Fontanelle ultrasound

1. Bedside vignettes: why this matters#

Vignette A. Preterm 26-weeker with sudden bradycardia#

A 26-week, 850-g infant, day 3 of life. Acute desaturations and bradycardia overnight. Anterior fontanelle slightly tense. TFUS at the bedside: bilateral grade III IVH with dilated ventricles (left VI 18 mm, right 16 mm), small parenchymal echodensity in the right frontal periventricular white matter (early PVHI / Papile IV). Neurosurgery is consulted; serial TFUS over the next 4 days tracks ventricular dilation; an Ommaya reservoir is placed when VI reaches 22 mm. Without TFUS, this would have been a CT-and-transport day; with it, the management decision happens in the cot. 1 2

Vignette B. 6-month-old with bulging fontanelle, vomiting#

A previously well 6-month-old, 4 days of progressive vomiting and irritability. Anterior fontanelle bulging and tense. TFUS shows markedly enlarged lateral and third ventricles with effaced cisterns; an apparent posterior-fossa mass distorting the fourth ventricle. The team escalates straight to MRI without dwelling on CT; a posterior fossa tumour is confirmed and the child goes to theatre for an EVD that evening. TFUS converted an unclear clinical picture into a focused imaging plan. 1

Vignette C. Term newborn, day 2, post-asphyxial HIE under cooling#

A 39-week term newborn, severe HIE, day 2 of therapeutic hypothermia. The cot-side cranial ultrasound is normal: ventricles slit-like, basal ganglia echogenicity normal, no IVH. Through-fontanelle Doppler ACA: PI 0.42, EDV 32 cm/s, MFV 75 cm/s. The "lush" diastolic flow paradox: low PI with high EDV in early HIE means luxury perfusion with collapsed metabolism, a poor prognostic signature. The team waits for the MRI at day 4–7 but already has the bedside Doppler data to inform the family conversation. 3 4

2. What TFUS is, and what it is not#

The infant skull has two open fontanelles at birth: the anterior (diamond-shaped, frontal-parietal junction, ~3 × 3 cm at term, closes 9–18 months) and the posterior (triangular, parieto-occipital, ~1 × 1 cm, closes 6–12 weeks). The anterior is the main acoustic window for routine bedside imaging.

The principle is straightforward: the fontanelle is a window of soft tissue rather than bone. High-frequency ultrasound (5–10 MHz curvilinear or sector probe, 7–15 MHz linear for superficial structures) penetrates through this window and returns a clear image of the brain immediately beneath. Lateral views through the temporal bone provide complementary access to the basal cisterns and middle cerebral arteries.

What TFUS does well#

• No radiation, no contrast, no transport: ideal for the unstable neonate.
• Serial repeatability: same machine, same windows, repeatable trend across days.
• IVH grading: the Papile classification (I–IV) was developed on TFUS and remains the standard for preterm IVH grading. 1
• Ventricular size tracking: Levene VI and anterior horn width are the bedside metrics for post-haemorrhagic hydrocephalus management.
• Detection of PVL and PVHI: echodensities, cystic evolution, and timing visible at the bedside.
• Through-fontanelle Doppler: ACA and basilar artery PI / RI as flow surrogates; the only non-invasive cerebral flow tool in the preterm.

What TFUS cannot do#

• Image the convexity well: the curved skull and the soft-tissue gap mean lateral cortex (and the most common cortical infarct territory) is poorly seen.
• Resolve subtle injury: small punctate lesions, mild PVL, mild diffuse injury require MRI.
• Detect early HIE injury reliably in the first 24 h: changes appear later; MRI day 4–7 is the gold standard.
• Replace MRI for prognostic anatomical detail.
• Be performed after fontanelle closure: the window narrows from ~12 months and is gone by ~18 months in most infants.

3. Anatomy and windows#

3.1 The anterior fontanelle#

The anterior fontanelle, ~3 × 3 cm at term, narrows progressively and closes between 9 and 18 months. It is the primary window for routine TFUS.

3.2 The temporal window#

Above the zygomatic arch, in front of the tragus, the thin temporal bone permits ultrasound access. This is also the principal TCD window in adults. In infants it complements the anterior fontanelle by providing views of the basal cisterns, brainstem, and Circle of Willis.

3.3 The posterior fontanelle#

Small (~1 cm), triangular, closes 6–12 weeks. Useful for occipital horn views in the first month of life.

3.4 The mastoid window#

Below and behind the ear; provides views of the posterior fossa, including the fourth ventricle and cerebellar hemispheres. Useful when the anterior fontanelle view is suboptimal.

3.5 Standard sweeps#

Two sweeps, six planes coronal, three planes sagittal, complete a screening study:

4. The signal: what TFUS shows#

4.1 Normal anatomy#

In a healthy term newborn, the lateral ventricles are slit-like, the third ventricle barely visible, the choroid plexus echogenic, the basal ganglia and thalami isoechoic with cortex, and the cerebellum normal in shape.

4.2 IVH (Papile grading)#

The four-grade Papile classification is the standard for preterm IVH:

The acoustic shadow of fresh haemorrhage is bright (echogenic); as it matures it becomes hypoechoic and can develop cystic changes.

4.3 PVL#

Periventricular leukomalacia progresses through stages: early echogenic foci in the periventricular white matter, evolving to cystic cavities over 2–6 weeks. Subtle PVL requires MRI confirmation; established cystic PVL is visible on TFUS.

4.4 Hydrocephalus / ventriculomegaly#

The Levene Ventricular Index (VI) is the bedside measurement: width of the lateral ventricle at the level of the third ventricle, on a coronal view through the foramen of Monro. The 97th centile by GA defines hydrocephalus.

4.5 Mass effect / midline shift#

Subarachnoid blood, large subdural or extradural haematomas, and tumours produce mass effect with midline shift; TFUS detects gross shift but underestimates subtle effects.

4.6 Through-fontanelle Doppler#

A pulsed-wave Doppler gate placed over the ACA (in the interhemispheric fissure on a sagittal view) or basilar artery (in the midline sagittal posterior fossa view) generates a velocity envelope. PI (Gosling) and RI (Pourcelot) are calculated:

In a term newborn ACA, raised ICP shows as falling diastolic velocity with the Pourcelot resistive index (RI) rising toward its 1.0 ceiling (absent or reversed diastolic flow in the extreme), rather than a fixed pulsatility-index cutoff; PI > 1.4 is an adult-MCA convention, not a neonatal ACA threshold (normal term cerebral PI is only ~0.7–1.1). RI > 0.8 is a comparable sign. Because transducer pressure on the fontanelle, a patent ductus arteriosus, low PaCO2, and hypovolemia all shift these indices, no single numeric RI cutoff is reliable: use the trend and the diastolic-flow pattern, not one number. PI < 0.6 with high EDV in HIE signals luxury perfusion.

5. The numbers to record: the TFUS six-pack#

Document study quality, fontanelle size, infant state (settled vs irritable), and the operator. Compare against the previous study rather than a single time point.

6. What is normal? Age- and GA-banded reference#

6.1 Ventricular size by GA (Levene centiles)#

Sources: Levene normative data; 1 for IVH grading. A VI > 97th centile or rising VI over serial studies is the trigger for intervention.

6.2 Doppler indices by age#

Sources: 5 6 7. ACA PI normally falls in the first 24 h of life as systemic resistance settles.

7. What is abnormal? Pattern library#

Decision tree: tense fontanelle in an infant#

flowchart TD
  Tense[Tense fontanelle / irritability] --> TFUS[TFUS at bedside]
  TFUS --> Vent{Ventricles}
  Vent -->|Dilated| Hydro[Hydrocephalus; neurosurgical consult]
  Vent -->|Slit-like| Mass{Mass effect / midline shift?}
  Mass -->|Yes| MRI[Urgent MRI / CT; consult]
  Mass -->|No| Doppler{ACA diastolic flow}
  Doppler -->|Low EDV, high RI| ICP[Likely raised ICP; treat; serial imaging]
  Doppler -->|Preserved EDV| Other[Other cause: meningitis, oedema, metabolic; broader workup]

8. Try it: interactive widget#

Through-fontanelle ultrasound is operator-dependent and not easily reduced to a single interactive demo. Future expansion: an annotated image library and a virtual probe-positioning trainer. For now, refer to the TCD waveform explorer for the Doppler-envelope mechanics, which transfer directly to through-fontanelle Doppler.

9. Management: how TFUS drives bedside decisions#

9.1 Post-haemorrhagic ventricular dilatation#

The most common management driver. The progression is: IVH → ventricular dilation → arrested or progressive hydrocephalus → intervention.

1. Serial TFUS every 1–3 days during the first 2 weeks after IVH; weekly thereafter while stable.
2. VI threshold: rising > 97th centile or > 4 mm/week increase prompts neurosurgical consult.
3. Bedside CSF removal via lumbar puncture if accessible; ventricular reservoir (Ommaya) placed if persistent and VPS too premature.
4. Definitive VPS delayed until infant size and CSF protein permit.
5. Re-image at every clinical deterioration: bulging fontanelle, increased irritability, apnoeas, vomiting. 1

9.2 HIE serial monitoring#

TFUS day 1–3 in HIE is often normal or non-specific; the value is serial:

1. Day 1: baseline; usually unremarkable.
2. Day 2–3: may show early oedema (effaced sulci, narrow ventricles, increased parenchymal echogenicity).
3. Day 4–7: MRI is the gold standard; TFUS may show evolving changes.
4. Through-fontanelle Doppler PI / RI trends over the first 72 h carry prognostic information; a luxury-perfusion pattern (low PI, high EDV) is unfavourable. 3 4

9.3 Acute deterioration in the infant#

A previously stable infant who becomes irritable, bradycardiac, or apnoeic in the NICU / PICU gets a bedside TFUS as part of the immediate workup, in parallel with blood gas, glucose, and clinical exam. The most common findings are evolving IVH, hydrocephalus, or mass effect; the rapid bedside answer changes the next decision.

10. Clinical contexts#

10.1 Preterm IVH and PVH#

The single largest TFUS use case. Routine screening at days 3, 7, and 14 (or per local protocol) in infants < 32 weeks GA or < 1500 g identifies IVH in 20–25% (down from 40–50% in the pre-surfactant era). Serial TFUS guides Ommaya / EVD / shunt decisions. 1 2

10.2 Neonatal HIE and post-arrest#

TFUS is a screening tool in HIE; MRI day 4–7 is the gold standard for prognostication. Through-fontanelle Doppler adds early information about luxury perfusion. The combined aEEG + TFUS + MRI + clinical exam bundle is the standard neonatal HIE neuroprognostic stack. 4 2 3

10.3 Hydrocephalus monitoring#

Beyond preterm IVH, TFUS is the workhorse for serial monitoring of any infant hydrocephalus: post-meningitic, post-haemorrhagic, congenital (aqueduct stenosis), or post-tumour resection. Used to time shunt placement and to evaluate post-shunt ventricular size.

10.4 Bacterial meningitis#

Complications of meningitis (subdural empyema, subdural effusion, ventriculitis, infarct, oedema) are visible on TFUS in the infant. TFUS is the first imaging response to a clinically deteriorating infant with meningitis, supplemented by MRI when feasible. 8 9 10

10.5 Pediatric ECMO (neonatal)#

Daily TFUS is part of the standard neuro-monitoring for neonates on ECMO; stroke incidence is high (8–15%) and TFUS is the bedside imaging modality of choice. New IVH or large infarct on TFUS may be a contraindication to continued anticoagulation. 11 12

10.6 Pediatric arterial ischaemic stroke#

In the infant with suspected AIS, TFUS is a triage tool; MRI / MRA is definitive. The fontanelle window is inadequate for most cortical infarct territories. 13

10.7 Severe TBI in the infant#

A rare scenario in the infant (non-accidental trauma is the dominant aetiology in this age group). TFUS detects acute subdural haemorrhage, mass effect, and ventricular involvement; CT / MRI is required for the full assessment. 14

10.8 Brain-death determination (limited)#

TFUS may show absent intracranial flow as supportive evidence; the formal World Brain Death Project framework requires the clinical exam and apnoea test with ancillary imaging (CTA, MRA, EEG, BAER), not TFUS alone. 15 16

10.9 Aneurysmal SAH and DCI (rare in infants)#

Aneurysmal SAH is rare in this age group; AVM rupture and trauma are more common. TFUS may detect IVH and ventriculomegaly; CTA / MRA is required for the underlying lesion. 17

11. Multimodal integration: TFUS in the MMM/MNM stack#

12. Setup and technique#

12.1 Equipment#

• Ultrasound machine: portable bedside cart with neonatal preset.
• Probes: 5–10 MHz curvilinear or sector probe for routine TFUS; 7–15 MHz linear probe for surface structures and superficial views; phased-array probe for through-fontanelle Doppler.
• Coupling gel: warmed before application to a small infant.
• Image storage: PACS upload for serial comparison.

12.2 Standard scan: 8-step protocol#

1. Position the infant: supine, head midline; the bedside nurse may help maintain position.
2. Warm the gel to body temperature; apply over the anterior fontanelle.
3. Begin the coronal sweep: probe held perpendicular to the skull, swept from anterior (frontal lobes) to posterior (occipital horns), capturing the six standard coronal planes.
4. Switch to sagittal: rotate the probe 90°; capture midline plus right and left parasagittal planes.
5. Document VI on the standard coronal plane at the foramen of Monro; measure on the captured image.
6. Doppler interrogation: pulsed-wave gate over the ACA in the interhemispheric fissure (sagittal midline) and basilar artery (sagittal posterior fossa); record PI and RI over ≥ 5 cardiac cycles.
7. Temporal window: optional; useful for MCA / circle of Willis if fontanelle window is suboptimal.
8. Documentation: capture the six coronal + three sagittal + Doppler images; PACS upload; structured report.

12.3 Reading routine#

• Compare with the previous study wherever possible.
• Measure VI on the same coronal plane for serial comparability.
• Note infant state: an irritable infant produces higher ACA PI than a settled one.
• Document study limitations: small fontanelle, motion artefact, suboptimal window.

12.4 Training and credentialing#

TFUS interpretation is operator-dependent. Most NICUs maintain a small group of credentialed bedside operators (neonatologists, fellows, ultrasonographers); routine inter-rater reliability checks against pediatric radiology are standard practice. The skill is portable but acquired over months of supervised scanning.

12.5 When to escalate to MRI / CT#

• Suspected mass lesion with limited TFUS resolution: MRI.
• Suspected cortical infarct outside the fontanelle window: MRI / MRA.
• HIE prognostic timepoint (day 4–7): MRI with DWI is gold standard.
• Suspected non-accidental trauma: CT for acute haemorrhage, then MRI for parenchymal injury.
• Persistent unexplained findings on TFUS: MRI for definitive characterisation.

12.6 Limitations of the modality#

• Convexity poorly seen: cortical infarcts, small subdural haematomas may be missed.
• Posterior fossa partial: mastoid window improves but does not fully resolve.
• Subtle PVL requires MRI.
• Operator skill varies: inter-rater agreement on subtle findings is moderate.

13. Pitfalls#

• Settling the infant matters: motion artefact destroys subtle findings; gentle handling and warm gel pay off.
• Cold gel triggers crying, raises ACA PI, and worsens scan quality.
• Measuring VI on a non-standard plane introduces variability; commit to one plane (foramen of Monro) for serial comparability.
• IVH evolves: fresh haemorrhage is bright; mature haemorrhage is hypoechoic. A "disappearing" IVH may mean evolution, not resolution.
• Cystic PVL takes weeks to develop: early PVL is echogenic foci that may persist or evolve.
• Through-fontanelle Doppler indices depend on infant state: an irritable infant has higher PI than a settled one.
• Doppler angle: with through-fontanelle ACA Doppler the angle is favourable but not necessarily zero; recognise that absolute velocities are subject to the same angle assumptions as adult TCD.
• Fontanelle closure: a partly-closed fontanelle gives partial views; recognise the limitation and consider MRI.
• Operator dependence: a study read as "normal" by an inexperienced operator may have missed subtle findings; serial studies by the same operator are more useful than single comparisons.
• Convexity blindness: a child with sudden hemiparesis may have an MCA-territory infarct that TFUS does not see; do not dismiss the clinical picture on a "normal" TFUS.

14. Combine with…#

• Amplitude-integrated EEG: electrophysiology + structure for HIE.
• NIRS: cerebral oxygenation; SafeBoosC and HIE paradigms.
• TCD: formal Doppler interrogation; through-fontanelle Doppler is its infant analogue.
• ICP: direct measurement when fontanelle-based estimates are insufficient.
• Non-invasive ICP: ONSD, B4C, TCD-PI as complementary tools.
• Foundations: pediatric physiology: why the infant brain is different.
• Integration: HIE monitoring bundle: the neonatal multimodal stack.
• Integration: resource-limited bedside monitoring: TFUS as the structural channel.

15. Evidence summary#

16. Recent literature (2022–2025)#

• Sansevere 2019 (neonatal cEEG review): positions TFUS as the bedside structural channel paired with aEEG and cEEG in the neonatal monitoring bundle. 19
• Slovis 2024 (pediatric brain injury post-arrest): TFUS as serial bedside imaging in the infant post-arrest, complemented by MRI day 4–7. 21
• Laws 2022 (pediatric MMM review): TFUS as a bedside imaging modality in the neonatal and infant neuro-ICU. 18
• Laws 2022 (pediatric MMM review): describes TFUS as a bedside imaging modality for the neonate and infant. 18
• ECMO neuro complications (Pisano 2024): neurologic injury is common on ECMO; bedside cranial ultrasound in neonates can inform anticoagulation decisions. 12
• Through-fontanelle Doppler in HIE continues to be refined as a prognostic tool, paired with aEEG, MRI, and clinical exam in the term newborn neuroprognostic bundle. 3

17. Self-check#