Free-breathing non-contrast-enhanced flow-independent MR angiography using REACT: A prospective study for pediatric vessel assessment

Abstract

Purpose

To evaluate the non-contrast-enhanced relaxation-enhanced angiography without contrast (REACT) sequence for the assessment of extrathoracic vessels in pediatric patients compared to contrast-enhanced (CE), multiphasic magnetic resonance angiography (MRA).

Materials and Methods

In this prospective, single-center study, pediatric patients referred for clinically indicated contrast-enhanced MRI of various extrathoracic body regions underwent additional free-breathing REACT and multiphasic, free-breathing CE-MRA at 1.5 T (Philips Ingenia). REACT was acquired using Cartesian k-space order, except in the abdomen, where it was acquired using a radial stack of stars k-space sampling (REACT VANE). The acquisition time was recorded. Image quality (Likert scale 1–5, with 5 being the best) and vessel diameter were evaluated by two independent readers in four predefined vessels in each body region. Furthermore, a quantitative analysis of SNR and CNR was performed.

Results

30 patients (age: 12.3 ± 4 years) successfully completed REACT and CE-MRA. The acquisition time for REACT was 2:49 ± 1:03 min, while abdominal REACT VANE required 4:51 ± 0:52 min. The CE-MRA acquisition time was 3:49 ± 1:03 min. The median image quality ratings were good to excellent (Likert scale 4–5) for both readers. No significant difference in the image quality ratings was found (p = 0.12 – 0.58). Interobserver agreement of image quality ratings of the two readers was moderate to substantial (Cohen’s kappa REACT: 0.58, CE-MRA: 0.64). Vessel diameter measurements showed a strong correlation (r = 0.93) between REACT and CE-MRA with high intraclass correlation coefficients (REACT: 0.97, CE-MRA: 0.97). Quantitative analysis showed a higher venous SNR and higher arterial and venous CNR in REACT (p = 0.001–0.018).

Conclusion

Given the good and comparable image quality, REACT can be useful in vascular imaging in children under free-breathing, while potentially eliminating the need for contrast agent injection.

Key Points

• MR angiography is widely used in pediatric imaging for vessel assessment.

• Contrast-enhanced MRA has limitations due to the use of gadolinium-based contrast agents.

• REACT is a novel contrast-free MRA technique performed during free breathing.

• REACT provides image quality comparable to contrast-enhanced free-breathing MRA.

Citation Format

• Spogis J, Tsiflikas I, Katemann C et al. Free-breathing non-contrast-enhanced flow-independent MR angiography using REACT: A prospective study for pediatric vessel assessment. Rofo 2026; 10.1055/a-2781-8861

Zusammenfassung

Zweck

Ziel dieser Studie war die Evaluation der kontrastmittelfreien Relaxation-Enhanced Angiography without Contrast (REACT)-Sequenz zur Beurteilung extrathorakaler Gefäße bei pädiatrischen Patienten im Vergleich zur kontrastmittelbasierten (CE), multiphasischen Magnetresonanzangiografie (MRA).

Material und Methoden

In dieser prospektiven, monozentrischen Studie erhielten pädiatrische Patienten, die zur klinisch indizierten kontrastmittelverstärkten MRT verschiedener extrathorakaler Körperregionen überwiesen wurden, zusätzlich eine kontrastmittelfreie REACT sowie eine multiphasische CE-MRA unter freier Atmung an einem 1,5 T MRT (Philips Ingenia). Die REACT wurde mittels kartesischem k-Raum-Sampling akquiriert, außer im Abdomen, wo eine radiale Stack-of-Stars-Akquisition (REACT VANE) angewendet wurde. Die Akquisitionszeit wurde aufgezeichnet. Die Bildqualität (Likert-Skala 1–5, wobei 5 die beste Qualität darstellt) und der Gefäßdiameter wurden durch zwei unabhängige Untersucher an vier vordefinierten Gefäßen pro Körperregion bewertet. Darüber hinaus wurde eine quantitative Analyse von SNR und CNR durchgeführt.

Ergebnisse

30 Patienten (Alter: 12,3 ± 4 Jahre) erhielten sowohl eine REACT als auch CE-MRA. Die Akquisitionszeit für REACT betrug 2:49 ± 1:03 min, während die abdominale REACT VANE 4:51 ± 0:52 min erforderte. Die Akquisitionszeit für CE-MRA lag bei 3:49 ± 1:03 min. Die mediane Bildqualität wurde von beiden Untersuchern als gut bis exzellent bewertet (Likert-Skala 4–5) ohne signifikante Unterschiede zwischen den Sequenzen (p = 0,12–0,58). Die Interobserver-Übereinstimmung der Bildqualitätsbewertungen war moderat bis substanziell (Cohen’s kappa: REACT 0,58; CE-MRA: 0,64). Die Gefäßdiameter zeigten eine hohe Korrelation zwischen REACT und CE-MRA (r = 0,93) mit hohen Intraklassen-Korrelationskoeffizienten (REACT: 0,97; CE-MRA: 0,97). In der quantitativen Analyse zeigte die REACT höhere venöse SNR-Werte sowie höhere arterielle und venöse CNR-Werte im Vergleich zur CE-MRA (p= 0,001 – 0,018).

Schlussfolgerung

Angesichts der guten und vergleichbaren Bildqualität könnte die REACT-Sequenz eine nützliche Technik für die Gefäßbildgebung unter freier Atmung bei Kindern darstellen – mit dem Potenzial, auf die Gabe von Kontrastmitteln zu verzichten.

Kernaussagen

• Die MR-Angiografie ist eine etablierte Bildgebung zur Beurteilung der Gefäße bei pädiatrischen Patienten.

• Die kontrastmittelangehobene MRA ist durch den Einsatz gadoliniumhaltiger Kontrastmittel limitiert.

• REACT ist eine neuartige, kontrastmittelfreie MRA-Technik unter freier Atmung.

• REACT zeigt eine mit CE-MRA vergleichbare Bildqualität.

Keywords

vascular - MR angiography - MR imaging

Introduction

Magnetic resonance angiography (MRA) is a noninvasive method commonly used for examining the blood vessels in children and adults. MRA is particularly attractive for pediatric patients due to its lack of ionizing radiation. Contrast-enhanced MRA (CE-MRA) is widely accepted as the gold standard for evaluating vascular pathologies. However, obtaining MRA images of diagnostic quality in pediatric patients can be challenging due to factors such as patient movement, the inability to hold their breath, higher heart rates, short circulation times, and smaller volumes of contrast material, which complicate capturing the peak enhancement [1]. Additionally, despite the excellent safety profile of macrocyclic gadolinium-based contrast agents, there is an ongoing debate about gadolinium retention in tissues after repeated administration [2]. While no evidence of associated toxicity has been found to date, and the clinical significance remains unclear, reducing gadolinium exposure is especially important in children [3] [4]. Therefore, there is a need to obtain high-quality, non-contrast-enhanced images of the vessels. Several non-CE-MRA techniques are available for evaluating vessels in different body regions and disease contexts, including time-of-flight, phase contrast, and balanced steady-state free precession, though each has its own limitations [5].

Relaxation-enhanced angiography without contrast (REACT) was introduced by Yoneyama et al. in 2019 to address the limitations of non-CE-MRA and CE-MRA by developing a robust, free-breathing technique with high contrast and spatial resolution, without the need for contrast agent administration [6]. This technique starts with a T2-prep pulse consisting of four adiabatic refocusing pulses to suppress the signal from tissues with short T2 [7]. Following this, a non-selective inversion recovery pulse is applied with a short inversion time to suppress signals from tissues such as internal organs or muscles. A 3D dual-echo Dixon acquisition with semi-flexible echo times for water-fat separation is then performed, enabling effective fat suppression across large volumes of interest, thereby enhancing the vascular signal against background tissues [8]. The REACT sequence has already shown promising results with regard to the imaging of thoracic vasculature [9] [10] [11] and the evaluation and follow-up of congenital heart disease [12] [13] and with regard to extracranial arteries [14] [15] and other vascular territories [16]. However, apart from a feasibility study, systematic data on the use of REACT for extrathoracic vessel assessment in pediatric patients remain limited [17].

Therefore, the aim of this study was to evaluate non-CE REACT for the assessment of extrathoracic arteries and veins in pediatric patients, using contrast-enhanced free-breathing multiphasic MRA as the reference standard.

Materials and methods

Study design

This prospective, single-center study was approved by the institutional review board. Informed consent was obtained from all participants and their parents. Thirty-two patients referred from the pediatric department for clinically indicated contrast-enhanced MRI between February and October 2023 were included in this study and underwent additional REACT. As vascular imaging was not the primary indication in most cases, the cohort mainly represented normal vascular anatomy. The sample size was determined by feasibility within the study period, and multiphasic free-breathing CE-MRA served as the clinical reference standard.

MRI acquisition

All examinations were performed using a 1.5T whole-body clinical MRI scanner (Ingenia Evolution; Philips Healthcare, Best, the Netherlands). The acquisition protocol included the clinically indicated non-contrast-enhanced MRI sequences, which varied based on the body region and clinical indication. REACT was then acquired, followed by a multiphasic CE-MRA with 9–11 phases, utilizing 4D free-breathing radial stack-of-stars imaging with an efficient Diamond Sampling Strategy and view sharing with temporal domain k-space weighted image contrast (KWIC)[18]. This sequence was acquired before, during, and immediately after the administration of 0.1 mL/kg of gadobutrol (Gadovist; Bayer Vital, Leverkusen, Germany) at an injection rate of 0.5–1 mL/s, depending on the patient’s age. Subsequently, the remaining contrast-enhanced sequences from the clinical MRI protocol were acquired. REACT was acquired using Cartesian sampling of the k-space with compressed SENSE (CS-SENSE) acceleration, except in the abdominal and pelvic region, where pseudo golden angle radial stack of stars k-space sampling and respiratory gating were used (REACT VANE) with conventional SENSE to allow free-breathing acquisition [6] [16]. Radial acquisition was chosen for the abdomen because of its higher robustness to respiratory motion [19] [20]. Both REACT and CE-MRA were acquired in different planes depending on the body region: transverse (skull base, neck, upper abdomen), coronal (abdomen, pelvis, wrist, knee), or sagittal (ankle). Detailed acquisition parameters are provided in [Table 1].

Table 1 MRI acquisition parameters for REACT and multiphasic contrast-enhanced MRA (CR-MRA). mDIXON = modified DIXON; SENSE = SENSitivity encoding; CS-SENSE = compressed sensing SENSE; SPAIR = spectral attenuated inversion recovery.

	REACT VANE	REACT		Multiphasic CE-MRA
Body region	Abdomen/pelvis	Neck/skull base	Wrist/knee/ankle	Abdomen/pelvis	Neck/skull base	Wrist/knee/ankle
Time of echo (ms)	1.91/4.3	1.91/4.5	1.93/4.5	1.83	1.94	1.96
Time of repetition (ms)	6.8	6.6	6.9	4.4	4.6	4.7
Orientation	Coronal	Transverse	Coronal/coronal/sagittal	Coronal	Transverse	Coronal/coronal/sagittal
Voxel size (mm³), acquired	2 × 2 × 1.8	1.5 × 1.5 × 3	1.5 × 1.5 × 2	1.3 × 1.3 × 4	1.3 × 1.3 × 4	1.3 × 1.3 × 3
Voxel size (mm³), reconstructed	0.85 × 0.85 × 0.9	0.68 × 0.68 × 1.5	0.65 × 0.65 × 1	0.54 × 0.54 × 2	0.69 × 0.69 × 2	0.65 × 0.65 × 1.5
Acquisition matrix	152 × 152	232 × 232	168 × 164	276 × 276	256 × 256	192 × 192
Field of view (mm³)	300 × 300 × 88	350 × 350 × 150	250 × 250 × 70	360 × 360 × 160	330 × 330 160	250 × 250 × 60
Sampling order of k-space	Radial	Cartesian	Cartesian	Radial	Radial	Radial
T2 prep time (ms)	50	50	50	n/a	n/a	n/a
Turbo field echo factor	32	60	58	26	30	30
Flip angle (°)	12	12	12	12	15	15
Fat suppression	mDIXON			SPAIR
SENSE reduction	2	-	-	2.6	1.7	1.7
CS-SENSE reduction	–	6	6	–	–	–
Respiratory gating	Yes	No	No	Yes	No	No
Dynamic time (s)	n/a	n/a		16	16	16

Subjective image analysis

REACT and CE-MRA images were independently evaluated by two radiologists with 5 and 18 years of MRI experience. The readings were conducted using a dedicated workstation (Centricity PACS RA1000; GE Healthcare, Wisconsin). Image quality (IQ) of predetermined vessels of interest (two arteries and two veins per body region; [Table 2]) was assessed using a Likert scale ranging from 1 to 5, with 5 indicating the highest quality (1: non-diagnostic, 2: poor, 3: intermediate, 4: good, 5: excellent). Additionally, the same vessels were selected for measurement of vessel diameter. Both readers performed the measurements separately, recording the maximal transverse diameter at predefined positions. All measurements were conducted on the original images. Vascular anomalies, if present, were documented during image evaluation.

Table 2 Vessels of interest for image quality analysis and diameter measurements.

Body region	Artery 1	Artery 2	Vein 1	Vein 2
Skull base	Right internal carotid artery	Basilar artery	Transverse sinus	Left internal jugular vein
Neck	Right common carotid artery	Left vertebral artery	Right internal jugular vein	Right subclavian vein
Upper abdomen	Infrarenal aorta	Celiac trunk	Portal vein	Left renal vein
Pelvis	Right external iliac artery	Left internal pudendal artery	Right external iliac vein	Left common femoral vein
Knee	Superficial femoral artery	Popliteal artery	Superficial femoral vein	Popliteal vein
Ankle	Posterior tibial artery	Dorsalis pedis artery	Posterior tibial vein	Peroneal vein
Wrist	Radial artery	Ulnar artery	Radial vein	Ulnar vein

Quantitative analysis

For signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) assessment, ROIs were placed in the same predefined vessels of interest (ROI size 0.07–3 cm², depending on the size of the vessel) as well as in subcutaneous fat (1 cm²), which served as the background reference. All measurements were performed by reader 1. Vascular ROIs were positioned in areas with a homogeneous signal, avoiding vessel margins. For each vessel, the mean signal intensity (SI_Vessel) was determined, while the mean signal intensity (SI_Background) and the standard deviation (SD_Background) were obtained from the background ROI. The SNR was calculated as the ratio of mean signal intensity to the standard deviation of the background (SNR = SI_Vessel/SD_Background). The CNR was defined as the absolute difference between the mean signal intensity of the vessel and subcutaneous fat, normalized to the standard deviation of the background ROI (CNR = (SI_Vessel – SI_Background)/SD_Background). Identical ROI positions were used for REACT and CE-MRA datasets. In the CE-MRA acquisition, ROIs were placed in the contrast phase with the highest vascular signal intensity. Subcutaneous fat was chosen as the background region as a homogeneous tissue area for stable noise estimation in images reconstructed with parallel imaging and compressed sensing.

Statistical analysis

Statistical analysis was performed using JMP 16.2.0 (SAS Institute Inc, Cary, NC) and MedCalc 18.10 (MedCalc Software Ltd, Ostend, Belgium). Parametric variables are presented as mean ± standard deviation (SD), while nonparametric variables are expressed as median and interquartile range (IQR). The Wilcoxon signed-rank test was used for paired comparisons of image quality, and the Mann-Whitney U test for unpaired data. The chi-squared test was applied to compare the presence of poor or non-diagnostic image quality. Inter-reader agreement was assessed using weighted Cohen’s kappa. Paired t-test, Bland–Altman analysis, and Pearson’s correlation were used to evaluate differences in vessel diameter measurements. Intraclass correlation coefficients (ICC) were calculated to assess interobserver agreement, and the Kruskal-Wallis test was used to compare image quality and vessel diameter. P-values below 0.05 were considered statistically significant. All p-values were adjusted for multiple comparisons using Holm’s method [21]. The multiple comparison adjustment was applied separately to the group of tests comparing SNR/CNR, image quality, and vessel diameter. To address potential pseudo-replication from multiple vessels per patient, an additional analysis was performed at the patient level. For vessel diameters, patient-level paired differences were calculated separately for arteries and veins per reader and tested against zero using paired t-test. For image quality, median Likert scores across all four vessels per patient were compared between REACT and CE-MRA using the Wilcoxon signed-rank test.

Results

Thirty-two patients were initially enrolled in the study. Due to technical errors in acquiring REACT, two patients were excluded, resulting in a final study cohort of thirty patients. The mean age was 12.3 ± 4 years (range: 5–19 years), with 16/30 (53%) female patients. The most frequently examined body region was the abdomen (15/30; 50%), followed by the neck (4/30; 13%). Patient characteristics and body regions are summarized in [Table 3].

Table 3 Patient characteristics.

Characteristics	Values
Number of patients	N = 30
Age, mean ± standard deviation	12.3 ± 4 years
Sex	16 (53%) female, 14 (47%) male
Body regions
Skull base	N = 3
Neck	N = 4
Abdomen	N = 15
Pelvis	N = 3
Wrist	N = 2
Knee	N = 1
Ankle	N = 2

Sequence acquisition

The total acquisition time was 3:50 ± 1:24 min for REACT and 3:49 ± 1:03 min for multiphasic MRA (including pre-scan and 9–11 phases, p = 0.88). The acquisition time for REACT was significantly longer in the abdomen and pelvis with respiratory gating compared to the other body regions acquired without respiratory gating (4:51 ± 0:52 min for the abdomen, 2:49 ± 1:03 min for other body regions; p < 0.001).

Image quality analysis

Image quality was assessed by each reader for 120 vessels on REACT and multiphasic MRA. There was no significant difference in overall image quality ratings between REACT and multiphasic MRA (reader 1: adjusted p = 0.12; reader 2: adjusted p = 0.58; [Fig. 1], [Fig. 2]). Across body regions, image quality was consistently rated good to excellent for both techniques ([Table 4]). Non-diagnostic or poor image quality (IQ < 3) was observed in 4 out of 120 (3.3%) vessels on REACT and 3 out of 120 (2.5%) vessels on multiphasic MRA, with no significant difference between the two (p = 0.7). A scrotal vascular malformation was detected in one patient ([Fig. 3]), and compression of the left internal jugular vein due to a large neuroblastoma in another. In all remaining patients, no vascular pathologies were identified.

Fig. 1 Transverse images of the neck (A: REACT, voxel size = 0.68 × 0.68 × 1.5 mm; B: Multiphasic CE-MRA, voxel-size = 0.69 × 0.69 × 2.0 mm). Both sequences provide good image quality with clear visualization of the carotid arteries (small arrows) and jugular veins (large arrows). Note that in REACT, the cerebrospinal fluid and spinal cord appear hyperintense.

Fig. 2 Coronal maximum intensity projections of REACT (original voxel size = 0.65 × 0.65 × 1 mm; A) and multiphasic MRA in the arterial phase (original voxel size = 0.65 × 0.65 × 1.5 mm; B). Both REACT and contrast-enhanced MRA clearly depict the radial and ulnar arteries (small arrows), as well as the carpal arteries (large arrows). In REACT MRA, the vessels are partially overlaid by the carpal veins.

Table 4 Median (interquartile range) image quality for REACT and multiphasic CE-MRA. Likert-scale ranging from 1–5 with 5 being the best.

Characteristics		Reader 1			Reader 2
		REACT	Multiphasic MRA	adjusted p-value	REACT	Multiphasic MRA	adjusted p-value
Overall		4 (4–5)	5 (4–5)	0.12	4 (4–5)	4 (4–5)	0.58
	Arteries	5 (4–5)	5 (3–5)		4 (4–5)	4.5 (4–5)
	Veins	4 (4–5)	4 (4–5)		4 (4–4)	4 (3–5)
Abdomen/pelvis	Arteries	5 (4–5)	5 (4–5)		5 (4–5)	5 (4–5)
	Veins	4 (4–5)	4.5 (4–5)		4 (4–5)	4 (3–5)
Skull base/neck	Arteries	5 (5–5)	5 (5–5)		4 (4–5)	5 (4–5)
	Veins	4 (4–5)	5 (4–5)		4 (3–4)	5 (4–5)
Wrist/ankle/knee	Arteries	4 (3–4)	5 (4–5)		4 (4–4)	4 (3–4)
	Veins	3 (3–4)	3.5 (2–4)		3 (3–4)	4 (3–4)

Fig. 3 Coronal maximum intensity projections of the pelvis of a 5-year-old boy with a scrotal vascular malformation. Both REACT (original voxel size = 0.55 × 0.55 × 1.3 mm; A) and multiphasic MRA in the venous phase (original voxel size = 0.79 × 0.79 × 2 mm; B) demonstrate superior image quality, with clear delineation of the pelvic arteries and veins, as well as the ectatic vessels of the vascular malformation (asterisk). Note that in REACT, arteries are more hyperintense than veins (arrow).

Vessel measurements

Vessel diameter measurements were obtained for 120 vessels by each reader. For reader 1, a strong correlation and close intermethod agreement were observed between REACT and multiphasic MRA (r = 0.93; bias = 0.12 ± 0.22 mm, 95% limits of agreement: −2.25 to 2.49 mm, [Fig. 4]). Although the mean arterial diameters were slightly larger on multiphasic MRA than on REACT (reader 1: 6.2 ± 3.0 vs. 5.9 ± 2.8 mm; reader 2: 6.1 ± 2.9 vs. 5.8 ± 2.7 mm), this difference was not statistically significant (adjusted p = 0.136 to 0.78). Venous diameters were nearly identical across both techniques for both readers (all adjusted p = 1.0; [Table 5]). Inter-rater agreement for image quality was moderate to substantial (weighted Cohen’s kappa: REACT, 0.58; multiphasic MRA, 0.64), while vessel diameter measurements showed high intraclass correlation coefficients (REACT: 0.97; multiphasic MRA: 0.97; [Table 6]).

Fig. 4 (A) Bland-Altman plot and (B) scatter plot showing the correlation of vessel diameter measurements between REACT and multiphasic MRA for reader 1. Pearson’s correlation coefficient (r = 0.93) indicates a strong correlation. REACT = relaxation-enhanced angiography without contrast.

Table 5 Vessel diameter measurements by both readers of REACT and multiphasic MRA, presented as mean ± standard deviation.

Characteristics	Reader 1			Reader 2
	REACT	Multiphasic MRA	Adjusted p-value	REACT	Multiphasic MRA	Adjusted p-value
Arteries (mm)	5.9 ± 2.8	6.2 ± 3.0	0.136	5.8 ± 2.7	6.1 ± 2.9	0.78
Veins (mm)	8.1 ± 3.2	8.1 ± 3.2	1.0	8.1 ± 3.5	8.1 ± 3.3	1.0

Table 6 Interobserver agreement for image quality for vessel diameter measurements for REACT and multiphasic MRA. Values in parentheses are the 95 % confidence intervals. *Weighted Cohen’s kappa. **Intraclass correlation coefficient (ICC).

	REACT	Multiphasic MRA
Image quality*
Overall	0.583 (0.450–0.717)	0.639 (0.553–0.726)
Arteries	0.464 (0.244–0.684)	0.523 (0.337–0.710)
Veins	0.618 (0.448–0.789)	0.650 (0.551–0.750)
Vessel diameters**
Overall	0.965 (0.949–0.976)	0.965 (0.950–0.976)
Arteries	0.973 (0.955–0.984)	0.983 (0.971–0.990)
Veins	0.952 (0.919–0.971)	0.944 (0.906–0.967)

In the patient-level analysis of vessel diameters, Reader 1 observed slightly smaller arterial diameters on REACT compared to CE-MRA (mean difference = −0.27 mm, adjusted p = 0.329), while venous diameters were nearly identical across both techniques (mean difference = 0.03 mm, adjusted p = 1.0). Reader 2 similarly identified no significant arterial or venous differences, with the mean difference ranging from −0.09 to −0.08 mm (all adjusted p = 1.0). For image quality, patient-level median differences between REACT and CE-MRA were zero for both readers (Reader 1: p = 0.984; Reader 2: p = 0.501; both adjusted p = 1.0).

Quantitative analysis

Both arterial and venous SNR values were higher in REACT compared with multiphasic CE-MRA, although only the venous SNR difference reached statistical significance (arteries: median 41.3 [IQR 32.1–60.2] vs. 35.0 [IQR 22.2–49.1]; adjusted p = 0.072; veins: 41.2 [IQR 29.4–60.2] vs. 32.1 [IQR 22.9–48.9]; adjusted p = 0.018). Accordingly, arterial and venous CNR values were significantly higher in the case of REACT compared with multiphasic CE-MRA (arteries: 38.3 [IQR 29.1–57.2] vs. 28.7 [IQR 17.2–40.5]; adjusted p = 0.002; veins: 38.3 [IQR 26.1–57.2] vs. 27.3 [IQR 17.5–44.1]; adjusted p = 0.001).

Discussion

In this prospective study, we compared the free-breathing non-contrast-enhanced REACT sequence with multiphasic free-breathing MRA with contrast enhancement. This study evaluated extrathoracic arteries and veins in pediatric patients, including the application of REACT VANE for abdominal imaging, which allowed robust free-breathing acquisition. We demonstrated that the implementation of REACT is feasible across multiple extrathoracic body regions, including the abdomen, pelvis, neck, and limbs. The overall image quality was not significantly different from that of multiphasic contrast-enhanced MRA, although comparisons across individual body regions should be interpreted descriptively due to the small sample sizes within each subgroup. Both sequences received only a few non-diagnostic ratings, indicating good diagnostic performance. Vessel diameter measurements were consistent between the two readers, with no tendency for under- or overestimation of small or large vessels and no significant differences between arteries and veins. Measurements of arteries were slightly larger in the case of first-pass CE-MRA with a mean difference of 0.3 mm and no trend observed in the Bland-Altman plot for small or large vessels, suggesting that the lumen of the arteries is displayed slightly smaller on REACT. However, the difference of 0.3 mm was below the voxel size of both sequences and the differences were not statistically significant. For veins, no difference was observed. Additional patient-level analyses for vessel diameters and image quality confirmed the main findings. REACT showed superior SNR and CNR for both arteries and veins, even though the contrast phase with the highest vascular signal intensity was selected for CE-MRA. However, these results should be interpreted with caution, as SNR and CNR measurements are not fully comparable between the two sequences due to differences in reconstruction algorithms and slice thickness.

The novel REACT sequence offers several advantages over conventional contrast-enhanced and non-contrast MRI sequences commonly used for vascular imaging. Unlike traditional contrast-enhanced MR angiography, REACT can be performed during free-breathing and does not require gadolinium-based contrast agents. The avoidance of contrast agents eliminates the need for peripheral intravenous cannulation (which is particularly beneficial for the comfort and compliance of young children), prevents allergic reactions, mitigates potential risks associated with gadolinium retention, and reduces costs [3] [22]. Moreover, the REACT sequence is independent of the timing of contrast injection. In pediatric patients, the short circulation times and small amounts of contrast make the precise timing of image acquisition challenging in CE-MRA. REACT can help avoid non-diagnostic scans caused by incorrect timing. Compared to non-contrast time-of-flight techniques, REACT is a flow-independent technique based on the specific relaxation properties of blood, and is, therefore, sensitive to both in-plane and slow blood flow [5]. Like most multiphasic contrast-enhanced angiography examinations and 3D time-of-flight imaging, REACT acquires high-resolution isotropic images, allowing for multiplanar reformations and 3D reconstructions.

There are multiple indications for vascular MR imaging in children. In head and neck imaging, REACT can assess both the carotid arteries and jugular veins. In adults, this sequence has already shown potential for evaluating the external carotid artery and for stroke assessment [14] [15]. However, the assessment of intracranial vessels using REACT is limited by surrounding cerebrospinal fluid. The REACT technique enhances tissues with long T2 times, leading to hyperintensity in fluids such as cerebrospinal fluid, ascites, or pleural effusions ([Fig. 5]). In the neck and upper thorax, REACT can also evaluate venous thromboses and stenoses [11]. Children with chronic illnesses often require central venous catheterization for nutrition or medication administration, which can lead to blood clot formation and occlusion of the jugular and brachiocephalic veins. Prior to recatheterization or replacement of catheter materials, REACT can be employed as a noninvasive technique for evaluating central veins without contrast agents or radiation.

Fig. 5 Transverse images of the skull base in a 12-year-old boy (A: REACT, voxel size = 0.68 × 0.68 × 1.5 mm; B: Multiphasic MRA, voxel size = 0.69 × 0.69 × 2 mm). REACT shows good image quality with visualization of the carotid arteries and the transverse sinus. The basilar artery (arrow) is also depicted, but image quality is reduced due to hyperintense cerebrospinal fluid and brain tissue. In contrast, multiphasic MRA clearly depicts the basilar artery against a hypointense background.

In the abdomen, REACT is effective for imaging and evaluating the aorta and its major branches, the inferior vena cava, and the portal vein. Assessing abdominal vessels is crucial before tumor surgeries, such as those for liver or renal tumors and retroperitoneal tumors like neuroblastoma, where vessel encasement is common and assessment of so-called image-defined risk factors is part of preoperative evaluation [23]. While the image quality of retroperitoneal vessels was excellent in our study, that of intraperitoneal vessels (e.g., the hepatic artery and portal vein) was more variable, although still of diagnostic quality ([Fig. 6]). This variability likely stems from inconsistent patient breathing patterns and the fact that breathing movements in the ventral abdomen occur in both craniocaudal and ventrodorsal directions. In the abdomen, the REACT technique was employed in combination with the radial k-space orders (REACT VANE). Radial k-space acquisitions are generally less sensitive to respiratory movements [24] [25]. As a result, diagnostic image quality for intraperitoneal vessels was achieved in most cases with REACT VANE, although a systematic comparison between Cartesian and radial image acquisition in REACT was not done. Notably, REACT VANE in the abdomen required significantly longer image acquisition times than REACT in other body regions due to respiratory gating.

Fig. 6 Abdominal REACT (voxel size = 0.99 × 0.99 × 0.9 mm) and multiphasic MRA (voxel size = 0.67 × 0.67 × 2 mm) in coronal orientation. Maximum intensity projections (A: REACT, B: Multiphasic MRA in the venous phase) show comparable image quality, with clear visualization of all major intra-abdominal vessels. The intrahepatic vessels are more distinctly visible in REACT compared to multiphasic MRA (arrows). Coronal images (C: REACT, D: multiphasic MRA in the arterial phase) demonstrate good image quality, allowing for precise delineation of the proximal renal arteries and visualization of the lumbar vessels (arrows). In contrast, the image quality of the portal vein (E: REACT, F: Multiphasic MRA in the portal venous phase, asterisk) is diagnostically adequate but is not on par with that of the retroperitoneal vessels. Both sequences were acquired using a radial k-space acquisition technique during free-breathing leading to streak artifacts at the edges of the large field of view.

Out of thirty patients, eight received MRI of the pelvis or extremities. Our study demonstrates that REACT is feasible for imaging peripheral vessels, such as those in the wrist or ankle, and for assessing vascular malformations ([Fig. 3] shows a scrotal vascular malformation in a 5-year-old boy). However, image quality ratings for these peripheral vessels tended to be slightly lower with REACT compared to CE-MRA, and several factors need to be considered when evaluating peripheral vasculature using this technique. First, the REACT technique cannot exclusively enhance either arterial or venous vessels. As such, all vascular structures are hyperintense, and especially in peripheral vessels, it might be difficult to differentiate them using anatomical knowledge ([Fig. 7]). Although arteries are theoretically more hyperintense than veins due to slight differences in relaxation times and T2 preparation pulses, in our experience this distinction is not always helpful in peripheral vessels. Second, peripheral vessels are often quite small, and the resolution of the REACT sequence may be insufficient to accurately assess stenoses or minor caliber alterations. Third, and most importantly, REACT only provides static images of vessel lumens without dynamic information about blood flow, which might be clinically important since vascular malformations are common indications for visualizing peripheral vessels. Thus, REACT is more suitable for assessing size progression over time than for initial diagnosis.

Fig. 7 Sagittal images of the ankle (A: REACT, voxel size = 0.65 × 0.65 × 1 mm; B: Multiphasic MRA, voxel size = 0.65 × 0.65 × 1.5 mm). In REACT, both arteries and veins appear hyperintense. Although the dorsalis pedis artery is clearly depicted (arrow), the vessels are partially blurred, and the overlapping of arteries and veins hampers accurate measurements. The arterial phase of multiphasic MRA provides better image quality with sharper delineation of the dorsalis pedis artery (asterisk).

Recent advances have combined compressed sensing with deep learning–based reconstruction to accelerate and improve REACT acquisitions, particularly in thoracic and pulmonary applications [26] [27]. Such approaches highlight promising future directions for further improving REACT beyond the present analysis. Especially in children, short acquisition times are essential for successful image acquisition.

There are several limitations to this study: First, vascular imaging or suspected vascular pathology was not the primary indication for MRI in most patients. As a result, the present study focused mainly on normal vascular findings and did not systematically assess vascular abnormalities. Future research is required to evaluate the diagnostic performance of REACT with regard to detecting vascular pathologies. Second, no external reference standard such as CT angiography or digital subtraction angiography was used. Instead, multiphasic CE-MRA served as the comparator, as it often represents the clinical standard of care and avoids additional radiation or invasive procedures in children. While this approach is pragmatic and ethically justified, it may limit the generalizability of the findings, and future studies including multimodality comparisons would be valuable. Third, the reference standard we used (CE-MRA) may not serve as an ideal gold standard, as CE-MRA was also acquired during free breathing with keyhole imaging. Typically, high-resolution imaging of abdominal vessels is performed with breath-holding, but this was not always feasible in our patient population. Moreover, the lower spatial resolution of free-breathing CE-MRA may have artificially narrowed the differences observed between REACT and CE-MRA. On the other hand, this approach allows for a more realistic comparison in the common clinical scenario of young children or sedated patients who are unable to hold their breath, where REACT performed comparably to CE-MRA. In addition, multiphasic CE-MRA enables evaluation of arteries and veins, while REACT provides static images of both simultaneously, which makes multiphasic CE-MRA a meaningful comparator. Fourth, the sample size was limited to 30 patients across various anatomical regions, with some regions represented only by one or two patients, precluding dedicated subgroup analyses. The study population was determined by practical feasibility and the study period and was representative of a typical pediatric cohort. Consequently, several findings should be interpreted as exploratory, and larger studies will be required to confirm these observations and allow more robust subgroup analyses. Nevertheless, the primary aim of this study was to demonstrate that REACT MRA can be applied across different body regions and to highlight that REACT VANE yields good results in the abdomen. Fifth, given the clear visual distinction between the two sequences, blinded reading was not feasible, which may have introduced expectancy bias.

In conclusion, the REACT technique demonstrates an image quality comparable to multiphasic free-breathing CE-MRA. Notably, with radial k-space acquisition in REACT VANE, it achieves good image quality for abdominal vessels despite free breathing. In general, key advantages of this technique include high image quality without the need for intravenous gadolinium-based contrast agents, and the ability to perform imaging without breath-holding – a significant benefit in pediatric MRI. Therefore, REACT may serve as a valuable alternative to contrast-enhanced angiography in clinical settings where the use of contrast agents is undesirable or unnecessary, although further studies in pediatric patients with vascular pathologies are warranted.

Conflict of Interest

The authors declare that they have no conflict of interest.

• References

• 1 Chung T. Magnetic resonance angiography of the body in pediatric patients: experience with a contrast-enhanced time-resolved technique. Pediatr Radiol 2005; 35: 3-10
• 2 Mathur M, Jones JR, Weinreb JC. Gadolinium Deposition and Nephrogenic Systemic Fibrosis: A Radiologist's Primer. Radiographics 2020; 40: 153-162
• 3 McDonald JS, McDonald RJ, Jentoft ME. et al. Intracranial Gadolinium Deposition Following Gadodiamide-Enhanced Magnetic Resonance Imaging in Pediatric Patients: A Case-Control Study. JAMA Pediatr 2017; 171: 705-707
• 4 McDonald RJ, Levine D, Weinreb J. et al. Gadolinium Retention: A Research Roadmap from the 2018 NIH/ACR/RSNA Workshop on Gadolinium Chelates. Radiology 2018; 289: 517-534
• 5 Chavhan GB, Babyn PS, John P. et al. Pediatric Body MR Angiography: Principles, Techniques, and Current Status in Body Imaging. AJR Am J Roentgenol 2015; 205: 173-184
• 6 Yoneyama M, Zhang S, Hu HH. et al. Free-breathing non-contrast-enhanced flow-independent MR angiography using magnetization-prepared 3D non-balanced dual-echo Dixon method: A feasibility study at 3 Tesla. Magn Reson Imaging 2019; 63: 137-146
• 7 Andia ME, Henningsson M, Hussain T. et al. Flow-independent 3D whole-heart vessel wall imaging using an interleaved T2-preparation acquisition. Magn Reson Med 2013; 69: 150-157
• 8 Eggers H, Brendel B, Duijndam A. et al. Dual-echo Dixon imaging with flexible choice of echo times. Magn Reson Med 2011; 65: 96-107
• 9 Pennig L, Wagner A, Weiss K. et al. Imaging of the pulmonary vasculature in congenital heart disease without gadolinium contrast: Intraindividual comparison of a novel Compressed SENSE accelerated 3D modified REACT with 4D contrast-enhanced magnetic resonance angiography. J Cardiovasc Magn Reson 2020; 22: 8
• 10 Pennig L, Wagner A, Weiss K. et al. Comparison of a novel Compressed SENSE accelerated 3D modified relaxation-enhanced angiography without contrast and triggering with CE-MRA in imaging of the thoracic aorta. Int J Cardiovasc Imaging 2021; 37: 315-329
• 11 Betz LH, Dillman JR, Towbin AJ. et al. Respiratory-Triggered Flow-Independent Noncontrast Non-ECG-Gated MRV (REACT) Versus CE-MRV for Central Venous Evaluation in Children and Young Adults: A Six-Reader Study. AJR Am J Roentgenol 2023; 221: 240-248
• 12 Isaak A, Luetkens JA, Faron A. et al. Free-breathing non-contrast flow-independent cardiovascular magnetic resonance angiography using cardiac gated, magnetization-prepared 3D Dixon method: assessment of thoracic vasculature in congenital heart disease. J Cardiovasc Magn Reson 2021; 23: 91
• 13 Isaak A, Mesropyan N, Hart C. et al. Non-contrast free-breathing 3D cardiovascular magnetic resonance angiography using REACT (relaxation-enhanced angiography without contrast) compared to contrast-enhanced steady-state magnetic resonance angiography in complex pediatric congenital heart disease at 3T. J Cardiovasc Magn Reson 2022; 24: 55
• 14 Hoyer UCI, Lennartz S, Abdullayev N. et al. Imaging of the extracranial internal carotid artery in acute ischemic stroke: assessment of stenosis, plaques, and image quality using relaxation-enhanced angiography without contrast and triggering (REACT). Quant Imaging Med Surg 2022; 12: 3640-3654
• 15 Janssen JP, Rose S, Kaya K. et al. Non-contrast-enhanced MR-angiography of Extracranial Arteries in Acute Ischemic Stroke at 1.5 Tesla Using Relaxation-Enhanced Angiography Without Contrast and Triggering (REACT). Clin Neuroradiol 2024;
• 16 Tan EJ, Zhang S, Tirukonda P. et al. REACT – A novel flow-independent non-gated non-contrast MR angiography technique using magnetization-prepared 3D non-balanced dual-echo dixon method: Preliminary clinical experience. Eur J Radiol Open 2020; 7: 100238
• 17 Dillman JR, Trout AT, Merrow AC. et al. Non-contrast three-dimensional gradient recalled echo Dixon-based magnetic resonance angiography/venography in children. Pediatr Radiol 2019; 49: 407-414
• 18 Spogis J, Katemann C, Zhang S. et al. Feasibility and Implementation of a 4D Free-Breathing Variable Density Stack-of-Stars Functional Magnetic Resonance Urography in Young Children Without Sedation. Invest Radiol 2024; 59: 271-277
• 19 Chandarana H, Feng L, Block TK. et al. Free-Breathing Contrast-Enhanced Multiphase MRI of the Liver Using a Combination of Compressed Sensing, Parallel Imaging, and Golden-Angle Radial Sampling. Investigative Radiology 2013; 48: 10-16
• 20 Endler CH-J, Kukuk GM, Peeters JM. et al. Dynamic Liver Magnetic Resonance Imaging During Free Breathing: A Feasibility Study With a Motion Compensated Variable Density Radial Acquisition and a Viewsharing High-Pass Filtering Reconstruction. Investigative Radiology 2022; 57: 470-477
• 21 Holm S. A Simple Sequentially Rejective Multiple Test Procedure. Scandinavian Journal of Statistics 1979; 6: 65-70
• 22 Forbes-Amrhein MM, Dillman JR, Trout AT. et al. Frequency and Severity of Acute Allergic-Like Reactions to Intravenously Administered Gadolinium-Based Contrast Media in Children. Invest Radiol 2018; 53: 313-318
• 23 Brown EG, Adkins ES, Mattei P. et al. Evaluation of Image-Defined Risk Factor (IDRF) Assessment in Patients With Intermediate-risk Neuroblastoma: A Report From the Children's Oncology Group Study ANBL0531. J Pediatr Surg 2024; 161896
• 24 Harder FN, Budjan J, Nickel MD. et al. Intraindividual Comparison of Compressed Sensing-Accelerated Cartesian and Radial Arterial Phase Imaging of the Liver in an Experimental Tumor Model. Investigative Radiology 2021; 56: 433-441
• 25 Yoon JH, Lee JM, Yu MH. et al. Evaluation of Transient Motion During Gadoxetic Acid–Enhanced Multiphasic Liver Magnetic Resonance Imaging Using Free-Breathing Golden-Angle Radial Sparse Parallel Magnetic Resonance Imaging. Investigative Radiology 2018; 53: 52-61
• 26 Janssen JP, Gertz RJ, Tristram J. et al. Accelerating non-contrast MR angiography of the thoracic aorta using compressed SENSE with deep learning reconstruction. Eur J Radiol 2025; 192: 112403
• 27 Erdem S, Jack A, Doctor P. et al. Feasibility of magnetization-transfer-contrast relaxation-enhanced angiography without contrast and triggering (REACT) imaging at 1.5 T combined with deep learning-based reconstruction for cardiovascular visualization. Quant Imaging Med Surg 2025; 15: 3222-3236

Publication History

Received: 04 June 2025

Accepted after revision: 02 January 2026

Article published online:
30 January 2026