the TNC reference scan. This layout is chosen for visual compilation and does not represent a direct statistical comparison of accuracy between VNC-AP and VNC-VP. CT, computed tomography; HU, Hounsfield units; LoA, limits of agreement; TNC, true noncontrast; VNC, virtual noncontrast; VNC-AP, virtual noncontrast image reconstructed from the arterial phase; VNC-VP, virtual noncontrast image reconstructed from the venous phase.

Table 3 summarizes the agreement analysis of CT attenuation. TNC demonstrated moderate-to-excellent agreement with both VNC phases. Moreover, the agreement between the two VNC phases (VNC-AP and VNC-VP) ranged from good to excellent.

Subjective image quality evaluation

As ordinal categorical variables, subjective image quality scores were expressed as medians with interquartile ranges (IQRs). Interreader agreement between the two radiologists was good across the TNC and both VNC phases (κ>0.6). Although the VNC-VP phase produced slightly higher scores than the VNC-AP phase, both VNC phases scored lower than the TNC phase. Nevertheless, with the exception of a single VNC-AP image that received a score of 2, all VNC images achieved a score of ≥3, successfully meeting diagnostic requirements (Table 4).

Comparison of radiation dose

Table 5 and Figure 5 present the radiation dose comparisons stratified by age group. The cumulative DLP for the standard four-phase protocol was 392.66±172.78 mGy·cm, yielding an ED of 7.98±1.71 mSv. In contrast, the cumulative DLP for the three contrast-enhanced phases alone was 296.25±129.79 mGy·cm, corresponding to an ED of 6.02±1.28 mSv. Consequently, substituting the TNC scan with the VNC images resulted in an overall radiation dose reduction of approximately 24.6%.

Figure 5 Boxplots comparing radiation dose distribution across different age groups. (A) DLP, (B) ED. Total DLP and total ED represent the cumulative values for all scanning phases (the standard four-phase protocol). CE-DLP and CE-ED represent the cumulative values for the three CE phases alone. CE, contrast-enhanced; CE-DLP, contrast-enhanced dose-length product; CE-ED, contrast-enhanced effective dose; DLP, dose-length product; ED, effective dose.

Discussion

The purpose of this study was to evaluate the feasibility of substituting TNC with VNC imaging across a spectrum of pediatric urinary diseases (including congenital malformations, tumors, hydronephrosis, calculi, and trauma). To achieve this, a standard four-phase protocol was used to ensure a direct, intraindividual comparison of image quality between VNC and TNC under identical anatomical conditions. Our results demonstrated that the image quality between the VNC-AP and VNC-VP phases differed only slightly, with the vast majority of cases fully meeting diagnostic requirements. Given the robust consistency in image quality and diagnostic performance between VNC and TNC, the dedicated TNC scan can be safely omitted in routine clinical practice, and only the contrast-enhanced phases can be retained. Through the adoption of this simplified protocol, the effective radiation dose was reduced from 7.98±1.71 to 6.02±1.28 mSv (a substantial reduction of 24.6%), representing significant dose optimization without a reduction in diagnostic confidence. For pediatric patients requiring longitudinal follow-up, this translates to a marked decrease in cumulative radiation exposure risk and is a finding that strongly aligns with the VNC advantages outlined by Gallo-Bernal et al. (19). Although the reduction percentage may vary depending on the specific scanning protocols, scanner hardware, and patient age distribution, our data unequivocally confirm that VNC is a highly effective tool for the optimization of dose in pediatric CT.

Image quality is a key metric for assessing the viability of VNC as a TNC surrogate. Our objective evaluations revealed that, with the exception of renal calculi, VNC images yielded higher CNR and lower SD compared to TNC images, indicating superior tissue contrast resolution. The CT attenuation of VNC images showed moderate-to-excellent agreement with TNC. Specifically, subcutaneous fat demonstrated slightly higher attenuation on VNC, whereas other anatomical regions exhibited a systematic, yet slight, underestimation (mean difference: <8.7 HU). These variations remain well within clinically acceptable limits and corroborate findings from previous adult abdominal studies (20,21). The slight overestimation of fat attenuation can be primarily attributed to the inherent characteristics of the DECT material decomposition algorithm. Due to its low effective atomic number, the spectral attenuation profile of adipose tissue differs significantly from that of water and soft tissues, leading to fitting deviations in the material decomposition model. Furthermore, residual trace iodine signals present within the interstitial spaces of adipose tissue postenhancement are challenging to subtract completely, resulting in a systematic overestimation of CT attenuation. Clinically, although the CT attenuation of renal calculi on VNC was significantly lower than it was on TNC (a difference of approximately 330 HU), their inherent high contrast with surrounding structures ensured that their detectability was not compromised. This suggests that radiologists should interpret calculus-related findings on VNC images in conjunction with the broader clinical context. These pediatric findings mirror established data in adults indicating that VNC can adequately replace TNC in renal mass evaluations (22) and abdominal CT angiography (23), demonstrating robust multiphase attenuation consistency (24). Subjective image quality scores further confirmed that VNC images can satisfy diagnostic requirements. Interreader agreement was good (kappa value =0. 67 for VNC-AP images and 0.62 for VNC-VP images), and nearly all VNC images scored ≥3. Notably, a single VNC-AP image received a suboptimal score of 2 (Figure 6). This case involved a 2-month-old infant who received contrast medium via a pedal vein injection. During the cortical (arterial) phase scan, the highly concentrated contrast bolus was peaking within the inferior vena cava (IVC) (Figure 6A). This extreme, undiluted iodine concentration, bypassing the cardiopulmonary circulation, exceeded the linear dynamic range of the DECT material decomposition algorithm, resulting in severely incomplete iodine subtraction in the IVC and iliac veins on the VNC-AP images (Figure 6C). Conversely, during the medullary (venous) phase, the contrast medium had been sufficiently diluted through the systemic circulation, which brought the iodine concentration back into the algorithm’s optimal working range, thus yielding excellent iodine subtraction on the VNC-VP images (Figure 6D). This finding prompts a crucial clinical caveat: for young infants receiving pedal vein contrast injections, delayed lower-extremity venous return can cause localized extreme iodine concentrations that severely degrade early-phase VNC image quality. In such scenarios, reconstructing VNC images from the later medullary phase (VNC-VP) is highly recommended to ensure optimal diagnostic quality.

Figure 6 A 2-month-old boy with bilateral hydronephrosis. Images from the (A) AP, (B) VP, (C) VNC-AP, (D) VNC-VP, and (E) TNC. Note the severe incomplete iodine subtraction in the IVC on the VNC-AP image (C) due to the extreme local iodine concentration during the AP image (A). This artifact is successfully resolved on the later VNC-VP image (D), which closely resembles the TNC reference (E). AP, arterial phase; IVC, inferior vena cava; TNC, true noncontrast; VNC-AP, virtual noncontrast image reconstructed from the arterial phase; VNC-VP, virtual noncontrast image reconstructed from the venous phase; VP, venous phase.

A major strength of this study is the comprehensive evaluation of VNC across a wide array of pediatric urinary tract diseases. However, several limitations should be noted. First, the CT attenuation of the renal calculi on VNC images was significantly underestimated, with maximum discrepancies reaching several hundred HU. This occurred because the DECT material decomposition algorithm typically utilizes iodine, fat, and soft tissue as base materials. The primary component of renal calculi is calcium, which exhibits a spectral attenuation profile that partially overlaps with that of iodine at the 80 kV/Sn150 kV energy levels. During the iodine subtraction process, the algorithm erroneously identifies high-density calcium as iodine, leading to an inadvertent subtraction and a subsequent reduction in the apparent density of the calculi (25,26). Therefore, when the calculus composition is being characterized based on CT attenuation, we recommend interpreting VNC images in conjunction with the original contrast-enhanced images, iodine maps, and the broader clinical context. If precise attenuation measurements are strictly required, preserving the TNC scan may be necessary. Second, this is a single-center, retrospective study. Second, we employed single-center, retrospective design, and although the overall total cohort (217 cases) was substantial for a pediatric study, certain subcohorts—specifically those composed of cases involving renal calculi (n=6), ectopic kidneys (n=7), and renal lacerations/contusions (n=8)—were small. Consequently, the statistical power for these pathologies is limited, and these findings must be interpreted with caution and as preliminary observations. Future prospective, multicenter trials featuring larger targeted pathological subgroups are warranted to confirm these clinical evaluations. Third, while our four-phase scanning protocol was optimized to ensure a direct intraindividual comparison, it does not represent the absolute minimum achievable radiation dose. For instance, Chen et al. (27) demonstrated that employing a split-bolus contrast injection technique combined with DECT in pediatric CTU can simultaneously capture a well-defined corticomedullary phase within a single enhanced acquisition. This approach condenses the protocol to only two phases (corticomedullary and excretory), offering a clear, actionable pathway for further dose reduction.

Conclusions

VNC imaging derived from DECT provides diagnostic image quality equivalent to that of TNC scans for the majority of pediatric urinary tract conditions, enabling a significant radiation dose reduction of approximately 24.6% in adherence to the ALARA principle (28). However, a dedicated TNC scan remains necessary for the accurate quantitative assessment of renal calculi. Radiologists should remain cognizant of inherent algorithmic limitations, such as localized high-iodine artifacts in young infants, to ensure absolute diagnostic accuracy.

Acknowledgments

We extend our sincere thanks to the Wuhan Clinical Research Center for Children’s Medical Imaging for project approval and platform support in this diagnostic accuracy and agreement study. We would also like to express our gratitude to the clinical radiographers and the pediatric urology data management teams at Wuhan Children’s Hospital for their dedicated collaboration during patient scanning and text formatting.

Footnote

Reporting Checklist: The authors have completed the GRRAS reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0695/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0695/dss

Funding: This study was supported by the Wuhan Clinical Research Center for Children’s Medical Imaging (No. WSTB2022-38).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0695/coif). J.W., B.Z., H.Y., Y.G., and X.P. report that this study was supported by the Wuhan Clinical Research Center for Children’s Medical Imaging (No. WSTB2022-38). W.P. is an employee of Siemens Healthineers, Shanghai. The authors have no other conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board (IRB) of Wuhan Children’s Hospital (Wuhan Maternal and Child Healthcare Hospital), Tongji Medical College, Huazhong University of Science & Technology (approval No. 2026R011-E01), and written informed consent was obtained from the legal guardians of all pediatric patients.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Saini S, Bhatia A, Bansal A, Saxena AK, Sodhi KS. Dual-energy computed tomography in children: Technique and clinical applications. World J Clin Pediatr 2025;14:108823. [Crossref] [PubMed]
2. Aldughiem A. Imaging Diagnosis of Major Kidney and Urinary Tract Disorders in Children. Medicina (Kaunas) 2025;61:696. [Crossref] [PubMed]
3. Xie M, Wang H, Tang S, Chen M, Li T, He L. Application of dual-energy CT with prospective ECG-gating in cardiac CT angiography for children: Radiation and contrast agent dose. Eur J Radiol 2024;170:111229. [Crossref] [PubMed]
4. Mileto A, Ananthakrishnan L, Morgan DE, Yeh BM, Marin D, Kambadakone AR. Clinical Implementation of Dual-Energy CT for Gastrointestinal Imaging. AJR Am J Roentgenol 2021;217:651-63. [Crossref] [PubMed]
5. Cellina M, Cè M, Grimaldi E, Mastellone G, Fortunati A, Oliva G, Martinenghi C, Carrafiello G. The role of dual-energy computed tomography (DECT) in emergency radiology: a visual guide to advanced diagnostics. Clin Radiol 2025;83:106836. [Crossref] [PubMed]
6. Tu B, Jia J, Yu L, Li H, Wang D. Correlative investigation between routine clinical parameters of dual-energy computed tomography and the outcomes of extracorporeal shock wave lithotripsy in children with urolithiasis: a retrospective study. Abdom Radiol (NY) 2021;46:4881-7. [Crossref] [PubMed]
7. Noda SM, Romberg EK, Wishart J, Kim HHR, Stanescu AL, Phillips G. Dual energy CT of pediatric genitourinary disorders. Abdom Radiol (NY) 2026;51:3971-9. [Crossref] [PubMed]
8. Reimer RP, Zaytoun H, Klein K, Sonnabend K, Lennartz S, Zopfs D, Heidenreich A, Maintz D, Große Hokamp N. Detection and size measurements of kidney stones on virtual non-contrast reconstructions derived from dual-layer computed tomography in an ex vivo phantom setup. Eur Radiol 2023;33:2995-3003. [Crossref] [PubMed]
9. Dudás I, Schultz L, Benke M, Szücs Á, Kaposi PN, Szijártó A, Maurovich-Horvat P, Budai BK. The reliability of virtual non-contrast reconstructions of photon-counting detector CT scans in assessing abdominal organs. BMC Med Imaging 2024;24:237. [Crossref] [PubMed]
10. Chen CY, Hsu JS, Jaw TS, Shih MC, Lee LJ, Tsai TH, Liu GC. Split-Bolus Portal Venous Phase Dual-Energy CT Urography: Protocol Design, Image Quality, and Dose Reduction. AJR Am J Roentgenol 2015;205:W492-501. [Crossref] [PubMed]
11. Si-Mohamed S, Dupuis N, Tatard-Leitman V, Rotzinger D, Boccalini S, Dion M, Vlassenbroek A, Coulon P, Yagil Y, Shapira N, Douek P, Boussel L. Virtual versus true non-contrast dual-energy CT imaging for the diagnosis of aortic intramural hematoma. Eur Radiol 2019;29:6762-71. [Crossref] [PubMed]
12. Corwin MT, Altinmakas E, Asch D, Bishop KA, Boge M, Curci NE, et al. Clinical Importance of Incidental Homogeneous Renal Masses That Measure 10-40 mm and 21-39 HU at Portal Venous Phase CT: A 12-Institution Retrospective Cohort Study. AJR Am J Roentgenol 2021;217:135-40. [Crossref] [PubMed]
13. Sharma SP, Budde RPJ, Groen JW, Dijkshoorn ML, Manintveld OC, Hirsch A. Comparison of image quality between photon-counting detector CT and energy-integrating detector coronary CT angiography in heart transplant patients. Int J Cardiovasc Imaging 2026;42:393-401. [Crossref] [PubMed]
14. Adhikari K, Regmi N, Devkota K, Koirala S. The Impact of Pelvicalyceal System Anatomy on Stone Formation in Patients with Single Lower Pole Calyceal Stone on Computed Tomography Urography. J Nepal Health Res Counc 2024;22:91-5. [Crossref] [PubMed]
15. Syrgiamiotis V, Ploussi A, Rallis S, et al. Pediatric Local Diagnostic Reference Levels for CT Examinations: A Study Based on the European Guidelines on Diagnostic Reference Levels for Pediatric Imaging. Cureus 2025;17:e85284. [Crossref] [PubMed]
16. Vañó E, Miller DL, Martin CJ, Rehani MM, Kang K, Rosenstein M, Ortiz-López P, Mattsson S, Padovani R, Rogers A. Authors on behalf of ICRP. ICRP Publication 135: Diagnostic Reference Levels in Medical Imaging. Ann ICRP 2017;46:1-144. [Crossref] [PubMed]
17. Oechsner T, Soschynski M, Schlett CL, Krauss T, Schupppert C, Müller-Peltzer K, Vecsey-Nagy M, Kravchenko D, Varga-Szemes Á, Emrich T, Scheu R, Taron J, Bamberg F, Hagar MT. Feasibility of very low iodine dose aortoiliac CT angiography using dual-source photon-counting detector CT. Eur J Radiol 2025;183:111919. [Crossref] [PubMed]
18. Koo TK, Li MY. A Guideline of Selecting and Reporting Intraclass Correlation Coefficients for Reliability Research. J Chiropr Med 2016;15:155-63. Erratum in: J Chiropr Med 2017;16:346.
19. Gallo-Bernal S, Peña-Trujillo V, Gee MS. Dual-energy computed tomography: pediatric considerations. Pediatr Radiol 2024;54:2112-26. [Crossref] [PubMed]
20. Ananthakrishnan L, Rajiah P, Ahn R, Rassouli N, Xi Y, Soesbe TC, Lewis MA, Lenkinski RE, Leyendecker JR, Abbara S. Spectral detector CT-derived virtual non-contrast images: comparison of attenuation values with unenhanced CT. Abdom Radiol (NY) 2017;42:702-9. [Crossref] [PubMed]
21. Javadi S, Elsherif S, Bhosale P, Jensen CT, Layman RR, Jacobsen MC, Le O, Jia S, Parikh RJ, Tamm EP. Quantitative attenuation accuracy of virtual non-enhanced imaging compared to that of true non-enhanced imaging on dual-source dual-energy CT. Abdom Radiol (NY) 2020;45:1100-9. [Crossref] [PubMed]
22. Graser A, Johnson TR, Hecht EM, Becker CR, Leidecker C, Staehler M, Stief CG, Hildebrandt H, Godoy MC, Finn ME, Stepansky F, Reiser MF, Macari M. Dual-energy CT in patients suspected of having renal masses: can virtual nonenhanced images replace true nonenhanced images? Radiology 2009;252:433-40. [Crossref] [PubMed]
23. Lee MH, Park HJ, Kim JN, Kim MS, Hong SW, Park JH, Kang CH. Virtual non-contrast images from dual-energy CT angiography of the abdominal aorta and femoral arteries: comparison with true non-contrast CT images. Br J Radiol 2022;95:20220378. [Crossref] [PubMed]
24. Sauter AP, Muenzel D, Dangelmaier J, Braren R, Pfeiffer F, Rummeny EJ, Noël PB, Fingerle AA. Dual-layer spectral computed tomography: Virtual non-contrast in comparison to true non-contrast images. Eur J Radiol 2018;104:108-14. [Crossref] [PubMed]
25. Greffier J, Villani N, Defez D, Dabli D, Si-Mohamed S. Spectral CT imaging: Technical principles of dual-energy CT and multi-energy photon-counting CT. Diagn Interv Imaging 2023;104:167-77. [Crossref] [PubMed]
26. Peña-Trujillo V, Gallo-Bernal S, Tung EL, Gee MS. Pediatric Applications of Dual-Energy Computed Tomography. Radiol Clin North Am 2023;61:1069-83. [Crossref] [PubMed]
27. Chen H, Deng Z, Tan T, Li S, Qian F, He L, Tang S. Application value of split-bolus contrast injection combined with dual-energy CT scanning technology in pediatric CTU imaging. Eur J Radiol 2025;184:111949. [Crossref] [PubMed]
28. Nagy E, Tschauner S, Schramek C, Sorantin E. Paediatric CT made easy. Pediatr Radiol 2023;53:581-8. [Crossref] [PubMed]