Facial ultrasound training is being sold to aesthetic clinicians the way cannulas were sold a decade ago: as the thing that closes the gap between where you think your product is going and where it actually goes. The pitch is not fraudulent. High-frequency facial ultrasound does things no other tool in the room can do. But the gap between what it has been shown to do and what injectors are being told it does is wide, and it is worth measuring carefully before you spend the money.
Melissa Pulcini-Buttine, PA, gives it about two sentences when she teaches complications, and both of them are load-bearing. "Ultrasound can be something that we could add to help with safety. The issue is ultrasound is very difficult in the face, and takes not just one certification class but a lot of practice with it. So we can't rely on anything to be 100% safe — but it can definitely help if you're very well skilled with ultrasound."
She prefaces it with the phrase she uses as a general rule: "preference is not greater than proof." That is the correct test to apply here, so this article applies it. What follows is what the literature actually supports as of September 2026, what it does not, and how to decide whether this belongs in your career.
What ultrasound has unambiguously proved
The strongest contribution facial ultrasound has made is not procedural. It is epistemic. It has demonstrated, in living patients rather than cadavers, that the arterial maps injectors memorise describe a minority of faces.
A systematic review and meta-analysis of Doppler studies of the facial artery pooled ten studies from 1,087 screened records and found that, relative to the nasolabial fold, the facial artery ran medial in 46.2%, crossed medial-to-lateral in 22.5%, and ran lateral in 12.0% of cases (Pourani MR, Ebrahimzade M, Goudarzi E, et al. Evaluation of Facial Artery Course Variations, Diameters, and Depth Using Doppler Ultrasonography: A Systematic Review and Meta-Analysis. J Cosmet Dermatol. 2025;24(9):e70431. PMID 40874402). The terminal branch was the angular artery in 71.8%, the lateral nasal in 27.9% and the superior labial in 5.7%.
Fewer than half of faces follow the modal course. The atlas is not wrong; it is a mode being taught as a fact.
A Doppler study of normal variants went further and found arteries missing: the transverse facial artery was absent in 31.39% of subjects and the angular artery in 11.62%, and the two sides of the same face differed significantly in facial artery depth at two of three measured levels (Khorasanizadeh F, Delazar S, Gheidari O, et al. Anatomic evaluation of the normal variants of the arteries of face using color Doppler ultrasonography. J Cosmet Dermatol. 2023;22(6):1844-1851. PMID 36718833).
And in the region that matters most for the worst outcome, an ultrasound investigation of 50 subjects measured where the supratrochlear and supraorbital arteries change plane as they ascend the forehead. The mean was around 13–14 mm above the superior orbital rim — with a range of 4.0 to 27.0 mm (Cotofana S, Velthuis PJ, Alfertshofer M, et al. The Change of Plane of the Supratrochlear and Supraorbital Arteries in the Forehead — An Ultrasound-Based Investigation. Aesthet Surg J. 2021;41(11):NP1589-NP1598. PMID 33652475). The authors concluded that the superficial lower forehead should be avoided "due to the unpredictability and inconsistent presence of the central and paracentral arteries."
A mean of 14 mm with a range of 4 to 27 mm is the cleanest illustration in the entire literature of why an average is useless at the individual patient. That finding alone has changed how the glabella and lower forehead should be reasoned about, and it came from ultrasound.
So the honest first conclusion is this: ultrasound has already improved your safety, whether or not you ever buy a machine. It produced the variability data that should be changing your margins.
What it does during a live vascular event
The second area where the evidence is real, if thin, is management rather than prevention.
The most-cited comparison examined ultrasound-guided targeted hyaluronidase against conventional regional flooding in 39 patients with vascular adverse events. Targeted treatment used a mean of 122.5 ± 34 IU of hyaluronidase against 1,519.4 ± 1,137 IU for flooding (P = 0.028), and where flooding produced no improvement in 92.3% of cases, targeted treatment produced immediate improvement in 100% (Schelke LW, Velthuis PJ, Decates T, et al. Ultrasound-Guided Targeted vs Regional Flooding. Aesthet Surg J. 2023;43(1):86-96. PMID 35951759).
Those numbers are striking and they deserve a caveat printed in the same size. This is a retrospective series of 39 patients in which each patient served as their own sequential control — flooding was attempted first and targeted treatment followed only after failure. That design is inherently biased toward the second treatment, a limitation the authors themselves acknowledge.
The size of the whole evidence base is the number worth carrying away. A PRISMA systematic review of ultrasound-guided hyaluronidase for filler-induced arterial ischaemia screened 88 studies, included 9, and reported that the entire published world literature amounted to 83 cases (Azizi N, Tootoonchi N, Khorasanizadeh F, et al. Aesthet Surg J Open Forum. 2025;7:ojaf125. PMID 41200269). Level of evidence 3. The authors call for large-scale randomised studies, which is another way of saying none exist.
And a finding that should be quoted to anyone who describes ultrasound as a rule-out test: an international multicentre study of 100 clinically diagnosed vascular adverse events found no Doppler abnormality at all in 12% of them (Sigrist RMS, Gonzalez C, Schelke L, et al. Doppler Ultrasound Findings in Filler-Related Facial Vascular Adverse Events. Diagnostics. 2026;16(11):1587. PMID 42279455). A normal scan in a symptomatic patient does not exclude an occlusion. If you would have treated on clinical grounds, treat.
What it has never been shown to do
Here is where the marketing outruns the evidence, and the gap is not subtle.
No published study — controlled or otherwise — demonstrates that pre-injection Doppler screening lowers the rate of vascular occlusion.
What exists is a three-year retrospective single-arm series of 503 patients receiving 1,109 cc of hyaluronic acid with pre-injection vascular mapping, reporting no vascular complications (Teixeira DA. A New Time-Saving Technique for Ultrasound-Guided Filler Injection. Aesthet Surg J Open Forum. 2025;7:ojaf061. PMID 40657559). The study contains a genuinely useful incidental finding — arteries larger than 0.8 mm were present at the intended injection site in 34.7% of glabellas and 13.2% of nasolabial folds — but it is a single operator, single arm, with no control group. Vascular occlusion is a low-base-rate event, and zero events in 503 patients cannot distinguish prevention from chance. The author's own conclusion is properly hedged: Doppler "can help physicians reduce the risk."
An earlier clinical-experience paper reporting 448 patients used a 2 MHz handheld pencil Doppler — an audible flow detector, not an imaging transducer, and not the modality being sold today — and describes only "potential benefits" without reporting a comparative complication rate (Lee GSK. J Cosmet Dermatol. 2019;18(5):1267-1270. PMID 30706697).
The expert position is stronger than the data. A modified Delphi consensus of 15 experts across 11 countries found that panellists "find ultrasound imaging prior to injection mandatory in certain areas with high risk for vascular adverse events" (Velthuis PJ, Alfageme F, Cartier H, et al. J Plast Reconstr Aesthet Surg. 2025;104:33-37. PMID 40101353). That is expert agreement, which is a legitimate form of guidance and is not outcome data. Melissa's test applies exactly here: preference is not greater than proof.
The learning curve nobody has measured
Melissa's specific warning is that this is "not just one certification class but a lot of practice." She is right, and the literature's treatment of the point is remarkable: no study quantifying the learning curve for facial aesthetic ultrasound exists.
Not a long one. Not a disputed one. None. There is no published work reporting the number of scans, the hours of training, or a validated competency benchmark required for a clinician to become reliable at facial vascular sonography.
What the literature contains instead is the phrase used rhetorically. A best-practices paper notes that "the learning curve associated with mastering real-time ultrasound imaging can be steep" (Sigrist R, Desyatnikova S, Chammas MC, Vasconcelos-Berg R. Best Practices … Part 1: Upper Third of the Face. Diagnostics. 2024;14(16):1718. PMID 39202206). Device papers mention a learning curve for novices in their limitations sections. The Delphi consensus provides "an extensive list of training and practice topics" that should be learned before using ultrasound in filler practice — a list of topics, not a dose of training.
The one training-intervention study with a sample size taught facial toxin anatomy to 115 clinicians across four sessions in a year and reported self-rated anatomy understanding rising from 6.5 ± 2.7 to 8.9 ± 1.8 (Zhang YX, Sun YX, Liang S, et al. Aesthetic Plast Surg. 2026;50(1):37-53. PMID 40938410). That is a self-rated Likert score for anatomy education, not a measured sonographic skill, and it cannot tell you how long competence takes.
The practical consequence for anyone comparing courses is straightforward. Any figure quoted for time to competence — a weekend, a number of scans, a certification — is an estimate rather than a measured benchmark, because no measured benchmark has been published. Plan for a longer runway than a brochure describes, and for ongoing scanning volume rather than a single course. That is precisely Melissa's point: "not just one certification class but a lot of practice with it."
Ready to put this into practice?
Explore Empire's hands-on, CME-accredited Aesthetic & Injectable Training courses — live patients, expert faculty, and ongoing mentorship.
The equipment floor is real, and higher than most POCUS units
If you do proceed, the consensus specification is explicit and it disqualifies a good deal of the hardware clinicians already own.
The Delphi consensus states that a minimum linear 15 MHz transducer is recommended for both learning and regular diagnostics; that B-mode, colour Doppler, spectral Doppler and picture/video archiving are mandatory with power Doppler recommended; that handheld devices are acceptable; and that sterile probe covers and sterile gel are required for guided injection (PMID 40101353).
The working groups publishing the best-practice literature report using linear probes ranging from 6 to 24 MHz, and the anatomical studies cited above were performed at 18 MHz.
Two practical points follow. First, a general POCUS unit whose linear probe tops out below 15 MHz, or which lacks spectral Doppler or archiving, falls below the published consensus specification. Second, handheld is not the disqualifier — the consensus explicitly permits handheld devices. Frequency, Doppler modes and archiving are.
One clarification worth making because it causes genuine confusion in the market: diagnostic high-frequency facial ultrasound is not microfocused ultrasound, the energy-based skin-tightening modality. They share a word and nothing else. Consensus statements about microfocused ultrasound say nothing about vascular safety.
It is also worth knowing what governs this space. The relevant documents are a World Federation position paper on aesthetic dermatologic ultrasound (Chammas MC, Sigrist R, Alfageme F, et al. Ultrasound Med Biol. 2025;51(11):2173-2193. PMID 40866164), a European federation position statement on dermatologic ultrasound (Alfageme F, Wortsman X, Catalano O, et al. Ultraschall Med. 2021;42(1):39-47. PMID 32380567), the DERMUS group's examination guidelines, and the Delphi consensus above. The European statement concedes in its own abstract that "the scientific literature regarding this application is still not sufficient for evidence-based guidelines." A 2026 narrative review of the guideline landscape screened 7,981 records and concluded that guidelines on this specific topic remain "quite limited" (Catalano O, Leggieri C, Scarpa AM. J Ultrasound. 2026, online ahead of print. PMID 42509454).
So should you learn it?
This is a career-investment decision, not a purchasing one, and it should be made on four honest questions.
What will you actually use it for? The evidence is meaningfully stronger for diagnosis and targeted management of a vascular event than for routine pre-injection screening. If your motivation is "I will scan before every syringe and therefore be safe," the literature does not support that expectation. If your motivation is "when something goes wrong I want to see where the product is before I flood the region with hyaluronidase," you are aligned with the better evidence — thin as it is.
What is your injection volume? Sonography is a perishable psychomotor skill with a pattern-recognition component. A clinician injecting a few days a month will not maintain competence at reading facial vessels, and a partially competent scan is worse than no scan, because it produces confident misreadings. This is the specific form Melissa's warning takes in practice.
Can you meet the equipment floor? 15 MHz linear minimum, spectral Doppler, archiving. Budget for the device honestly rather than for the course alone.
Where is it in your sequence? This is the part Melissa would not let pass. Ultrasound is an adjunct layered on top of anatomical knowledge, not a substitute for it. A clinician who cannot reason about which carotid system a territory belongs to, where a vessel changes plane, and what a region anastomoses with will not be rescued by a probe — they will be looking at a grey image without a framework to interpret it. "Anatomy first," as she puts it. "I cannot stress that enough."
What changes on Monday, with or without a machine
Even if you never buy one, three things should change because of what ultrasound has revealed.
Treat published arterial courses as distributions rather than positions. Fewer than half of faces follow the modal facial artery course, roughly a third are missing a transverse facial artery, and the two sides of one patient's face differ.
Treat the lower central forehead and glabella as regions where the plane of the vessel is genuinely unpredictable — a 4 to 27 mm range around a 14 mm mean is not a landmark you can act on.
And if you do have ultrasound and a symptomatic patient scans normal, treat anyway. Twelve per cent of confirmed vascular adverse events showed no Doppler abnormality.
The teaching framing in this article reflects Melissa Pulcini-Buttine's clinical practice as taught in Empire Medical Training's hands-on curriculum. Technique is learned under supervision; this article is educational and is not a substitute for training.
Empire does not teach ultrasound, and the foundation this article argues has to come first is anatomical. Clinicians who want to build it on dissection rather than diagrams should look at anatomical based aesthetics training and special anatomical cadaver aesthetics training. Core injectable competence is covered in complete dermal filler training and the highest-risk territories in master eye and nose injection training. Related reading: filler dissolve and can Botox cause blindness.
Melissa Pulcini-Buttine, PA — physician assistant of two decades; professor of anatomy and physiology for ~14 years; faculty member, Empire Medical Training; founder of an aesthetics practice in Greenwich, Connecticut.
Frequently Asked Questions
Does ultrasound prevent vascular occlusion from filler?
No study has shown that it does. As of September 2026 there is no controlled or comparative evidence that pre-injection Doppler screening lowers occlusion rates. The available support is a single-arm 503-patient series with no control group and expert consensus opinion. The evidence is meaningfully stronger for diagnosing and targeting treatment of an event than for preventing one.
How long does it take to learn facial ultrasound?
Nobody has measured it. No published study reports the number of scans, training hours or validated competency benchmark required for facial aesthetic ultrasound. Papers describe the learning curve as "steep" in passing without quantifying it, and the consensus documents list training topics rather than a dose of training. Treat any specific number in a course brochure as unsupported.
What equipment do you need for facial vascular ultrasound?
Consensus specifies a minimum linear 15 MHz transducer, with B-mode, colour Doppler, spectral Doppler and image archiving mandatory, and power Doppler recommended. Handheld devices are acceptable. A general point-of-care unit below 15 MHz, or lacking spectral Doppler or archiving, falls below that specification. Sterile probe covers and gel are required for guided injection.
If the ultrasound looks normal, can I rule out an occlusion?
No. An international multicentre study of 100 confirmed vascular adverse events found no Doppler abnormality in 12% of them. A normal scan in a symptomatic patient does not exclude occlusion. If the clinical picture warrants treatment — disproportionate pain, blanching, livedo, delayed capillary refill — treat on clinical grounds regardless of the image.
Is ultrasound-guided hyaluronidase better than flooding the area?
A retrospective series of 39 patients found targeted treatment used far less hyaluronidase (122.5 IU vs 1,519 IU) with better response. But each patient served as their own sequential control, with flooding attempted first — a design biased toward the second treatment. The entire published literature on ultrasound-guided hyaluronidase amounts to roughly 83 cases.
Disclaimer
This article reflects the clinical opinions and experience of Melissa Pulcini-Buttine, PA, an independent faculty member contributing to Empire Medical Training's curriculum. The views expressed are the author's own and do not necessarily represent those of Empire Medical Training.
It is professional education, not medical advice, and is no substitute for hands-on training or independent clinical judgment. Licensed clinicians remain responsible for their own patient selection, technique and outcomes, for verifying current product labelling, and for practising within their scope and applicable law. Empire Medical Training accepts no liability for reliance on this content.
