ASL-Based Non-Contrast MR Angiography — Labeling Strategies (PASL, pCASL, Time-Encoded, Superselective), Time-Resolved Readouts, Philips 4D-TRANCE and pCASL-Based 4D-PACK / 4D-S-PACK

ASL-Based Non-Contrast MR Angiography — Labeling Strategies (PASL, pCASL, Time-Encoded, Superselective), Time-Resolved Readouts, Philips 4D-TRANCE and pCASL-Based 4D-PACK / 4D-S-PACK

MRIninja Knowledge Base | Sequence/Technique Deep Dive Builds on: ASL Labeling Scheme Techniques and Parameters · Triggered Non-Contrast-Enhanced MRA (Section 9) · ASL Generic Standard Protocol Version 1.0 — October 2026

1. Introduction and Scope

Two pages already cover parts of this subject. The ASL labeling parameters page (Section 6) documents a Philips 4D-TRANCE console protocol parameter by parameter, and the triggered NCE-MRA page (Section 9) places ASL-based MRA in the taxonomy of non-contrast angiography. This page does not repeat either. It addresses how ASL is turned into an angiographic technique: how the angiographic use differs from perfusion ASL, how time resolution is created, how the labeling families (pulsed, pseudo-continuous, time-encoded, superselective) behave in angiography, and — as a focused section — what distinguishes the Philips pCASL-based angiographic techniques (4D-PACK, 4D-S-PACK, 4D-APACK) from the pulsed 4D-TRANCE.

Information status. The clinical literature on ASL angiography is dominated by single-centre studies, conference abstracts and one narrative review. No guideline or consensus statement specific to ASL angiography was found. Parameters are quoted only where a source reports them, with platform and field strength, as examples of reported practice and not as recommended settings.

2. Principle — Perfusion ASL and Angiographic ASL Use the Same Labeling Physics for Opposite Goals

ASL-based MRA is a subtraction technique: inflowing arterial blood is labelled with RF pulses, a control image is acquired without labeling, and the label–control difference leaves signal only from labelled blood, in the manner of digital subtraction angiography [1,2]. The labeling physics is common to perfusion ASL; the use of the signal is opposite.

Reasoning, not a quoted statement: a perfusion protocol and an angiographic protocol of the same labeling scheme are therefore tuned in opposite directions — long label and delay versus short label or delay and several sampled phases — which is why the console parameters of the two differ so markedly on the ASL labeling parameters page.

3. Labeling Families Used in Angiography

The Gomyo review lists the vendor attribution as Philips for 4D-TRANCE (PASL) and for 4D-PACK and 4D-S-PACK (pCASL), and Canon for mASTAR (PASL) and its ultrashort-TE and high-temporal-resolution variants [1]. The Philips 4D-TRANCE product page describes it as "a time-resolved technique for non-contrast angiography" with temporal resolution "down to 160 msec" and does not name the labeling scheme [10]; the PASL attribution therefore comes from the review [1] and from the site's own console review of a "STAR Angio" protocol on the labeling parameters page. The ISMRM comparison of 4D-TRANCE with 4D-SNAP likewise states that the labeling pulse of 4D-TRANCE is not specified in its abstract [3]. Time-encoded pCASL is described in the review only as under development for simultaneous perfusion and angiography and not clinically established [1]. Silent MRA (GE) and VASC-ASL (Hitachi) are listed as ASL-based on the vendor-name table of the triggered NCE-MRA page [2] and are not developed here, because no source describing their labeling scheme was verified for this page.

4. How Time Resolution Is Created

4.1 Varying the Delay After a Single Label (PASL, 4D-TRANCE)

In the Philips 4D-TRANCE / ASL comparison abstract, 4D-TRANCE "varies the post-labeling delay", with a first PLD of 200 ms and an interval of 150 ms over 8 dynamic phases, using a 3D turbo field echo readout (TFE factor 23) or a 3D TFE-EPI readout (TFE factor 11, EPI factor 5) at 3 T on a Philips Ingenia with a 32-channel head coil [3]. The authors reported pulsation artefacts with the TFE-EPI readout and vessels fading in long-PLD images, and noted that ECG gating could reduce pulsation but would lengthen acquisition [3]. The Gomyo review explains the late-phase fading as signal decay with PASL due to RF saturation [1].

4.2 Varying the Label Duration with Keyhole and View-Sharing (pCASL, 4D-PACK)

4D-PACK combines pseudo-continuous labeling with a CENTRA-keyhole acquisition [5,6]. In the Moyamoya abstract, the label durations were 100, 200, 400, 600, 800, 1200, 1600 and 2200 ms, each phase including label and control images; the 100 ms data served as the reference for the first five phases and the 2200 ms data for the last three, with 75% keyhole and view-sharing [5]. The first abstract describing 4D-PACK states that the keyhole technique "can accelerate acquisition speed, while keeping flow dynamics information" in three volunteers [6]. Reasoning, not a quoted statement: in this scheme each time point corresponds to a bolus of different length, not a different delay, and the keyhole and view-sharing allow the central k-space to be refreshed more often than the periphery, which is what reduces the scan time.

4.3 Time-Encoded pCASL (Hadamard)

In time-encoded pCASL the labeling module is split into N segments whose label or control state changes across N+1 acquisitions following a Hadamard matrix; decoding yields N images with different labeling durations and delays [8]. In the Philips implementation by Suzuki and colleagues, an order-8 matrix was used, with a first sub-bolus of 1800 ms optimised for perfusion and six sub-boli of 130 ms for encoding the passage of the label in the angiogram, with a 3D TFE-EPI angiographic readout followed by a single-shot multislice EPI perfusion readout, at 3 T in six healthy volunteers [8]. The angiographic flip angle was a trade-off: 5° and 7.5° gave no significant CBF overestimation after correction, whereas 10° and above did, and higher flip angles produced ghosting of residual arterial signal; 7.5° was judged the best compromise [8]. Perfusion temporal SNR was significantly lower at all tested flip angles than with conventional time-encoded pCASL [8].

5. Focus — Philips pCASL and Its Distinguishing Characteristics

5.1 What the Verified Sources Establish for Philips

• Perfusion pCASL on the Philips console. The reviewed Philips protocol (documented on the labeling parameters page) exposes label distance (90 mm in the reviewed protocol) and post-label delay (1800 ms), and no separate editable labeling duration field in that protocol. This is a console observation from the site's own review of one protocol and is not a vendor specification.
• Angiographic pCASL implementations. 4D-PACK (pCASL with CENTRA-keyhole and view-sharing), 4D-S-PACK (superselective version) and 4D-APACK (4D-PACK combined with acceleration-selective ASL) are reported on Philips Ingenia 3 T systems with Philips-affiliated authors [5,6,7,9,11].
• Philips 4D-TRANCE is, per the review [1], PASL-based, so on this platform the pulsed and pseudo-continuous angiographic approaches coexist.

5.2 Characteristics of pCASL Relative to Pulsed Labeling

The consensus [4] concerns perfusion imaging and is cited here only for the physics of the labeling schemes; it is not an angiography guideline.

5.3 Reported Philips pCASL Angiographic Parameters

Reasoning, not a quoted statement: the 80 mm slab in both abstracts implies limited craniocaudal coverage, consistent with intracranial rather than extracranial or body use.

5.4 Superselective Labeling (4D-S-PACK)

In 4D-S-PACK, standard pCASL gradients along the flow direction are supplemented with perpendicular Gx and Gy gradients so that labeling is restricted to a circular or elliptical spot placed over the proximal ICA or ECA [9]. In a retrospective single-centre study of 21 patients with intracranial dural arteriovenous fistulas (DSA reference), feeding arteries were detected in 80.0% and 76.8% (two observers) of the 95 seen on DSA with 4D-S-PACK, against 41.1% and 48.4% with non-selective 4D-PACK; shunt location was identified in all 21 patients with both methods; cortical venous reflux sensitivity and specificity were 90.9% for both; Borden classification was correct in 20 of 21 with both [7]. The authors list as limitations the small sample, motion sensitivity of a small labeling spot, harder ECA labeling, faint signal from unlabelled vessels, operator training for labeling-area planning, and an acquisition of about 20 minutes in total; three authors were Philips employees [7]. Lower CNR with 4D-S-PACK was reported in the earlier seven-patient abstract (feeding arteries 47.7 ± 12.8 versus 58.3 ± 15.3; draining veins 48.7 ± 18.5 versus 71.9 ± 24.7) [9]. The two feeding-artery sensitivities (71.0% in the abstract, about 80% in the paper) come from different cohorts and are not necessarily comparable.

5.5 Slow Flow and Acceleration-Selective ASL (4D-APACK)

The 4D-APACK abstract proposes replacing the last time point of 4D-PACK with AccASL to visualise very slow flow independently of arterial transit time; in six healthy volunteers and one AVM patient, transit-time correlation with 4D-pCASL was R = 0.971, MCA M1–M4 CNR did not differ significantly, and the AVM nidus and feeding artery were visible at the AccASL time point [11]. This is a methodological abstract and not a clinical validation.

6. Evidence Summary

For extracranial arteries, Edelman and Koktzoglou describe ASL-based MRA as well suited to fast-flowing vessels such as the carotids, taking about 5 minutes, and vulnerable to misregistration between acquisitions [2]. No study retrieved for this page reports a head-to-head diagnostic accuracy trial of Philips 4D-TRANCE against DSA or CE-MRA.

7. Pitfalls and Failure Modes

8. Choosing Between Approaches (Reasoned Synthesis)

Reasoning, not a published algorithm: the following follows from the mechanisms and limitations above and from the single-centre evidence, and should not be read as a guideline.

• Short examination time and hemodynamic phases of fast-filling arteries: a pulsed approach such as 4D-TRANCE, accepting late-phase signal loss.
• Distal vessels, collaterals and slow flow (the Moyamoya setting): pCASL-based 4D-PACK, whose later phases retain signal [5].
• Identification of the feeding artery of a shunt: superselective labeling (4D-S-PACK), accepting longer time and operator-dependent planning [7].
• Very slow flow independent of transit time: AccASL addition (4D-APACK), still methodological [11].
• Fine anatomical detail of the arterial lumen: not the strength of ASL-MRA; the review states lower spatial resolution than TOF-MRA [1].

9. Evidence Gaps and Ongoing Debate

• No guideline or consensus statement for ASL angiography was found. The pCASL recommendation [4] applies to perfusion imaging.
• The labeling scheme of Philips 4D-TRANCE is not stated by the manufacturer page or the ISMRM comparison abstract [3,10]; the PASL attribution rests on a review [1] and on the console review on the labeling parameters page.
• Prospective diagnostic-accuracy trials are lacking. The accuracy figures come from retrospective single-centre work with small samples [7], and several techniques are described only in conference abstracts.
• Vendor-affiliated authors. The 4D-PACK family papers and abstracts involve Philips employees [6,7,8,9,11].
• Parameter values for 4D-APACK, and the 4D-TRANCE parameter tables of the comparison abstract beyond those quoted, were not available in the retrieved text.
• The review (Gomyo 2025) was read through its abstract and a processed summary of the accessible page, not as a full text. Quantitative statements taken from it (for example the temporal-resolution figures) are attributed to it and were not independently checked against the primary studies it cites.
• Extracranial and body ASL-MRA is only summarised from a single review [2]; no primary study was verified for this page.

10. Evidence-Based References

No official guideline, consensus statement specific to ASL angiography, systematic review or meta-analysis was verified. References are listed by evidence tier and recency.

A. Consensus Statement (perfusion ASL; physics of labeling only)

C. Original Studies

Conference Abstracts (vendor-affiliated or abstract-only)

D. Reviews and Technical Sources

End of document — ASL-Based Non-Contrast MR Angiography — MRIninja v1.0 — October 2026