Nasal Dermoid

“This is a clinical photograph of a young child, taken from below, showing the nasal dorsum and tip. There are two pigmented puncta in the midline of the nasal dorsum. I would want to know whether hair is protruding from either of them, whether they discharge, and whether there is a palpable subcutaneous swelling deep to them extending up towards the glabella.

The history I need is short: has this been there since birth, has it grown, has it ever been red, discharged or needed antibiotics, and has the child ever had meningitis.

My working diagnosis is a congenital midline nasal dermoid sinus cyst. The two things I am formally excluding are an encephalocele and a nasal glial heterotopia — a so-called nasal glioma.

The important point is that the size of what I can see tells me nothing about the depth of what I cannot. I would not probe it or biopsy it before imaging.”

“Three congenital lesions account for almost all midline nasal masses in a child, and they are embryologically related — dermoid, encephalocele and nasal glial heterotopia. Dermoid is at least 60 per cent of them.

I separate them clinically before I image. A dermoid is firm, non-compressible, does not transilluminate and does not change with crying. An encephalocele is soft, compressible, may be pulsatile, transilluminates and enlarges with crying or jugular compression — a positive Furstenberg sign — because it retains a connection with the subarachnoid space. A glioma is firm and non-compressible like a dermoid, but is often reddish or telangiectatic and is commonly mistaken for a capillary haemangioma.

Beyond those three I would consider a teratoma, a nasolacrimal duct cyst at the medial canthus, an infantile haemangioma, a venous or lymphatic malformation, a sinus pericranii, a pilomatricoma, and — the one I would not want to miss — a rhabdomyosarcoma.

But I would not spend long on that list. The clinically decisive distinction is dermoid versus encephalocele, because getting that wrong and probing or biopsying the lesion risks a CSF leak and meningitis.”

“Because a strip of ectoderm gets left behind when the dura withdraws from the nasal skin.

In the developing fetus there are two potential spaces at the nasal root: the fonticulus frontalis, the membranous gap between the frontal and nasal bones, and the prenasal space, between the nasal bones above and the cartilaginous nasal capsule below. A diverticulum of dura projects forwards through the developing foramen cecum into the prenasal space and contacts the skin of the nasal dorsum. Normally that dural projection then involutes, the prenasal space obliterates, and the foramen cecum and fonticulus close.

If the dura stays adherent to skin as it retreats, it drags ectoderm — and sometimes mesoderm — back with it. That trapped ectoderm becomes the cyst; the tract it is dragged along becomes the sinus, which is why the tract can run all the way from a skin punctum on the dorsum to the foramen cecum and, in a minority, through it.

The same anatomy explains the other two lesions: a persistent patent herniation through the fonticulus or foramen cecum is an encephalocele; one that then loses its intracranial connection is a glial heterotopia. That is why the three sit together in the differential.”

“I would examine the lesion, the nose, the eyes and the whole child, with the parents present and with the child on a parent’s lap if that is easier.

The lesion: site relative to the midline, size, consistency, whether it is compressible, whether it is pulsatile, whether it transilluminates, whether there is a punctum, whether hair or sebaceous material is expressed, and whether there is any surrounding erythema suggesting current infection. I would palpate along the dorsum up to the glabella for a subcutaneous cord, which is the tract.

Then the Furstenberg test. I would compress both internal jugular veins, or simply watch the lesion while the child is crying, and look for the mass to enlarge or become more tense. A dermoid does not change — a negative Furstenberg. An encephalocele does, because it communicates with the subarachnoid space.

The rest of the child: intercanthal distance and nasal width, the nasal airway, and a general examination for other congenital anomalies.

And then the sentence that matters: none of this excludes intracranial extension. A dermoid tract does not transmit CSF pressure, so a negative Furstenberg tells me it is not an encephalocele — it tells me nothing about whether the tract reaches the skull base.”

“MRI is my primary investigation, and its purpose is a single binary question: does the tract reach the anterior cranial fossa.

I would ask for thin-slice, high-resolution MRI with contrast, including coronal images through the crista galli and the foramen cecum, and I would want the brain imaged as well, because associated intracranial anomalies are described. A dermoid is typically hyperintense on T1 because of its lipid content, variable on T2, and does not enhance — the rim may enhance if it has been infected.

I would add fine-cut CT where I need the bone anatomy for surgical planning, and in a child under five many units, including the Polish group who published on this in 2021, argue for both. Amin’s 2024 series of 129 — the largest — found CT and MRI comparable and recommended single-modality imaging. I would take the decision with a paediatric neuroradiologist rather than by protocol.

The signs of intracranial extension are a bifid or bulbous crista galli and a widened foramen cecum — but I would call those indirect evidence, not proof.

And I would be explicit with the family and with myself that imaging cannot exclude intracranial extension. In Winterton’s Leeds series the negative predictive value of both CT and MRI was 50 per cent, which is why they used an approach that allowed them to convert intra-operatively.”

“I would recommend excision, and I would give them the number rather than an impression.

Kubba’s Glasgow group published a survival analysis of a 22-year series in Clinical
Otolaryngology in 2023. The risk of infection is around 7 per cent per year through childhood — roughly half of children have had at least one local infection by the age of four, and over 90 per cent by their ninth birthday. So watchful waiting is not a neutral choice; it is a choice with a quantified downside.

The reason infection matters is not the cellulitis. It is that the tract can be a route to the anterior cranial fossa — meningitis, frontal osteomyelitis and brain abscess are all reported, and there is a published case of a child of eighteen months left with irreversible brain injury after an infected nasal dermoid.

Infection also makes the definitive operation harder — scarring and adhesions mean a wider exposure and, on the published data, a worse cosmetic result after previous incomplete surgery.

So my position is elective excision, planned, in a quiet field, in a specialist unit — not urgent, but not deferred indefinitely either. I would be honest that there is no randomised evidence on the optimal age, and no series stratifies outcome by age at operation.”

“That is a reasonable question and there is good evidence to answer it with.

The FDA issued a warning in 2016, but it is specific: it concerns repeated anaesthetics, or single anaesthetics lasting more than three hours, in children under three. A single, relatively short anaesthetic was described even in that communication as unlikely to affect behaviour or learning.

Since then, the GAS trial — the only randomised trial in this area — followed 722 infants to the age of five and found full-scale IQ equivalent after around an hour of sevoflurane compared with awake regional anaesthesia. PANDA, a sibling-matched study, found no difference after a single exposure before 36 months. MASK found no IQ difference either, though it did find small differences in processing speed and fine motor skills in children with multiple exposures.

And I would turn that round: the exposure pattern the FDA actually warns about — repeated anaesthetics — is what she risks if we leave this and she needs incision and drainage two or three times before a harder definitive operation. One planned operation, in a specialist paediatric centre, with a paediatric anaesthetist, is the lower-exposure option.”

“First, who I am consenting. I would confirm who holds parental responsibility — under the Children Act 1989 one person with parental responsibility can consent, but I would want both parents involved where there are two. I would involve the child herself to the extent she can understand, and if she were older I would consider whether she was Gillick competent to consent in her own right.

Second, the standard. Since Montgomery, my duty is to ensure the parents are aware of any risk a reasonable person in their position would attach significance to, and of the reasonable alternatives — which here includes doing nothing, quantified.

Third, the specific risks, and I would group them: recurrence, around 12 per cent in Rahbar’s series at a mean of seven years’ follow-up, higher after incomplete excision; the need to extend or convert the approach intra-operatively, up to and including a craniotomy; CSF leak and the need for dural repair, possibly a lumbar drain; meningitis; anosmia, which is anatomically plausible near the cribriform plate although I am not aware of a published rate; scar and nasal contour irregularity, with the possible need for cartilage grafting; blood transfusion if we go transcranial in a small child; and the possibility of further surgery. Fourth, the process. Two-stage consent — clinic and again on the day — written information, a professional interpreter if English is not the parents’ first language, and separate written consent for clinical photography beyond the medical record.”

“I would choose the smallest approach that gives me complete clearance of the cyst and the entire tract, because recurrence is a function of residual tract, not of incision length.

I would classify the lesion first, using Hartley’s four types from the GOSH series of 103 — superficial, intraosseous, intracranial extradural, intracranial intradural.

• Superficial: direct midline or elliptical excision, taking the punctum and the abnormal skin with it.
• Intraosseous: an external, open-tip rhinoplasty approach, which gives wide access to the dorsum, the septum and the skull base, and lets me drill the frontonasal bones to follow the tract. It is particularly suited to the young child with a short nose.
• Intracranial extradural: this does not automatically mean a bicoronal craniotomy. A brow incision with a small-window anterior craniotomy is the GOSH alternative and is what Hartley’s paper explicitly advocates.
• Intracranial intradural, extensive disease, or equivocal imaging: a combined approach — bicoronal flap and frontal craniotomy with the neurosurgeons, with dural repair and anterior skull base reconstruction.

Whichever I choose, the principles are Pollock’s: access to the whole cyst and tract, access to the skull base, exposure adequate to reconstruct the dorsum, and an acceptable scar. I would dissect the tract in continuity, under magnification, and send it whole.”

“I would treat the infection and I would not excise it today. Assess the child first — is she systemically well, is there any orbital or intracranial concern, any neck stiffness, any altered consciousness. Bloods and cultures, discuss with paediatrics and microbiology, and start antibiotics with staphylococcal and streptococcal cover according to local policy. If there is a collection, incision and drainage — in Bishop’s series three of seven infected children needed drainage before definitive surgery.

Then I would let it settle completely and bring her back for planned definitive excision with full imaging and neurosurgical availability. Operating on an acutely infected field means an incomplete excision in scarred tissue, which is the single strongest predictor of recurrence.

The one thing that would change that plan is intracranial sepsis — meningitis, or imaging showing an intracranial collection. That becomes an emergency, managed jointly with neurosurgery.

Reassuringly, Bishop’s data suggest prior infection does not itself increase the recurrence rate — but it is a series of fourteen, so I would hold that lightly.”

“I stop, and I do not chase it blindly.

This is exactly the scenario I consented for, so the first thing is that the parents already know it is possible. I would confirm the anatomy, and if I have neurosurgical support available — which is why I would be doing this in a craniofacial unit — I would extend the approach appropriately to achieve complete clearance under direct vision, with dural repair if the dura is breached.

If I did not have that support available, I would take what I can safely and completely remove, mark and document the level at which I stopped, close, re-image, and bring the child back for a planned combined procedure. A staged, complete excision is better than a single-stage incomplete one, and far better than a dural injury I cannot repair.

This is precisely why Leeds argued for an approach that permits intra-operative conversion — because imaging has a negative predictive value of only 50 per cent.”