Triggered Non-Contrast-Enhanced MR Angiography — Cardiac-Gated Subtraction, Inflow-Inversion and Quiescent-Interval Techniques (TRANCE, FBI, NATIVE, Inhance Deltaflow, QISS and Equivalents)

Triggered Non-Contrast-Enhanced MR Angiography — Cardiac-Gated Subtraction, Inflow-Inversion and Quiescent-Interval Techniques (TRANCE, FBI, NATIVE, Inhance Deltaflow, QISS and Equivalents)

MRIninja Knowledge Base | Sequence/Technique Page Builds on: MRI Sequences Overview · Flow, Motion and Diffusion Physics · Stationary Tissue vs. Flowing Fluid · Controlling Fluid Signal · 2D vs 3D vs MOTSA Version 1.0 — October 2026

1. Introduction and Scope

"TRANCE" (TRiggered Angiography Non-Contrast-Enhanced) is the Philips product name for one specific member of a much larger family of angiographic techniques that share a single organising idea: the arterial pulse is used as the contrast mechanism, in place of an injected contrast agent. Data acquisition is synchronised to the cardiac cycle (ECG or peripheral pulse trigger), so that the flow state of arterial blood at the moment of readout is known and controllable. Every vendor offers one or more members of this family under a different name — Siemens NATIVE SPACE / NATIVE TrueFISP / QISS, Canon Fresh Blood Imaging (FBI) / Time-SLIP, GE Inhance 3D Deltaflow / Inhance Inflow IR, Hitachi / Fujifilm VASC-FSE / VASC-ASL — which is why this page is organised by mechanism first and by vendor name second [1].

This page assumes as already known, and does not re-derive, the flow-related enhancement and time-of-flight washout physics of Stationary Tissue vs. Flowing Fluid, the parameter-level levers of Controlling Fluid Signal (including presaturation and black-blood preparation), and the 2D / 3D / multi-slab inflow comparison of 2D vs 3D vs MOTSA. The untriggered inflow technique (time-of-flight, treated on the Time-of-Flight MRA page) and phase-contrast angiography are the other two classical non-contrast methods; phase-contrast is treated on the Phase-Contrast MRI page, the underlying physics on the Flow, Motion and Diffusion Physics page, and both are used here only as comparators.

Clinical rationale, stated neutrally. Triggered NCE-MRA is attractive wherever gadolinium-based contrast is undesirable or unnecessary (impaired renal function, pregnancy, paediatric patients, repeated follow-up, patient preference) or where contrast-enhanced MRA is technically limited [1,4]. This page does not argue that these techniques replace contrast-enhanced MRA or CT angiography; the evidence base (Section 11) does not support such a general claim, and the available data are predominantly single-centre diagnostic-accuracy studies against a contrast-enhanced or digital-subtraction reference.

Information status of this page. The vendor product names in Section 8 are taken from the published review by Edelman and Koktzoglou [1] and from Siemens technical documentation [5]. Numerical sequence parameters are quoted only where a source reports them, and are attributed to that source and to its field strength and platform; they are examples of reported practice, not recommended settings. Where a statement is general physical reasoning not drawn from a specific cited source, it is labelled as such.

2. Taxonomy — Where Triggered Techniques Sit Among Non-Contrast MRA Methods

Non-contrast MRA methods differ in what physical property of flowing blood they convert into signal. The triggered methods are those in which the cardiac cycle is part of that conversion.

Two points of classification matter for everything that follows. First, subtractive techniques (cardiac-gated 3D FSE, FSD-bSSFP, ASL) rely on registration between two acquisitions and therefore inherit a vulnerability to motion between them; non-subtractive techniques (inflow-IR bSSFP, QISS) do not. Second, techniques differ in whether the flow dependence is velocity-based (systolic dephasing) or inflow-based (replacement of saturated or inverted spins); this decides which vessels and which flow states they fail in (Section 10).

3. Why Trigger — the Arterial Pulse as a Contrast Mechanism

3.1 Arterial Flow Is Not Constant Within One Heartbeat

The prerequisite physics page treats flow as a steady quantity. In the extremity arteries it is not: flow is fast during systole and much slower during diastole, whereas venous signal stays relatively constant between systole and diastole [6]. That difference between arterial systolic and diastolic flow, and the absence of the same difference in veins and static tissue, is the entire basis of cardiac-gated 3D FSE subtraction angiography.

3.2 How a Fast Spin-Echo Train Converts Fast Flow Into Signal Loss

A fast (turbo) spin-echo train forms each echo by refocusing transverse magnetisation with a 180° pulse. A spin that has moved between excitation and refocusing, or that has accumulated flow-dependent phase in gradients that are not velocity-compensated, is not fully refocused, and its signal is lost; this is the flow void familiar from routine T2-weighted imaging. Fast systolic arterial flow produces a pronounced flow void; slow diastolic flow does not [6,10]. Blood has a long T2, so when flow is slow enough to avoid dephasing, arterial lumen signal is high on a T2-weighted FSE acquisition.

The published description of the technique therefore reads as follows [6,10]: acquire one matched 3D data set at diastole, where arterial and venous signal are both high, and one at systole, where arterial signal is largely lost while venous and background signal remain essentially unchanged; subtract the systolic set from the diastolic set so that static tissue and veins cancel and predominantly arterial signal remains. The dependence on readout direction is a consequence of the same physics: motion sensitivity is highest when the artery runs along the readout gradient, because of endogenous spoiling from alternating echoes; additional flow-spoiling gradients can be added to accentuate systolic signal loss, while flow-compensating gradients can be used in large central arteries where diastolic stasis is small, in order to boost diastolic signal [6]. Miyazaki and colleagues demonstrated the separation of arteries from veins with flow-spoiled gradient pulses in an ECG-triggered 3D half-Fourier FSE sequence in 15 healthy volunteers: veins stay bright throughout the cardiac cycle and are removed by subtracting the diastolic bright-blood from the systolic black-blood images, and stronger flow-spoiling improved depiction of slow-flowing arteries, including in the distal foot and hand [7].

3.3 Trigger Source: ECG vs Peripheral Pulse

Siemens documentation states that, for NATIVE SPACE, ECG triggering is essential because data must be collected at consistent points of the cardiac cycle [5]; a peripheral pulse unit (PPU) has been used instead on a Philips 1.5 T system in a published TRANCE protocol [15], and the educational literature describes gating as "ECG or peripheral pulse" [6]. General physiological reasoning, not drawn from the cited sources: a peripheral pulse sensor registers the pulse wave later than the R-wave of the ECG, so a delay measured with one trigger source cannot be transferred to the other, and the pulse-wave delay differs between the proximal and the distal leg.

4. Cardiac-Gated 3D Fast Spin-Echo Subtraction MRA (TRANCE, FBI, NATIVE SPACE, Inhance 3D Deltaflow, VASC-FSE)

4.1 Sequence Design

All members of this family are T2-weighted 3D FSE, turbo spin-echo or half-Fourier FSE acquisitions, usually accelerated with partial Fourier and parallel imaging [6]. Siemens describes NATIVE SPACE as a flow-sensitive 3D turbo spin-echo sequence with non-selective refocusing pulses and ultra-short echo spacing, whose image contrast "is based on the pulsatility of the blood flow" [5]. Because the contrast is a systole-minus-diastole difference, fat suppression is not technically required (subtraction removes static fat), although STIR or SPAIR may optionally be applied to the source images [6].

The reported TR values look different ("2700 ms" versus "1 beat" versus "3–4 × R–R") because they describe different ways of expressing the same constraint: TR is tied to the cardiac cycle, so that each excitation falls at the same trigger delay. General reasoning, not drawn from the cited sources: because each shot is tied to the cardiac cycle, acquisition time scales with heart rate and with the number of shots; empirically, reference [15] reports a total protocol time of 60 minutes (35 minutes arterial, 25 minutes venous) in a patient cohort, and Edelman and Koktzoglou list approximately 12 minutes of acquisition for the cardiac-gated 3D FSE family [1], so published times vary widely with coverage, resolution and heart rate.

4.2 Trigger-Delay Determination — the Step That Makes or Breaks the Examination

The two delays (systolic and diastolic) must be set per vessel region, because the timing of peak flow differs between proximal and distal vessels and can differ between the two legs in the presence of unilateral disease [1,6]. Each vendor provides a timing scout:

The sources agree on the principle: systolic delay = maximal flow void; diastolic delay = maximal arterial brightness. Practical reasoning, not drawn from the cited sources: a wrong delay does not make the subtraction fail loudly; it yields an angiogram with weak arterial signal, so the timing scout is the first thing to inspect when an examination is disappointing.

4.3 Subtraction and Output

Siemens NATIVE SPACE returns a "Peak Flow" image (dark arteries), a "Min Flow" image (bright arteries), an inline subtraction and an inline coronal MIP; two acquisitions at different trigger delays are subtracted voxel-wise to remove background and venous signal [5]. FBI processing produces coronal MIPs stitched across stations, with semi-automated analysis software (FBI-Navi) [10].

4.4 A Variant: TRANCE for the Veins

The same triggering logic can be inverted. In the venous TRANCE protocol of reference [15], a 3D TSE-STIR sequence is acquired with systolic triggering so that arterial blood is dephased and black, leaving a veins-only data set; STIR additionally suppresses background fat and bone. In that single-centre cohort of 25 patients, sensitivity was 85.7%, specificity 88.9% and accuracy 88%, with substantial agreement with ultrasonography for thigh deep-vein thrombosis (κ = 0.72) [15]. The authors list a small non-randomised sample, absent inter-observer analysis, insufficient validation against other modalities, possible false positives in the left common iliac vein (especially in the elderly) and a protocol length that is unsuitable for critically ill or agitated patients [15].

4.5 Evidence for the Peripheral Arterial Application

These two studies illustrate the consistent pattern of the literature: good-to-excellent performance in the thigh, degrading in the calf and feet, where flow is slower and the systolic-diastolic difference smaller [1,9,10]. Lim and colleagues' early clinical experience (36 patients, calf and pedal arteries, gadolinium-enhanced consensus reference) reported 85.4% sensitivity, 75.8% specificity and 79.4% accuracy for stenosis; serious artefacts — including motion, trigger-delay errors, linear artefact and vessel blurring — reduced diagnostic confidence in 47.2% of patients, and accuracy rose to 92.2% when only cases with adequate confidence were analysed [8]. This is direct evidence that trigger-delay quality (Section 4.2) is a major practical failure point.

5. Flow-Sensitive Dephasing (FSD) — the bSSFP Variant of Gated Subtraction

Flow-sensitive dephasing keeps the systolic/diastolic subtraction logic but replaces the FSE readout with a balanced SSFP readout; dephasing gradients along the vessel length create the flow-dependent systolic signal loss [1]. Edelman and Koktzoglou list its primary use as calf and foot imaging, with an acquisition time of about 5 minutes [1]. The motivation for a bSSFP readout is the high intrinsic blood signal and short acquisition time of that readout; the price is the characteristic bSSFP sensitivity to off-resonance (banding), treated in Section 10 [1].

6. Triggered Inflow Inversion-Recovery bSSFP (NATIVE TrueFISP, B-TRANCE, Inhance Inflow IR, Time-SLIP)

6.1 Mechanism

The physics of the bSSFP readout itself — T2/T1 contrast, the transient phase and off-resonance banding — is developed on the bSSFP sequence page; Philips names this readout "balanced FFE" or "b-TFE" (B-TRANCE is the Philips product name for the inversion-prepared angiographic use). Inversion geometry, TI, respiratory strategy, acceleration and the renal evidence are developed in the inflow-IR bSSFP deep dive.

This family is not subtractive and not based on systolic flow void. A slab-selective (typically thick axial) inversion pulse inverts the stationary tissue and venous blood within the region of interest; after an inversion time TI, fresh non-inverted arterial blood that has flowed into the region appears bright on a 3D bSSFP readout, while the inverted background is suppressed; fat is suppressed chemically [1,5,11]. Siemens describes it in exactly these terms: an inversion pulse applied to the region of interest, fresh non-inverted blood flowing into the inverted slab appearing bright [5]. Canon's term Time-SLIP refers to a time-spatial labelling inversion pulse [1,4].

6.2 Triggering, Respiration and TI

• ECG triggering is "highly recommended" for NATIVE TrueFISP, although data can be acquired without it; with ECG, a new inversion is played out per trigger [5].
• Respiratory gating uses a 1D PACE navigator (alternatively a respiratory cushion) [5]; Edelman and Koktzoglou state that respiratory gating is typically used and that irregular breathing degrades image quality [1].
• TI must be long enough to always contain an inflow episode; the Siemens documentation gives "1200 to 1400 ms" as an example when running without ECG. When 1400 ms exceeds what a single R–R interval allows, every second trigger is used while the trigger-pulse parameter stays at 1; the upper limit of the inversion band should match the upper limit of the FOV; acquisition time is usually 2–5 minutes [5]. Edelman and Koktzoglou report inflow times of up to 1400 ms and scan times of about 5 minutes or less [1].

6.3 Evidence and Limits

The principal prospective validation is Wyttenbach et al. (53 patients with hypertension and suspected renal artery stenosis, 1.5 T; slab-selective inversion-prepared 3D SSFP versus breath-hold contrast-enhanced MRA): per artery, sensitivity 100% / 95%, specificity 93% / 95%, NPV 100% / 99% for the two readers; stenosis was overestimated (six and four false positives), and mean maximal visible vessel length was shorter with SSFP (61.1 mm versus 69.9 mm; P < .001) [11]. The same technique has been described in renal transplants (usually without respiratory gating) and in patients with reduced stroke volume or dehydration [5]. Limitations [1]: irregular breathing; slow flow or severe distal stenosis limits visible vessel length; and the technique is generally ineffective for peripheral arteries, which is precisely where Section 4 and Section 7 techniques are used.

7. Quiescent-Interval Single-Shot (QISS) MRA

7.1 Sequence Structure

QISS is an ECG-triggered, 2D, slice-by-slice technique. Within each cardiac cycle, saturation pulses are applied, followed by a quiescent interval during which unsaturated arterial spins flow into the imaging plane, followed by a single-shot 2D bSSFP readout [1,13]. One saturation pulse suppresses background signal, and a second, placed inferior to the slice in the original lower-limb implementation, suppresses venous inflow, so that the lumen signal comes from arterial blood that arrived during the quiescent interval. Edelman and Koktzoglou report approximately 300 ms of data collection per slice (making the sequence relatively insensitive to motion), a typical protocol of 40–50 axial 3-mm slices per station, and about 400 slices (10 stations) for a whole-leg examination in roughly 7–8 minutes, shimming excluded [1].

Compare this with the prerequisite physics: QISS is a deliberate engineering of the 2D inflow case described in 2D vs 3D vs MOTSA — each slice is a thin cross-section into which blood renews itself — with the quiescent interval guaranteeing time for that renewal and the triggering guaranteeing that renewal is timed to the same cardiac phase for every slice. Because there is no 3D slab, there is no progressive saturation with depth to correct with a ramp pulse.

7.2 Evidence

Hodnett et al. studied 53 patients with suspected or known chronic peripheral arterial disease in a two-centre prospective design, with contrast-enhanced MRA as reference: per-segment sensitivity was 89.7% (436/486) and 87.0% (423/486), and specificity 96.5% (994/1030) and 94.6% (973/1028), for the two reviewers [12]. The original technical-feasibility report is Edelman et al. 2010 [13], and a direct comparison of QISS with NATIVE SPACE was published by Ward et al. in 2013 [14].

7.3 Limitations

Edelman and Koktzoglou list: a minimum slice thickness of about 2 mm; a sinc-shaped slice profile (less sharp multiplanar reconstructions than 3D techniques); and greater sensitivity to off-resonance than fast spin-echo [1]. General physical reasoning, consistent with the in-plane saturation effect developed on the prerequisite pages: a vessel running within the slice plane (rather than crossing it) will tend to lose signal, which is one reason for acquiring axial slices along the long axis of the limb.

8. Vendor Nomenclature — One Mechanism, Several Names

Source: Table 1 of Edelman and Koktzoglou [1], as retrieved for this page. The Hitachi entry of the ASL-based row was retrieved as "VASL-ASL" while the inflow-IR row reads "VASC-ASL"; one of these may be a transcription difference and both should be checked against vendor documentation. Additional vendor names appear in secondary educational sources (for example Canon "CIA" and Fujifilm "VASC-FSE") [6]; the Hitachi / Fujifilm attribution of these names differs between sources and was not verified. Empty cells mean that [1] lists no commercial equivalent, not that no equivalent exists. Console-level parameter names (the exact labels of the delay, shot, flow-compensation and navigator fields on each platform) change between software releases and are intentionally not enumerated here.

9. ASL-Based Angiography (4D-TRANCE, Silent MRA, mASTAR, VASC-ASL)

ASL-based MRA is the subtractive member of the inflow family: inflowing arterial spins are labelled with RF pulses, and tagged and untagged images are subtracted, analogous to digital subtraction angiography; it is well suited to fast-flowing vessels such as the extracranial carotids, takes about 5 minutes, and is vulnerable to misregistration if the patient moves between the two acquisitions [1]. The labelling physics is developed on the Arterial Spin Labeling master and the ASL labeling parameters page and is not repeated here. The angiographic strategy of ASL (time axis, pulsed versus pseudo-continuous labeling, Philips 4D-TRANCE versus 4D-PACK) is developed in the ASL-based non-contrast MRA deep dive.

10. Pitfalls, Failure Modes and Artefacts

11. Technique Selection by Territory

This section distils the published pattern; it is guidance derived from the evidence in Sections 4–7, not a guideline recommendation (see Section 12).

Choice between gated-FSE and QISS in the periphery. Gated 3D FSE provides 3D data (with sharper multiplanar reformats than QISS [1]) but depends on accurate per-region timing, is sensitive to motion between systolic and diastolic acquisitions and takes minutes per slab; QISS is slice-by-slice, motion-tolerant per slice (about 300 ms per readout) (how QISS handles trigger delay was not documented in the retrieved sources), but has a sinc-shaped slice profile, a minimum slice thickness of about 2 mm and greater off-resonance sensitivity [1]. Ward et al. compared the two directly [14].

12. Evidence Gaps and Ongoing Debate

• No guideline statement located. A search for a society guideline recommendation specifically endorsing a triggered NCE-MRA technique for peripheral or renal imaging returned no verifiable text; the 2024 ESC key-points summary consulted contains no statement on lower-extremity MRA technique. This page therefore makes no guideline-level claim, and the absence of such a recommendation should not be read as evidence either way.
• Evidence level. The studies cited here are single-centre or two-centre diagnostic-accuracy studies with small to moderate sample sizes, mostly at 1.5 T, against contrast-enhanced MRA or DSA; the reference standards differ between studies, and no systematic review or meta-analysis was verified for this page.
• Site dependence. The consistent loss of accuracy from thigh to calf to foot [1,9,10] is the main applicability limit and should be stated in any protocol decision.
• Field strength. Most of the cited primary data derive from 1.5 T; Edelman and Koktzoglou note that early QISS studies were mostly at 1.5 T, and 3 T results were also reported [1,13]. General physical reasoning: bSSFP off-resonance sensitivity and SAR of high-flip-angle FSE both worsen with increasing field strength.
• Full TRANCE methods. The Methods section of the original TRANCE-versus-DSA study [9] was not accessible for this page; TRANCE parameters are quoted only from the venous protocol [15].
• Vendor naming. See Section 8 for unverified entries.

13. Evidence-Based References

No official guideline, consensus statement, systematic review or meta-analysis was verified for this specific topic. References are listed by evidence tier and then by recency.

C. Important Prospective / Original Studies

D. Technical MRI Papers and Reviews

End of document — Triggered Non-Contrast-Enhanced MR Angiography — MRIninja v1.0 — October 2026