Glomus Tumour of the Hand

Glomus tumours of the hand are uncommon, 1 to 5% of all hand tumours, and cause symptoms out of all proportion to their size. The classic triad of paroxysmal pain, pinpoint tenderness and cold sensitivity is so specific that, once recognised, the diagnosis is essentially clinical; when all three features are present the triad is pathognomonic.

Who. The peak is in the third to fifth decades, though the tumour is reported at any age, and women outnumber men by approximately 2-3 to 1. About 10% of patients have multiple lesions, often familial, the glomuvenous malformation described under Clinical Assessment.

Where. 75 to 80% are subungual. The remainder lie in the pulp, the volar finger and the palm.

The delay. The average interval from first symptom to correct diagnosis is around 7 years, because clinicians do not examine the nail bed carefully and do not perform the Love and Hildreth tests. The tumour is tiny (1-5 mm) and often invisible on inspection, the plain radiograph is frequently normal in early disease, and the symptoms are episodic, so the patient is well between attacks. Misdiagnosis is common: neuroma, Raynaud disease, neuropathy, arthritis, complex regional pain syndrome (CRPS) and psychiatric pain.

What it costs the patient. The pain is disabling, paroxysmal and triggered by cold and pressure, so patients avoid typing, writing, tool use and cold weather. Years of undiagnosed pain bring anxiety and sleep disturbance, multiple GP and specialist visits, and unnecessary imaging and treatment.

The glomus body. A specialised arteriovenous anastomosis, the Sucquet-Hoyer canal, in the reticular dermis. It is composed of an afferent arteriole, the Sucquet-Hoyer canal itself, an efferent venule and an intraglomerular reticulum of modified smooth muscle cells. Its function is thermoregulation: it shunts blood flow in the skin in response to temperature, particularly in the fingertips, toes and subungual regions, and glomus bodies are most concentrated in the fingertips and subungual region.

The tumour. A glomus tumour is a benign hamartomatous proliferation of these modified smooth muscle cells: a small, well-circumscribed, purple-red nodule that is rich in nerve fibres, hence the exquisite pain. Three components make up the lesion, and each accounts for part of the clinical picture:

• Modified smooth muscle cells - spindle or epithelioid glomus cells around vascular spaces form the bulk of the tumour; positive for SMA and vimentin
• Vascular channels - small, endothelium-lined channels within the lesion; they constrict in the cold, which is what triggers the paroxysmal pain and reflects the thermoregulatory origin of the parent cell
• Non-myelinated nerve fibres - abundant unmyelinated C fibres within and around the tumour; this dense innervation is the source of the pain and accounts for pinpoint tenderness, hyperalgesia, and Love and Hildreth positivity

View larger

Why pinpoint, and why the tourniquet helps. The nodule is small, discrete, encapsulated and superficial, so the tenderness is confined to a single spot. Exsanguination empties the vascular channels and removes the pressure on the nerve fibres, which is why the Hildreth test abolishes the pain.

Histology. Well-circumscribed nodules of uniform, round, epithelioid glomus cells arranged in nests around small vascular channels. A fibrous pseudocapsule is typical, which supports marginal rather than wide excision. The cells are positive for smooth muscle actin (SMA), vimentin and h-caldesmon and negative for S100, CD31 and cytokeratins, useful in distinguishing them from melanoma, neurofibroma and adnexal tumours.

Two classifications are in use and they answer different questions. Anatomic location decides the operation; Masson's histological subtype supports the diagnosis on pathology but does not change management, since every variant is treated by marginal excision.

Could it be malignant? The Folpe classification. The overwhelming majority are benign, but a malignant variant, the glomangiosarcoma (malignant glomus tumour), is rare but real, so "benign" is the rule rather than a guarantee. Malignancy is defined histologically; there is no such thing as "malignant transformation" of a benign glomus tumour. Folpe, from 52 unusual glomus tumours, proposed the four-category scheme the WHO adopted:

• Benign glomus tumour - the ordinary lesion
• Symplastic glomus tumour - high nuclear grade alone, with no other worrying feature; despite the alarming nuclei it behaves benignly
• Uncertain malignant potential (UMP) - one concerning feature only (high mitotic activity in a superficial lesion, or large size (over 2 cm) alone, or deep (subfascial or visceral) location alone) without meeting full malignant criteria
• Malignant glomus tumour (glomangiosarcoma) - deep location and size over 2 cm, or atypical mitotic figures, or moderate-to-high nuclear grade with 5 or more mitoses per 50 high-power fields

Behaviour tracks the category. In Folpe's series metastasis occurred in 38% of tumours meeting malignant criteria and in none classified symplastic, UMP or glomangiomatosis; the drivers of metastatic risk were deep location, size over 2 cm and atypical mitoses. Malignant glomus tumours are far more common in deep or extradigital sites than in the typical superficial subungual location, and they can metastasise (lungs). That is a second reason, beyond excluding melanoma, that every excised specimen goes to histology. A genuinely malignant or uncertain-potential lesion warrants wide excision, sarcoma-service referral and oncological follow-up, not simple marginal excision.

History. The pain is episodic and paroxysmal, sharp or burning, and triggered by cold exposure, light pressure and dependency. The patient points to a single fingertip or subungual spot. The history is typically years long and already carries several of the wrong labels listed in the overview.

Examination. Inspect the nail for a subungual blue-red spot, nail ridging or onycholysis, and palpate for a small, well-defined, exquisitely tender nodule. Then perform the two named tests.

Love test. With the patient seated and the hand rested, apply the tip of a paper clip, ballpoint pen, pencil or cotton-tipped applicator to the suspected spot. A positive test is disproportionate, sharp pain with minimal pressure; the patient typically withdraws the hand. It is easy to perform at the bedside and highly sensitive.

Hildreth (ischaemic) test. Apply a digital tourniquet (a Penrose drain or the finger of a glove) at the base of the affected digit, exsanguinate the digit by wrapping or milking blood proximally, then inflate or tighten the tourniquet. A positive test is complete abolition of the pain, and often of the ability to localise it, while the digit is exsanguinated: the patient who could previously point to a single spot can no longer find it. Release of the tourniquet causes immediate return of the pain.

Why Hildreth matters. It is the most specific bedside test for glomus tumour and helps differentiate it from other causes of fingertip pain, including neuroma and small nerve entrapment. This near-pathognomonic sign, combined with the classic triad, should be sufficient to justify surgery without further imaging, though MRI is still obtained for surgical planning.

Adjuncts. An ice cube or cold water on the digit reproduces the pain; the cold test is helpful but less specific, and positive in Raynaud too. Transillumination with a penlight through the distal phalanx may show a small reddish mass under the nail, a useful adjunct for subungual lesions.

Solitary glomus tumour versus glomuvenous malformation. The familial, multiple "glomus tumours" are a genuinely different entity, the glomuvenous malformation (GVM), and the distinction matters for both the exam and management. A solitary glomus tumour is single, tiny, exquisitely painful with the classic triad, usually subungual, sporadic, and cured by marginal excision. A GVM is multiple, often plaque-like or nodular, bluish and partially compressible, and frequently non-tender or only mildly tender rather than showing the paroxysmal triad. It is inherited autosomal dominant via the GLMN (glomulin) gene with a paradominant, "second-hit" mechanism, hence the patchy distribution and variable penetrance.

Why the distinction changes treatment. GVMs are venous-predominant histologically (the glomangioma pattern) and respond poorly to simple excision; symptomatic or cosmetically troublesome lesions may be treated with sclerotherapy or laser as well as selective excision. The take-home: a painful solitary subungual lesion is a classic glomus tumour, whereas multiple compressible, non-tender bluish lesions with a family history are a glomuvenous malformation, so examine the skin and ask about affected relatives.

The exception: neurofibromatosis type 1. NF1 patients develop multiple true glomus tumours in the digits, often in several fingers and frequently bilateral, and unlike glomuvenous malformations these are the painful kind, with the full triad in each lesion. So "multiple, therefore not painful, therefore a malformation" is a shortcut that fails precisely where the diagnosis is hardest. The mechanism is a genuine tumour-suppressor effect, biallelic inactivation of NF1, with loss of the second allele demonstrated within the glomus cells themselves, which is why these behave as real neoplasms rather than malformations.

Two practical consequences.

• Fingertip pain in a patient with NF1 is a glomus tumour until proven otherwise. These patients endure the longest delays of all, because digital pain is attributed to a neurofibroma or to the neurofibromatosis in general and the small subungual lesion is never sought. Ask specifically about cold sensitivity and pinpoint tenderness, and examine each fingertip with a pen tip.
• Run the inference the other way too. A patient with more than one painful digital glomus tumour, especially bilaterally, should prompt an examination for café-au-lait macules, axillary and inguinal freckling, cutaneous neurofibromas and Lisch nodules, and a family history; the hand may be where an undiagnosed NF1 first declares itself. Warn the patient that further lesions may appear over time, since a new painful digit later is a new tumour rather than a failed operation.

Plain radiographs. PA and true lateral views of the affected digit, obtained before MRI in every case to exclude exostosis, bone tumour or significant deformity. Look for a scalloped erosion of the dorsal cortex of the distal phalanx: it is highly suggestive of a long-standing subungual glomus tumour, the slow-growing lesion having remodelled the bone, and a useful pre-operative confirmatory sign when present. The film is normal in many cases, because the tumour is soft tissue and radiographs cannot see it, so a normal X-ray does not exclude a glomus tumour.

Ultrasound. High-resolution ultrasound is a quick, cheap screen in clinic, showing a small, well-defined, hypoechoic, hypervascular nodule. It is operator dependent, and a deeply subungual lesion may be obscured by the nail plate.

MRI. The imaging gold standard, obtained in all suspected glomus tumours before surgery, with a dedicated finger coil:

• T1
• T2 fat-sat or STIR
• Post-gadolinium T1 fat-sat

The lesion is a well-circumscribed, T2 hyperintense, strongly enhancing nodule beneath the nail. MRI localises occult or satellite lesions and guides the surgical approach, and modern high-resolution sequences (3 T with a dedicated surface coil) can identify lesions as small as 1 to 2 mm. Its limitations are cost and that it is not always available urgently.

Histopathology. Always send the excised specimen. Nests of uniform glomus cells around vascular channels, SMA positive and S100 negative, confirm the diagnosis and exclude melanoma or another lesion. It is available only after excision.

The operation is a marginal excision, and the approach is chosen by where the lesion sits.

The table gives the figures; the work-up of recurrent pain is in the Management section.

What to expect. Complete excision cures more than 90% of patients with immediate pain relief. The recurrence rate of 10-20% in the complications table usually reflects incomplete excision, a satellite lesion or a second glomus tumour, which is why the two figures are not in conflict.

Prognostic factors. The best prognosis is a single, well-localised subungual lesion treated by an experienced hand surgeon with intra-operative localisation and complete marginal excision. Prognosis is poorer with multifocal disease, an incomplete first excision, misdiagnosis (neuroma, CRPS), or a transungual approach that injures the nail matrix. Most patients are cured with one well-planned operation, and the single biggest predictor of a good outcome is accurate pre-operative localisation, which is why Love testing and MRI before surgery are not optional.