A non-healing socket with lip numbness: when 'one more antibiotic course' is the wrong answer

Leading diagnosis: squamous cell carcinoma of the alveolar ridge. Four findings drive it: a non-healing socket beyond 3–4 weeks, induration with rolled everted margins fixed to bone, heavy cumulative tobacco and alcohol exposure, and - the most ominous - unilateral mental nerve anaesthesia indicating perineural invasion. The failure of two adequate antibiotic courses is supportive but not diagnostic; antibiotics frequently produce partial, misleading improvement in necrotic tumours.

Differential, in order of plausibility.

1. Osteomyelitis or a sequestrating socket - possible with mobile adjacent teeth, but does not explain induration or sensory loss, and would usually show more pain, pus and a radiographic mottled pattern.
2. Medication-related osteonecrosis - excluded by a clear negative drug history, though the history must be taken twice; patients forget infusions.
3. Traumatic ulcer or a reactive lesion - excluded by duration, fixation and node.
4. Lymphoma, plasmacytoma or metastasis - cannot be excluded clinically; a non-healing alveolar mass with a node may be haematological or a deposit from lung or kidney.
5. Deep fungal infection (histoplasmosis, mucormycosis) - unlikely without immunodeficiency, but the ulcerated fixed base keeps it on the biopsy differential.
6. Major salivary gland malignancy - unlikely at this site.

The unchanged left level II node is almost certainly incidental and must not be allowed to dilute the significance of the 1.5-cm ipsilateral node.

Judgement point. Nothing here justifies a third antibiotic course; the correct response to a non-healing socket with paraesthesia is urgent biopsy, not another prescription.

Sequence.

1. Same-visit decision and referral: this is a suspected malignancy with paraesthesia, so the target is a tissue diagnosis and head-and-neck multidisciplinary review within 1–2 weeks, not watchful waiting.
2. Incisional biopsy under local anaesthesia, taken from the periphery of the lesion so that it includes the interface between abnormal and clinically normal tissue, deep enough to include epithelium and subepithelial connective tissue; avoid the necrotic centre, which yields only debris and inflammatory slough. Send in formalin with a completed history; request direct immunofluorescence or immunohistochemistry if the provisional reading is uncertain. Fine-needle aspiration of the 1.5-cm node at the same visit.
3. Imaging: contrast-enhanced CT of the neck for cortical erosion and nodal architecture; MRI if perineural spread along the mandibular canal is suspected, since MRI traces the nerve far better than CT. Chest imaging for a second primary or metastasis.
4. PET-CT and full staging only after histological confirmation - it does not make the diagnosis and delays it.
5. Nutrition, smoking cessation, dental assessment and multidisciplinary team discussion before any resection.

What not to do. Do not prescribe a further antibiotic trial; do not perform an excisional biopsy of a 2.5-cm lesion you cannot clear; do not extract the mobile 35 and 37 to 'clean up' the area, which risks seeding and destroys tissue needed for reconstruction; do not send a single superficial pinch; do not reassure him on the basis of one negative sampling. Document the paraesthesia objectively (light touch and two-point discrimination on both sides) because it is both prognostic and medico-legally important.

Staging consequences. In oral cavity carcinoma, invasion through cortical bone into the medullary space or producing a pathological fracture places the tumour at T4a (moderately advanced local disease); a 2.5-cm lesion confined to soft tissue would otherwise be T2. The node is single, ipsilateral, 1.5 cm and mobile, which is N1 provided cytology confirms metastasis. Clinical reasoning must not pre-empt histology, but the working stage governs the conversation.

Why the numbness matters. Mental nerve involvement means perineural spread, which is an independent adverse prognostic factor: higher rates of local recurrence, a higher risk of skip lesions along the canal well beyond the clinically visible margin, and a reason to plan imaging of the nerve course and to consider frozen-section control of the mandibular canal margin at resection. Numbness in a patient with a non-healing socket is a red flag that should move the case from 'routine oral surgery' to 'two-week suspected cancer pathway' on its own.

Node matters differently. A palpable node converts this from a local problem to a regional one, mandates neck dissection discussion at the multidisciplinary meeting and makes FNA an early, not a later, test.

Urgency. The practical consequences are: biopsy within days rather than weeks, MRI of the mandible to map nerve spread, and immediate referral to a unit able to perform segmental mandibulectomy with reconstruction. Delay of even 4–6 weeks while repeating antibiotics can upstage the neck and worsen survival; that delay is the commonest avoidable error in this presentation.