Marjolin's Ulcer

Definition. A Marjolin's ulcer is a malignancy arising in a chronic wound, scar, sinus or ulcer, most commonly an old burn. Between 75 and 95 percent are squamous cell carcinoma; the remainder include basal cell carcinoma, melanoma, sarcoma and Merkel cell carcinoma. Marjolin described it in 1828, observing tumour change in old burn scars, and the term now covers every aetiology.

Why it matters. Chronic wounds can undergo malignant change, so any non-healing ulcer needs a biopsy, and the tumour that results behaves more aggressively than a de novo SCC, so the treatment is wider. Lower-limb burns, chronic osteomyelitis sinuses and venous ulcers are the typical sources seen in fracture-clinic and sarcoma-unit practice worldwide.

Who. Burn survivors in low- and middle-income countries where childhood burns were managed conservatively; patients with chronic osteomyelitis and a long-standing draining sinus, often for decades; elderly or immobile patients with long-standing venous, pressure or diabetic ulcers of the leg and foot; patients with previous radiotherapy and radiation-induced chronic skin change; and patients with chronic fistulae or surgical scars that have never epithelialised properly.

Where. The lower limb is the single most common site, 40-60 percent of cases, because burns and chronic osteomyelitis are concentrated there. Long-standing leg ulcers are the second most common source after burn scars. Rarer sources are chronic scarring dermatoses (lupus, dystrophic epidermolysis) and vaccination scars.

Time is the risk factor. The single biggest risk factor is time since the burn. Patients burned in childhood, especially in regions with limited acute burn care, are now reaching the latency window in middle age, and the incidence of Marjolin's is expected to rise in these populations. Vigilance in this group is cost-effective: a single targeted biopsy can change management.

Latency. Typically two to four decades, with a mean close to 31 years. Two independent series converge: Kowal-Vern's 412 pooled burn-scar neoplasms gave a mean latency of 31 years (mean age 20 at the burn, 50 at diagnosis), and Hahn's clinical series gave 31.5 years in a cohort that was 32 percent chronic osteomyelitis fistula. The longer a scar has been stable, the more sinister a new change.

The osteomyelitis-sinus caveat. The common teaching that latency is shorter after an osteomyelitis sinus is widely repeated but poorly sourced, and Hahn's osteomyelitis-heavy cohort showed no such shortening. Treat latency as long whatever the aetiology. Acute Marjolin's, under one year from the injury, exists but is rare and behaves more aggressively.

Two pathways to malignant change. The first is the chronic-inflammation, or scar-niche, model: persistent wound healing, repeated epithelial injury, free-radical damage, local immunosuppression and bacterial colonisation produce cumulative DNA damage in a setting of impaired immune surveillance. This is the dominant model for burn-scar and osteomyelitis-sinus Marjolin's. The second is direct malignant transformation of unstable scar epithelium, supported histologically by dysplasia at the wound edge adjacent to frank carcinoma in many specimens.

The chronic wound environment. The cellular events described are:

• Repeated epithelial injury and repair
• Sustained cytokine release (TNF-alpha, IL-6)
• Local immunosuppression by regulatory T cells
• Free-radical and bacterial toxin exposure
• Hypoxia and lactate accumulation

The genetic events reported:

• p53 mutation in a high proportion of cases
• Loss of p16 and other cell-cycle regulators
• UV-signature mutations in sun-exposed sites
• Field change: dysplasia extends beyond the visible tumour

View larger

Histology. The tumour is most often well- to moderately-differentiated SCC, with keratin pearls, intercellular bridges and invasive nests. The diagnosis is made when SCC is identified arising from, or directly adjacent to, the chronic wound epithelium; the dysplasia-carcinoma sequence is often visible at the wound edge, with adjacent carcinoma in situ.

Why it behaves worse than de novo SCC. Part of the answer is the tumour: often of higher grade than a de novo SCC, deep invasion by the time the diagnosis is made, a high rate of perineural and lymphovascular invasion, and a deep margin that is frequently poorly circumscribed. Part is the patient: older, with more comorbidities, often poorly nourished and chronically unwell, slow to present because the wound was long-standing, and with reconstructive options limited by the surrounding scar. The net effect is lower survival than de novo SCC of equivalent stage.

Three classifications are in use, by timing of the malignant change, by histology and by the lesion it arose in. The histology is the one that sets the operation.

History. Establish the original injury (mechanism, age at injury, healing course, treatments received) and the time since the wound in years, decades being typical for a burn scar. Then what has changed: new bleeding, pain, discharge, mass, foul smell or failure to heal. Ask about previous treatment (skin grafts, dressings, surgery, compliance with dressings), past radiotherapy, immune suppression and smoking, and about weight loss, night sweats, anorexia or bone pain, which suggest metastasis.

Examination. Look at the wound and everything around it:

• The wound: size, depth, edges (raised, everted, rolled) and floor (granulation, necrotic, fungating)
• Surrounding skin: satellite nodules, dysplastic change, the quality of the old scar
• Palpation: induration, fixation to deep structures, bone involvement
• Regional lymph nodes: inguinal, axillary, popliteal, epitrochlear or cervical depending on the site
• Distal neurovascular status, especially before any amputation decision
• Distant sites: lung auscultation, abdominal examination for hepatomegaly

The nodes at the first visit. Compare the contralateral limb and assess the regional nodes at the same consultation. A Marjolin's on the lower leg that has travelled to the popliteal or inguinal nodes is a different operation from one that is node-negative. Document the nodal status and use ultrasound or cross-sectional imaging if the nodes are clinically equivocal.

View larger

What needs a biopsy. Any chronic wound that fails to heal, or any stable scar that ulcerates, bleeds, becomes painful or develops a mass. A chronic osteomyelitis sinus with increased discharge or a fungating mass at its mouth is biopsied along with the tract, and a venous or pressure ulcer that has outlasted multiple skin grafts is in the same category. The technique is set out under Investigations.

The differential. A non-healing ulcer in an old scar has other explanations, and the table separates them; the biopsy is what settles it.

Biopsy first. Take a 4 mm punch or wedge biopsy from the edge of the lesion, including both abnormal and adjacent normal-appearing tissue. The floor is granulation tissue, slough or necrotic debris without diagnostic epithelium, so floor biopsies are often false negative. Large or heterogeneous lesions need multiple samples, at least two or three from the most suspicious (raised, indurated, painful) areas, and a benign result from a lesion that still looks suspicious is grounds for a second opinion.

The pathology request. Give the clinical context (Marjolin's suspected), ask for SCC versus another tumour type, and request depth of invasion, grade, lymphovascular and perineural invasion, and margins if the specimen is excisional. Discuss the suspicion explicitly with the pathologist so that multiple levels and immunohistochemistry are considered; immunohistochemistry is useful when the tumour is poorly differentiated (consider p16, p53 and Merkel cell polyomavirus).

Local staging. Plain radiographs of the limb are often obtained to look for bony involvement (chronic osteomyelitis, cortical breach, pathological fracture) and as a baseline for future comparison, but they do not exclude early marrow involvement. MRI of the region is the local staging study of choice for every confirmed case before excision: it shows the enhancing soft-tissue mass, any bone marrow oedema, cortical breach or sinus tract, and it guides the margin and the reconstruction by identifying invasion of the neurovascular bundle or joint.

Nodal staging. Regional nodes are involved in 20-40 percent at presentation, against under 5 percent for de novo cutaneous SCC. The direction is well accepted, but note the provenance: the largest pooled series (Kowal-Vern, 412 cases) found 22 percent regional nodal involvement, the bottom of that range, and made no direct comparison with de novo SCC. No study has matched the two for size, depth and grade, so part of the excess may simply be that these tumours are large and deep by the time anyone biopsies them.

• Ultrasound of the regional nodes is first line in every confirmed case and guides fine-needle aspiration of suspicious nodes
• CT or MRI of the nodal basin for deep nodes (pelvis, retroperitoneum) or when ultrasound is equivocal
• Sentinel lymph node biopsy is increasingly used because of the high rate of microscopic nodal disease, but the evidence is mixed for SCC; consider it in high-grade, deep, large lesions, in conjunction with plastic surgery or nuclear medicine

Distant staging. CT of the chest and abdomen for high-grade lesions, a large primary, clinically positive nodes, recurrent disease or systemic symptoms; CT chest is the standard distant staging study for any high-risk tumour. The lung is the most common distant site, then liver and bone. PET-CT is increasingly used in advanced or recurrent disease, or when cross-sectional imaging is equivocal, to identify occult metastatic disease.

The complications are those of the tumour, of the reconstruction and of the amputation, and the table sets out the rates and the risk factors.

Late recurrence. Marjolin's ulcer is one of the few cutaneous malignancies that can recur 10-20 years after apparently successful treatment, which is why follow-up is lifelong. A healed scar that changes again is re-biopsied and treated as a new Marjolin's until proven otherwise, and the patient is told at discharge to seek urgent review for any new bleeding, mass, pain, nodal swelling or non-healing in the scar, even decades later.

Nodal status is the single most important prognostic factor, so document it carefully. The survival figures by stage:

Favourable factors are a small, well-differentiated, node-negative tumour without lymphovascular or perineural invasion or bone involvement, excised with clear margins from an immunocompetent host.

Unfavourable factors. Any of the following worsens the outlook:

• Deep invasion (greater than 4-5 mm)
• High grade
• Lymphovascular invasion
• Perineural invasion
• Positive nodes
• Bone or joint involvement
• Recurrent tumour
• Immunocompromised host
• Latent period under 1 year (acute Marjolin)
• Older age and comorbidities

Against de novo SCC. A de novo cutaneous SCC follows years of cumulative sun damage, is usually superficial at diagnosis and is variable in grade, often well-differentiated. A Marjolin's follows decades of latency, is often deep by the time of biopsy, is often of higher grade, and carries the nodal rate set out under Investigations. Even stage for stage, Marjolin's carries a lower survival than primary cutaneous SCC because of higher grade, deeper invasion at presentation and more frequent nodal disease.