Lymphoid and histiocytic infiltrates

Meet the patient

A 62-year-old man reaches dermatology clinic with an "eczema" on his buttocks, hips, and waistline — the bathing-suit distribution — that has smouldered for three years through four different topical steroids. The patches are vaguely annular, faintly scaly, and oddly fixed in place. Biopsy shows a band-like lymphoid infiltrate hugging the epidermis with scattered atypical cells. This is mycosis fungoides wearing an eczema mask, and the three-year steroid odyssey is the classic preventable delay.[5]

Across the hospital, an eight-month-old with a stubborn seborrhoeic-looking scalp and flexural eruption has begun passing vast volumes of dilute urine, and the blood film shows thrombocytopenia. The scalp biopsy is dominated by CD1a-positive histiocytes with coffee-bean nuclei. This is multisystem Langerhans cell histiocytosis with pituitary involvement — a dermatology clue to a multisystem emergency, and humour-off territory.[8]

Both vignettes ask the single question every lymphoid or histiocytic infiltrate asks: is this reactive or neoplastic, and how far has it gone? Hold that question and the rest of this page slots into place.[1]

Three worlds, one microscope — name the world first

Every cutaneous lymphoid or histiocytic infiltrate falls into one of three worlds, and naming the world is the first diagnostic act. A dermal band of lymphocytes could be reactive pseudolymphoma, a primary cutaneous lymphoma, or secondary systemic disease; a dermal nodule of histiocytes could be Langerhans cell histiocytosis, a juvenile xanthogranuloma, or Rosai–Dorfman disease. The microscope alone never answers this — the clinic sets the frame.[1]

The three worlds, with the one-line discriminator beneath each:[1]

[1]

The WHO-EORTC classification is the language pathology reports and exam answers must speak: it sorts primary cutaneous lymphomas by cell lineage, clinical behaviour, and site rather than by histology alone, because a CD30+ large-cell infiltrate behaves very differently in skin than the same cells would in a lymph node.[1]

The 2018 update refined entity boundaries and remains the classification currency today, building on the 2005 framework that first proved clinicopathologic correlation beat pure histology for predicting behaviour in skin.[2]

The architecture fork — read the pattern before the name

Before naming any disease, describe the architectural pattern — where the cells sit tells you which family to consider before you stain the slide. Five patterns cover almost every lymphoid infiltrate you will meet:[1]

The five patterns and the family each one points to:[1]

• Band-like and epidermotropic — lymphocytes lined up along the dermoepidermal junction and spilling into the epidermis. Think mycosis fungoides until proved otherwise, with spongiotic mimics as the caveat.[5]
• Superficial perivascular — the default inflammatory pattern; usually reactive or an early inflammatory dermatosis, but it can hide early MF, so clinical correlation decides.
• Nodular dermal — a bottom-heavy wedge of lymphocytes filling the dermis. The territory of primary cutaneous B-cell lymphoma and B-cell pseudolymphoma.[6][3]
• Diffuse dermal — sheets from papillary dermis to subcutis. Broad differential: aggressive lymphoma, leukaemia cutis, or a dense reactive process; IHC and context resolve it.
• Folliculotropic or syringotropic — lymphocytes homing to hair follicles and eccrine coils, often with mucin. Points to folliculotropic MF and the follicular mucinosis spectrum.[5]

The teaching point beneath the list: pattern narrows the family before the stains are even run. A band says T-cell and think MF; a nodule says B-cell and think CBCL until the IHC sorts it.[1]

Pseudolymphoma — the reactive lookalike that mimics lymphoma

Pseudolymphoma is a clinicopathologic diagnosis of a reactive lymphoid hyperplasia that can closely simulate lymphoma on H&E. The name is literal — pseudo- (false) plus lymphoma — and the falseness is proven only by clinicopathologic correlation, never by any single stain or gene rearrangement.[3]

Etymology viva gold: pseudolymphoma is Greek pseudēs, "false", bolted to lymphoma. The term exists precisely to flag that the lesion looks neoplastic but behaves reactive — Mitteldorf and Kempf's 2020 review formalised the spectrum and the triggers that produce it.[3]

The classic triggers cluster around sustained antigenic stimulation — learn them as a cluster, because naming the trigger often ends the diagnostic anxiety:[3]

• Arthropod bites — tick, mosquito, spider; often a single nodular lesion on an extremity that resists resolution.
• Tattoo pigment — red (mercury or cadmium) is the classic culprit, sometimes years after placement.
• Infections — borrelia, syphilis, viral; consider the local epidemiology.
• Drugs — anticonvulsants, beta-blockers, ACE inhibitors; lymphomatoid drug eruption resolves on withdrawal.[4]
• Idiopathic — a confident trigger is not always found, which is exactly when clinical follow-up matters most.[3]

Histology favours the reactive end of the spectrum when the infiltrate is mixed (lymphocytes with eosinophils, plasma cells, and macrophages), top-heavy (denser superficially than deep), and arranged around germinal-centre-like follicles — but none of these is absolute, and early lymphoma can show any of them.[3]

Management is conservative once the diagnosis is confident: remove the trigger (drug, antigen source), then potent topical or intralesional corticosteroid for symptomatic lesions, or simple excision of a solitary diagnostic nodule. The 2018 systematic review found the treatment evidence largely observational rather than randomised — many lesions are best watched after a confident diagnosis.[4]

The classic trap: over-calling a clonal reactive infiltrate as lymphoma on a single nodular biopsy. Reactive infiltrates can be oligoclonal or even monoclonal on T-cell-receptor or immunoglobulin rearrangement, and treating them as lymphoma exposes the patient to chemotherapy they never needed.[3]

WHO-EORTC cutaneous T-cell lymphoma — the great imitator and its kin

CTCL arises from skin-homing memory T cells and spans the indolent-to-aggressive spectrum within a single classification. The four entities an examiner expects — MF, Sézary, LyP, and pcALCL — are divided by morphology, behaviour, and the CD30 marker, not by T-cell origin alone.[5]

The great imitator — mycosis fungoides

MF is the most common primary CTCL and the disease that hides for years as "eczema". It evolves through patches, then plaques, then tumours, often in sun-protected bathing-suit skin — buttocks, hips, waistline, breasts — and develops poikiloderma (mottled atrophy, telangiectasia, and pigmentation) as it matures.[5]

Etymology viva gold: mycosis fungoides was coined in 1806 by Alibert for the mushroom-like tumours of late-stage disease — there is no fungus involved. The early patch stage looks exactly like eczema, which is precisely why it is missed.[5]

Diagnosis is clinicopathologic and often sequential — a single early patch biopsy is rarely diagnostic, and repeat biopsies over months to years may be needed before the architectural and phenotypic clues (a band-like CD4-predominant epidermotropic infiltrate with loss of CD7) align with the clinical course.[5]

The classic trap: diagnosing lymphoma from a single early patch biopsy without clinicopathologic correlation. Early MF and chronic spongiotic dermatitis share histologic features, and a clone — if present — is supportive but never standalone.[5]

Management of early-stage MF is skin-directed first — potent topical corticosteroids, phototherapy (nbUVB or PUVA), topical mechlorethamine, or local radiotherapy for limited tumours — escalating to systemic therapy only for advanced or refractory disease.[5]

Sézary syndrome — the leukaemic CTCL

Sézary syndrome is the leukaemic variant of CTCL: erythroderma, generalised lymphadenopathy, and circulating malignant Sézary cells. It is aggressive, and the erythrodermic patient is fragile — barrier failure, heat loss, sepsis risk, and high-output cardiac stress all compound the malignancy itself.[5]

Diagnosis rests on the triad plus blood flow cytometry showing a clonal CD4-positive T-cell population with loss of surface markers (a CD4-to-CD8 ratio above 10, with loss of CD7 or CD26). Skin biopsy may show a band-like infiltrate similar to MF, but the blood defines the syndrome.[5]

Management is systemic from the outset — skin-directed measures for comfort only, because the disease needs haemato-oncology: extracorporeal photopheresis, targeted agents such as romidepsin or mogamulizumab, and transplant considerations in selected patients. This is not a topical-steroid problem.[5]

The CD30+ pair — LyP and pcALCL

Two CD30-positive lymphoproliferative disorders share a marker and divide on behaviour. Lymphomatoid papulosis (LyP) self-heals in crops; primary cutaneous anaplastic large cell lymphoma (pcALCL) grows as persistent tumours. Both are CD30-positive, both are primary cutaneous, and both carry a far better prognosis than their systemic ALCL namesake.[1]

[1]

The one-line discriminator: crops that come and go are LyP; a tumour that stays and grows is pcALCL. Both need long-term skin surveillance, because LyP in particular carries a small but real risk of a second lymphoma.[1]

WHO-EORTC cutaneous B-cell lymphoma — two indolent, one that kills

Primary CBCL presents as red-to-violaceous papules, plaques, or nodules in the dermis, and the entity name — not the histology — sets the threat. The three exam entities are PCMZL and PCFCL (both usually indolent) and primary cutaneous large B-cell lymphoma, leg type (aggressive).[6]

[6]

The one-line discriminator: head or trunk means indolent; leg means aggressive. Site and immunophenotype together separate the two indolent entities from leg-type CBCL, and getting this wrong under-treats a disease that behaves like systemic DLBCL.[6]

Indolent CBCL is managed with local radiotherapy, excision, or watchful waiting for low-burden disease, with rituximab-based systemic therapy for multifocal or refractory cases.[6]

The leg-type trap — red-blue tumours on elderly legs

PCLBCL, leg type presents as rapidly enlarging red-blue tumours on the lower legs of elderly women and behaves like systemic DLBCL. It is the CBCL that kills, and it is the one most often dismissed as a cyst, a lipoma, or an infected boil for months before biopsy is finally done.[6]

Treat it with systemic immunochemotherapy (an R-CHOP-style regimen) and radiotherapy, coordinated with haemato-oncology — simple excision alone is inadequate, and delay costs lives.[6][1]

The classic trap: calling leg-type CBCL an infected boil. A rapidly growing leg tumour in an elderly woman is leg-type CBCL until biopsy proves otherwise.[6]

The histiocytoses — L, C, R, M

Emile's 2016 revised classification reorganised the histiocytoses into four groups by cell lineage, and the L/C/R/M letters are now exam currency. The framework replaced a chaos of eponyms with a lineage map and is the language modern pathology reports use.[9]

[9] [7]

Langerhans cell histiocytosis — the coffee-bean cell

LCH is a clonal neoplastic proliferation of Langerhans-type dendritic cells, frequently harbouring MAPK-pathway mutations — most commonly BRAF V600E. The histiocyte carries the characteristic coffee-bean nucleus with Birbeck granules on electron microscopy and stains CD1a-positive, langerin (CD207)-positive, and S100-positive.[8][7]

Etymology viva gold: the Langerhans cell is named for Paul Langerhans, the German medical student who described it in 1868 — he thought it was a cutaneous nerve ending; it was actually the skin's antigen-presenting dendritic cell. The same man gave his name to the pancreatic islets.[7]

Cutaneous LCH classically presents in infants as a seborrhoeic-like scalp and flexural eruption with purpuric papules, erosions, and crusting — easily dismissed as cradle cap until systemic signs emerge. The skin lesion may be the only organ involved (single-system skin LCH, usually benign) or the first clue to multisystem disease.[8]

Multisystem LCH — the infant with diabetes insipidus

Multisystem LCH is a paediatric oncology emergency when risk organs — liver, spleen, or marrow — are involved, and the dermatologist is often first to the diagnosis. The combination of a seborrhoeic scalp eruption in an infant with polyuria and polydipsia (pituitary diabetes insipidus), cytopenias from marrow involvement, hepatosplenomegaly, or respiratory failure demands same-day paediatric oncology referral.[8]

Staging — not histology — sets the threat: single-system skin LCH may need only observation or topical therapy, while multisystem risk-organ LCH needs oncology protocols with cytarabine-based chemotherapy and MAPK-targeted agents (trametinib, vemurafenib) in refractory BRAF-driven disease. Dermatology alone is insufficient for multisystem disease.[8][7]

The JXG family — Touton giants and yellow papules

Juvenile xanthogranuloma is the commonest non-LCH histiocytosis, a CD68-positive and CD163-positive lesion of infants and young children that typically self-resolves. Its signature is the Touton giant cell — a large cell with a wreath of nuclei surrounded by foamy cytoplasm — and the lesions are yellow-orange papulonodules, solitary or multiple, on the head, neck, or trunk.[7]

Etymology viva gold: the Touton giant cell is named for Karl Touton, the German botanist-pathologist who described it in 1885. A wreath of nuclei ringed by foamy cytoplasm makes the cell look like a floral centrepiece — fitting for a man who spent his career on plants.[7]

Most solitary JXG need only observation; ocular screening for iris lesions and glaucoma risk is mandatory in infants with multiple facial lesions — the exam pearl that is also the preventable-harm reminder.[7]

Multicentric reticulohistiocytosis sits in the broader non-LCH family and announces itself with coral-bead periungual papules and a destructive polyarthritis — a skin clue to a systemic histiocytic process that needs rheumatology and oncology collaboration.[7]

Rosai–Dorfman disease — the wandering histiocyte

Rosai–Dorfman disease (sinus histiocytosis with massive lymphadenopathy) is marked by S100-positive histiocytes that engulf intact lymphocytes — emperipolesis. Skin disease may be isolated (cutaneous RDD) or part of systemic RDD with massive lymphadenopathy, fever, and weight loss.[10]

Etymology viva gold: emperipolesis is Greek — em- (in) plus peripolein (to wander round) — literally "wandering in and out". The histiocyte swallows a lymphocyte whole, the lymphocyte survives unharmed inside it, and the whole parade is visible on H&E as a lymphocyte wearing a histiocyte coat. Pathology's most literal name.[10]

Cutaneous RDD presents as nodules or plaques, often yellowish or red-brown, on the face, trunk, or extremities. Management ranges from observation or excision for isolated cutaneous disease to systemic therapy (steroids, methotrexate, or imatinib in selected cases) for extensive systemic disease — but always exclude systemic RDD with clinical and imaging assessment first.[10]

The lineage IHC cheat-sheet — stain before you name

Immunohistochemistry resolves the architectural differential, and the panel is small enough to memorise. Run it as two questions: what lineage, then which entity within that lineage.[1]

[1]

The two highest-yield stains in the whole topic: CD7 loss in a band-like CD4-predominant epidermotropic infiltrate points to MF, and CD1a or langerin positivity in a coffee-bean histiocyte confirms LCH. Everything else is refinement on those two anchors.[1]

The clonality trap — clonal does not equal lymphoma

T-cell-receptor and immunoglobulin gene rearrangement results are supportive, never diagnostic alone, and the misuse of clonality is the commonest over-call in dermatopathology. A clonal result does not prove lymphoma, and a polyclonal result does not exclude it.[3][5]

The two directions of error, both of which harm patients:[3]

• False alarm — reactive infiltrates can be oligoclonal or even monoclonal. An arthropod-bite pseudolymphoma may show a dominant T-cell clone that vanishes when the lesion resolves; treating it as lymphoma is the classic iatrogenic harm.
• False reassurance — early MF can be polyclonal on a single early biopsy, with the clone emerging only on repeat sampling over months. A negative result never closes the file.[3][5]

The rule that saves patients: clonality is read only in context — architecture, IHC, and clinical course together. No single test names lymphoma.[1]

At the bedside — before you send the biopsy

The biopsy is only as good as the clinical question that accompanies it. Map the skin (full body, photographed, with lymph-node basins palpated), and take a focused history before the histology request is written.[5]

The history that changes the differential:[5]

• Drugs, arthropod exposure, tattoos, prior radiation — the pseudolymphoma triggers.
• B symptoms — fever, drenching night sweats, weight loss — that force systemic work-up of any lymphoid infiltrate.
• Polyuria and polydipsia — pituitary diabetes insipidus in LCH; the single question that catches multisystem disease early.[8]
• Immunosuppression — widens the differential to EBV-associated lymphoproliferative disease and atypical infiltrates.[1]

Biopsy strategy matters as much as history. Choose a developed lesion, not a fading one; for suspected MF, take multiple biopsies over time because a single early patch is often non-diagnostic. Request H&E plus the lineage panel up front, and add molecular only when it will change interpretation.[5]

Staging triggers — when to escalate from skin to system

Most indolent entities stay in the skin; three groups force systemic staging immediately. Recognising the trigger is what separates dermatology-as-only-care from dermatology-as-clue-to-systemic-disease.[1]

The three staging triggers, with the work-up each demands:[1]

• Erythrodermic CTCL or suspected Sézary — skin scoring, blood flow cytometry for Sézary cells, and nodal or visceral imaging; haemato-oncology from the outset.[5]
• Aggressive CBCL (leg type) or any CBCL with systemic suspicion — staging CT or PET and haemato-oncology labs; treat as a systemic DLBCL-equivalent.[6]
• Multisystem LCH — skeletal survey, pituitary MRI for diabetes insipidus, marrow, liver, and lung assessment; paediatric oncology leads.[8]

The non-negotiable signs that should make you stop treating the skin alone:[1]

• Erythroderma with circulating atypical cells or progressive tumours — Sézary work-up, not chronic eczema.
• B symptoms with any lymphoid infiltrate — systemic disease until excluded.
• Polyuria, polydipsia, cytopenias, hepatosplenomegaly, or lytic bone lesions with a histiocytic infiltrate — multisystem LCH.[1][8]

Management — observe, local therapy, or systemic

Management maps cleanly onto the three worlds. Reactive disease is observed or treated locally; indolent lymphoma gets skin-directed or local therapy with long follow-up; aggressive lymphoma and multisystem LCH go straight to systemic protocols.[1]

[1]

Prognosis follows the same fork. Early MF and indolent CBCL carry excellent long-term survival with skin-directed care; Sézary syndrome and leg-type CBCL are substantially worse and need specialist systemic management; single-system skin LCH does well, while multisystem risk-organ LCH drives the prognosis of that whole family.[1][6]

How patients come to harm — the preventable list

• Years of "eczema" treatment before MF is diagnosed — the bathing-suit patches smoulder through four topical steroids before anyone biopsies; the three-year delay is the classic preventable harm.[5]
• Missing multisystem LCH in an infant with diabetes insipidus or cytopenias — calling a seborrhoeic scalp "cradle cap forever" while the pituitary and marrow quietly fail.[8]
• Under-treating leg-type CBCL as a cyst or "infected boil" — simple excision alone is inadequate; the disease behaves like systemic DLBCL.[6]
• Over-calling a clonal reactive infiltrate as lymphoma — exposing a patient with an arthropod-bite pseudolymphoma to chemotherapy they never needed.[3]
• Diagnosing lymphoma from a single early patch biopsy — without clinicopathologic correlation and appropriate IHC and molecular context.[5]
• Forgetting the eye in multiple facial JXG — missing iris lesions and glaucoma risk in an infant.[7]
• Stopping CD30+ LPD surveillance — LyP carries a small but real risk of a second lymphoma, so long-term skin follow-up is mandatory.[1]

The mantra, and the mnemonic

[1] [1]

The mantra: Pattern first, lineage second, clonality only in context — never name lymphoma on a single early biopsy.[1]

Ward-round test — three stems, thirty seconds each

Stem 1 — the bathing-suit eczema (answer)

A 62-year-old man has scaly, vaguely annular patches on his buttocks, hips, and waistline that have not cleared after three years of topical steroids. Biopsy shows a band-like CD4-predominant epidermotropic infiltrate with loss of CD7. What is the diagnosis, and what was the preventable harm? Model: This is early mycosis fungoides — the bathing-suit distribution, the chronic steroid-resistant course, and the band-like CD4-predominant epidermotropic infiltrate with CD7 loss are characteristic. The preventable harm is the three-year delay before biopsy; MF is a clinicopathologic diagnosis and a single early patch may be non-diagnostic, but a fixed steroid-resistant eruption in sun-protected skin warrants repeat biopsy and specialist referral. Start skin-directed therapy — a potent topical corticosteroid, phototherapy — and stage clinically.[5]

Stem 2 — the infant with polyuria and a seborrhoeic scalp (answer)

An eight-month-old with a crusted seborrhoeic scalp and flexural eruption develops polyuria, polydipsia, and thrombocytopenia. Scalp biopsy shows CD1a-positive and langerin-positive histiocytes with coffee-bean nuclei. What is the diagnosis, and what is the immediate next step? Model: This is multisystem Langerhans cell histiocytosis with pituitary (diabetes insipidus) and marrow involvement — a paediatric oncology emergency. The skin lesion is the clue; the diabetes insipidus and cytopenia mark risk-organ disease. Arrange same-day paediatric oncology referral, staging (skeletal survey, pituitary MRI, marrow, liver and spleen assessment), and oncology-based therapy — cytarabine-based chemotherapy, with MAPK-targeted agents in refractory BRAF-driven disease. This is not cradle cap.[8]

Stem 3 — the leg boil in an elderly woman (answer)

A 78-year-old woman has a rapidly enlarging red-blue tumour on her lower leg, growing over six weeks. The GP treated it as an infected boil with two courses of antibiotics. Biopsy shows sheets of large B cells with strong BCL2 and MUM1. What is the diagnosis, and what was the management error? Model: This is primary cutaneous large B-cell lymphoma, leg type — the aggressive CBCL — confirmed by the strong BCL2 and MUM1 immunophenotype. The management error is treating it as an infected boil for six weeks; leg-type CBCL behaves like systemic DLBCL and needs systemic immunochemotherapy (an R-CHOP-style regimen) plus radiotherapy coordinated with haemato-oncology. Simple excision alone is inadequate. Any rapidly enlarging leg tumour in an elderly woman is leg-type CBCL until biopsy proves otherwise.[6]

References

1. [1]Willemze R, Cerroni L, Kempf W, et al. The 2018 update of the WHO-EORTC classification for primary cutaneous lymphomas Blood, 2019.PMID 30635287
2. [2]Willemze R, Jaffe ES, Burg G, et al. WHO-EORTC classification for cutaneous lymphomas Blood, 2005.PMID 15692063
3. [3]Mitteldorf C, Kempf W. Cutaneous pseudolymphoma-A review on the spectrum and a proposal for a new classification J Cutan Pathol, 2020.PMID 31237707
4. [4]Miguel D, Peckruhn M, Elsner P. Treatment of Cutaneous Pseudolymphoma: A Systematic Review Acta Derm Venereol, 2018.PMID 29136262
5. [5]Jawed SI, Myskowski PL, Horwitz S, et al. Primary cutaneous T-cell lymphoma (mycosis fungoides and Sézary syndrome): part I. Diagnosis: clinical and histopathologic features and new molecular and biologic markers J Am Acad Dermatol, 2014.PMID 24438969
6. [6]Goyal A, LeBlanc RE, Carter JB. Cutaneous B-Cell Lymphoma Hematol Oncol Clin North Am, 2019.PMID 30497672
7. [7]McClain KL, Bigenwald C, Collin M, et al. Histiocytic disorders Nat Rev Dis Primers, 2021.PMID 34620874
8. [8]Krooks J, Minkov M, Weatherall AG. Langerhans cell histiocytosis in children: History, classification, pathobiology, clinical manifestations, and prognosis J Am Acad Dermatol, 2018.PMID 29754885
9. [9]Emile JF, Abla O, Fraitag S, et al. Revised classification of histiocytoses and neoplasms of the macrophage-dendritic cell lineages Blood, 2016.PMID 26966089
10. [10]Bruce-Brand C, Schneider JW, Schubert P. Rosai-Dorfman disease: an overview J Clin Pathol, 2020.PMID 32591351