Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
The histiocytoses constitute a collection of rare hematologic diseases that resist easy classification, at least in part because of the imprecise definition of “histiocyte.” “Histiocyte” broadly refers both to cells of the macrophage lineage and to dendritic cells (DCs), only some types of which are derived from macrophages. The numerous descriptive and functional subsets of macrophages and DCs only add to the confusion. Nonetheless, their collective consideration is justified by a degree of uniformity in these childhood diseases that present with infiltration of bone, secondary lymphoid organs, or the liver, with or without concomitant involvement of other visceral organs. However, prognosis and treatment options are specific to the subtype, and thus an appreciation of the individual varieties of histiocytoses is mandatory.
This chapter is divided into two main sections that correspond to the two most common clinical histiocytic disorders, Langerhans cell histiocytosis (LCH) and hemophagocytic lymphohistiocytosis (HLH). Also included is a relatively brief discussion of non–Langerhans cell histiocytoses and Rosai-Dorfman disease. Each section will begin with a description of the normal counterpart(s) of the relevant histiocytes and their ontogeny to provide a context for describing how these cells are pathologically involved in the histiocytic disorders. Each section will also include a description of their clinical manifestations and current approaches to treatment.
Types of Dendritic Cells
The pathologic cells of LCH share some features with normal epidermal Langerhans cells (LCs), which are the primary antigen-presenting cells (APCs) of skin. LCs are one type of the general class of APCs known as DCs because of the characteristic dendritelike structure they assume when activated. DCs are, in turn, just one of several types of APCs that are called “professional” because they can, by themselves, fully activate naive T lymphocytes by providing both the antigen/major histocompatibility complex ligand for T-cell receptor binding and the accessory signals required for full activation.
Because DCs can also render T cells tolerant to specific antigens, they have become a major focus of basic immunologic research. This research has revealed a surprising level of functional and phenotypic diversity among DCs, which will undoubtedly be relevant to the proliferative disorders affecting these cells. Several authors have provided helpful guides to thinking about and classifying DCs. Overall, DCs can be divided into conventional and inflammatory types ( Box 64-1 ) . Conventional DCs are present in lymphoid and nonlymphoid organs in their basal state, and their primary function is to collect antigens for presentation to T cells in order to activate or tolerize them, depending on the antigen's source (i.e., nonself vs. self). Inflammatory DCs arise in response to specific inflammatory or infectious signals and are ordinarily not identifiable in the nonchallenged state. In addition to presenting antigen, these cells secrete cytokines and other mediators, including tumor necrosis factor (TNF), that enhance host defense.
Box 64-1
Modified from Shortman K, Naik SH: Steady-state and inflammatory dendritic-cell development. Nat Rev Immunol 7:19–30, 2007.
Dendritic Cells
Conventional Dendritic Cells
Migratory Dendritic Cells
Migratory dendritic cells (DCs) gather antigen in peripheral tissues, then migrate to regional lymph nodes to present antigen to T lymphocytes
Examples: Langerhans cells, dermal DCs, interstitial DCs
Lymphoid Tissue Dendritic Cells
Lymphoid tissue DCs gather antigen within lymphoid tissue and present antigen to T lymphocytes within the same lymphoid tissue
Examples: thymic DCs, splenic DCs
Inflammatory Dendritic Cells
Inflammatory DCs arise in response to inflammatory signals; they can gather antigen and present antigen to T lymphocytes and can secrete cytokines
Example: tumor necrosis factor and inducible nitric oxide synthase–producing DCs
Mucosal Dendritic Cells
Mucosal DCs are resident in mucosal surface tissues; they can gather antigen and present antigen to T lymphocytes and help direct responses toward activation or tolerance
Predendritic Cells
Predendritic cells can develop DC function directly in response to inflammatory stimuli
Example: plasmacytoid DCs, monocytes
Within the conventional DC category, one can distinguish between migratory DCs and DCs that reside in lymphoid tissue. Migratory DCs fulfill the sentinel role of this leukocyte class and include LCs, dermal DCs, and interstitial DCs of other organs. Migratory DCs shuttle between end organs that interface with the external world (e.g., skin or mucosa), in the case of LCs, and regional lymph nodes. This trafficking occurs at a low level in the basal state but can be greatly enhanced in the presence of foreign antigen or inflammatory challenges. The migratory DC takes up antigen in the periphery and, once activated, travels to regional nodes either to present antigen itself or to transfer antigen to resident DCs in the node, which then perform the presentation function to T cells. Although the molecular signals that control this migratory behavior are not understood in detail, chemokines and their receptors play a large role. In the case of LCs, resting cells express the chemokine receptor CCR6 whose ligand, CCL20, is secreted by cutaneous keratinocytes. Upon antigen uptake and activation, LCs downregulate CCR6 and in its place upregulate another chemokine receptor, CCR7, whose ligands, CCL19 and CCL21, are secreted by cells in regional lymph nodes. This receptor switch has the dual effect of neutralizing the LC's anchor to the skin and attracting them to regional lymph nodes.
In contrast, DCs that reside in lymphoid tissue are nonmigratory. They make up most of the DCs populating the thymus and spleen and about half of the DCs in lymph nodes. Unlike migratory DCs, which are mature when they appear in lymph nodes, resident DCs are immature, which allows them to take up, process, and present local antigens. Surface markers can distinguish several subsets of resident DCs, which are presumed to have specialized functions. In the mouse, these subsets include cluster of differentiation (CD)8+ and CD8− cells, and among the CD8− population are CD4+ and CD4− subpopulations. Specific DC subsets also have characteristic gene expression profiles.
LCs bear distinctive intracellular and extracellular markers that permit their distinction from DCs that reside in lymphoid tissue. The most characteristic surface marker is CD1a, and the most characteristic ultrastructural feature is Birbeck granules, which are pentilaminar “tennis racket”–shaped cytoplasmic organelles that appear to be uniquely present in LCs. Langerin, also known as CD207, is a cell surface marker for LCs that associates with Birbeck granules when internalized. Langerin is a C-type lectin with specificity for sugars that contain mannose, suggesting that it may be involved in processing and trafficking of antigens that bear these sugars. Dectin-2 is another C-type lectin restricted to DCs, and DEC-205 (CD205) is yet another lectin that is often used to identify LCs, although its expression is not as restricted to LCs as are the other lectins. In the appropriate histologic context, from a diagnostic standpoint, CD1a and Langerin are the most reliable lineage-specific markers.
Some authors also distinguish a group of so-called “predendritic cells” that do not ordinarily have DC functions or structure. However, these cells can develop directly into DCs upon stimulation without the need for proliferation. One well-studied example is the plasmacytoid DC, which circulates as a nondescript mononuclear cell but acquires APC function and secretes large amounts of alpha interferon when activated. Technically, using this definition, monocytes would also be predendritic cells. After exposure to interleukin (IL)-4 and granulocyte-macrophage colony-stimulating factor (CSF), monocytes become functional DCs, although their phenotype is closer to inflammatory DCs than conventional DCs.
Origins of Dendritic Cells
The availability of genetically modified mice that carry lineage-restricted markers has greatly expanded our understanding of the origins and development of DCs. Although mouse models have some limitations, the major insight these animals provide is that a salient characteristic of DC ontogeny is flexibility. For example, although DCs generally have myeloid characteristics, all DC subtypes in the mouse can arise from either committed lymphoid or committed myeloid precursors. Even though half of thymic DCs show evidence for immunoglobulin rearrangements, suggesting a lymphoid phenotype, they also transcribe the macrophage colony-stimulating factor (M-CSF) receptor gene, which is a myeloid characteristic. Activation of fms-like tyrosine kinase–3 (FLT3), a growth factor receptor, is required for DC development, and it has been suggested that committed myeloid or lymphoid precursors expressing FLT3 can differentiate into DCs in the presence of FLT3 ligand.
Splenic DCs have a relatively rapid turnover rate (every 3 days). These cells are repopulated both by DC division and splenic precursors, as well as by bone marrow precursors. In contrast, LCs are very long-lived. After their migration from the skin, the population is replenished from a pool of LC precursors in the skin, as well as by circulating monocytes. The latter is dependent on the action of the chemokine CCL2 acting on its receptor CCR2. Thus the ontogeny of LCs is distinct from that of splenic DCs. Immature DCs are more closely related to LCs but appear to be derived from noninflammatory monocytes that do not express CCR2. Plasmacytoid DCs split off from conventional DCs early in development, consistent with their distinct functional characteristics.
Langerhans Cell Histiocytosis
The hallmark of LCH is the accumulation of LC-like DCs in one or more tissues or organs. The clinical manifestations of LCH are highly diverse. They result both from the direct, local effects of the growth and accumulation of pathologic LCs and the indirect, secondary effects these activated cells have on normal tissues, particularly cells of the immune system. Although the most commonly involved sites of disease are the bone and skin, virtually any organ or system may be involved. The pattern of involvement is not necessarily related to patterns of normal LC or DC migration.
Although the histiocytes in LCH share certain characteristics with LCs, such as expression of CD1a and the presence of Birbeck granules, recent reports suggest that mature LCs may not be the cell of origin for LCH histiocytes. In particular, a global analysis of gene expression suggested that LCH cells are more closely related to myeloid dendritic precursors than to LCs. Whether this pattern of gene expression reflects a non-LC cell origin for LCH or is a consequence of LC transformation remains to be determined.
The fact that pathologic LCs are related to immunomodulatory cells and that they elicit inflammatory infiltrates suggests that LCH might be a reactive rather than a neoplastic disease. Indeed, ample precedent for this mechanism exists among the histiocytoses, including the secondary HLH syndromes that arise in the context of viral infection. However, no reproducible reports exist of viral genomes recovered from LCH cells, and epidemiologic studies are not consistent with an infectious or environmental cause of LCH. Rather, the preponderance of evidence indicates that LCH arises as a consequence of intrinsic genetic abnormalities. The most salient arguments that LCH is a neoplasm and not a reactive disease are as follows: (1) pathologic LCs are clonal, as demonstrated by nonrandom X chromosome inactivation both in whole LCH tissue (in a proportion corresponding to the proportion of CD1a+ cells in the lesion) and in sorted CD1a+ cells, and (2) nearly 60% of LCH samples carry the oncogenic BRAF V600E variant. Notably, the MAPK pathway, which would be activated constitutively in the presence of BRAF V600E, is uniformly activated in LCH cells regardless of the presence or absence of BRAF V600E, suggesting that alternative mechanisms of pathway activation are present in LCH samples that lack BRAF V600E. In some cases, other activating mutations of BRAF have been found. Additional known mechanisms of BRAF activation, such as gene amplification or translocation, which occur in other diseases, have not been described in LCH. Notably, although pulmonary LCH in adults is generally considered to be a polyclonal disease, more than 40% of pulmonary LCH cases contain BRAF V600E. This finding suggests that a significant proportion of these cases are clonal or that multiple independently developed clones of pulmonary LCH have BRAF V600E, giving the overall appearance of polyclonality.
Although BRAF mutations are among the most common molecular abnormalities found in all cancers, they are not present in other histiocytoses that affect children, such as juvenile xanthogranuloma (JXG) or Rosai-Dorfman disease. However, BRAF V600E mutations are present in the histiocytes of about 50% of patients with Erdheim-Chester disease, a rare and aggressive non-LCH of adults. Furthermore, treatment of patients who have BRAF V600E–positive Erdheim-Chester disease with a BRAF inhibitor led to dramatic clinical responses, indicating the pathogenetic importance of this mutation. Interestingly, some of these patients had concomitant LCH that also responded to BRAF inhibition, suggesting that this therapeutic approach may be effective in patients who have BRAF V600E–positive LCH as well.
Activating mutations of BRAF are insufficient by themselves to produce neoplastic disease, and therefore additional molecular alterations are likely to be required for the development of LCH. One of the most common molecular abnormalities in LCH is overexpression of TP53 , suggesting that mutational inactivation of this pathway may contribute to pathogenesis. More such changes in other genes are likely to be discovered as advanced genomic analyses are applied to LCH samples.
Incidence
The incidence of LCH is difficult to determine precisely because of the rarity and marked clinical variability of the disorder. Estimates are in the range of 2.6 to 8.9 cases per million per year for children younger than 15 years, corresponding to roughly one tenth the incidence of acute leukemia in childhood. LCH occurs in people of all races and all ages. Although the peak age at diagnosis is 2 years, LCH can present at any time from birth to old age. No evidence of seasonal variation has been noted in the time of presentation of LCH. Several studies have shown a slight male predominance.
No known predisposing factors exist for the development of LCH in the majority of cases. Studies have demonstrated concordance of the disorder in monozygotic twins, suggesting an inherited predisposition to LCH that might account for ~1% of cases. However, a positive family history is lacking in most cases. Interestingly, isolated pulmonary LCH in adults is closely linked to cigarette smoking, and it often resolves with smoking cessation. However, exposure to cigarette smoke has not been linked to pediatric LCH. As previously noted, a significant proportion of pulmonary LCH cases also carry the activated BRAF V600E allele.
An intriguing (although incompletely understood) association exists between malignancy and LCH. Children with a history of cancer have an increased risk of LCH. Conversely, children with a history of LCH appear to have an increased risk of cancer. This finding suggests the possibility that underlying genetic abnormalities may place individual patients at increased risk for LCH. Alternatively, the association between LCH and cancer may fall within the spectrum of secondary posttherapy effects, because irradiation and drugs such as etoposide are used to treat LCH and are known to induce tumor-promoting genotoxic injury. Several cases of LCH arising in the setting of T-cell acute lymphoblastic leukemia have been described, typically when the leukemia is in remission while the patient is receiving therapy or in the early posttherapy period. A clonal link between the initial leukemia and pathologic LCs of LCH has been established in some cases, suggesting that the LCH represents a manifestation of the malignant clone rather than a distinct reactive or neoplastic process related to the leukemia or its therapy. An intriguing link may exist to hairy cell leukemia, which also harbors BRAF V600E in the majority of cases.
Clinical Features
The clinical manifestations of LCH are protean. Although patterns of clinical presentation have been described, the disease is appropriately viewed as a spectrum. LCH may involve a single site, multiple sites in a single organ system, or multiple organ systems. Many patients present with localized pain, soft tissue swelling, or skin rash. Less commonly, patients have symptoms of diabetes insipidus (DI), respiratory insufficiency, cytopenias, lymphadenopathy, liver dysfunction, or organomegaly.
In the past, three distinct clinical syndromes were described: eosinophilic granuloma, Hand-Schüller-Christian disease, and Letterer-Siwe disease. “Eosinophilic granuloma,” a term still commonly used by radiologists and orthopedists, refers to the presence of one or more lytic bone lesions. Hand-Schüller-Christian disease consists of the triad of bone defects, exophthalmos, and polyuria. Letterer-Siwe disease, a fulminant disorder of the reticuloendothelial system, is characterized by hepatosplenomegaly, lymphadenopathy, skin rash, bone lesions, anemia, and the tendency to bleed. This classification has important historic significance, but its applicability in the clinical setting is limited. Once it was recognized that the diverse manifestations of the disease share common histopathologic features, the three syndromes were unified under the term “histiocytosis X.” Categorizing patients based on the number and location of lesions and the presence or absence of organ involvement was shown to be useful in predicting prognosis and determining therapy. Single-system LCH (SS-LCH) disease—which usually affects bone, less frequently affects skin, and rarely affects the lymph nodes, lung, or central nervous system—accounts for approximately two thirds of pediatric LCH cases. Involvement of two or more organ systems, referred to as multisystem LCH (MS-LCH), accounts for the remaining one third of cases. In about half of MS-LCH cases, “risk” organs (i.e., the liver, spleen, and hematopoietic system) are affected. Thus far the presence of BRAF V600E does not correlate with disease extent or prognosis.
Pathology
A biopsy of lesional tissue is required to establish the diagnosis of LCH. The pathologic diagnosis is usually straightforward, but because the disease is rare, varied, and may mimic many other conditions, delay in diagnosis is common. LCH should be considered in any patient who presents with skeletal lesions, a persistent rash, chronically draining ears, central DI, unexplained lymphadenopathy, respiratory insufficiency with a reticulonodular pattern on a chest radiograph, hepatosplenomegaly, or cytopenias.
Once clinical suspicion is raised, the workup should proceed with a biopsy of the most accessible site of disease. Complete surgical excision of the lesion is generally unnecessary either for diagnosis or treatment. LCH lesions at all sites share common histopathologic features: accumulation of large neoplastic LCs with a moderate amount of dense pink cytoplasm and plump, distinctive “C-shaped,” “coffee bean,” or cleaved nuclei admixed with variable numbers of inflammatory cells, including T lymphocytes, macrophages, plasma cells, and eosinophils. The composition and architecture of the infiltrate characteristically differ according to the location. Osteolytic bone lesions often contain many nonneoplastic osteoclasts, as well as multinucleated tumor giant cells and large numbers of eosinophils, the latter speaking to the origin of the term “eosinophilic granuloma.” Mature, “burned out” bone lesions may be difficult to distinguish from chronic osteomyelitis, which is often in the clinical and radiographic differential diagnosis. LCH infiltration of the skin typically involves the superficial dermis, with tumor cells infiltrating the epidermis (so-called “epidermotropism” or “exocytosis”), thus altering epidermal barrier function and leading to superinfection and the characteristic appearance of some LCH rashes. In all cases, the neoplastic LCs express the LC markers CD1a and langerin, as well as S100 protein and fascin ( Fig. 64-1 ). Electron microscopy, which historically is used to demonstrate the presence of Birbeck granules indicative of the LC lineage, has little diagnostic utility at the present time. Cytogenetics or molecular studies are predominately needed to exclude other diseases with similar clinical presentations or histologies. Furthermore, no immunohistochemical or other laboratory methods are useful in distinguishing clinically indolent from clinically aggressive forms of the disease.
Diagnostic Evaluation
In suspected LCH cases, evaluation for disease extent is indicated, starting with a history and physical examination. A skeletal radiographic survey is recommended to comprehensively assess skeletal involvement. Abnormalities detected on plain radiographs may be followed by axial imaging to document the presence of a soft tissue mass. Nuclear imaging (i.e., a bone scan or fluorodeoxyglucose positron emission tomography [FDG-PET]) complements the skeletal survey. Compared with plain radiographs, nuclear imaging tests detect active lesions on the basis of increased local metabolic activity. Active bone lesions usually show increased radiotracer uptake, whereas inactive lesions may be undetectable or appear “photopenic.” FDG-PET appears to be the most sensitive imaging modality for the detection of lesions in persons with LCH, and it is particularly useful for following up on patients. Unlike a bone scan, PET is capable of detecting extraosseous lesions.
Complete blood cell counts and liver function studies should be performed, and liver and spleen size should be assessed by physical examination. Abdominal ultrasound is indicated when laboratory studies are abnormal or if hepatosplenomegaly is present or suspected. The erythrocyte sedimentation rate may be elevated, but it is not a sensitive or specific disease indicator. Measurement of specific gravity of an early morning urine sample is an easy and inexpensive screening test for DI. A formal water deprivation test may be necessary to establish the diagnosis of DI. Magnetic resonance imaging (MRI) of the head should be performed in children with cranial bone lesions and when DI is suspected or confirmed. A chest radiograph should be performed in all cases, and a computed tomography (CT) scan of the chest should be performed if the radiograph findings are abnormal or if the patient has respiratory signs or symptoms. Additional evaluations such as upper/lower endoscopy and biopsies of bone marrow, skin, lymph nodes, and the liver should be considered in the appropriate clinical context.
Management
The approach to management of LCH must account for the variability in its clinical behavior. On one hand, LCH is an uncontrolled accumulation of a clonal population of cells, with the capacity to behave extremely aggressively and the potential to involve multiple sites and organ systems. The process may be driven, in part, by oncogenic variants of BRAF . In addition, it may be effectively treated with cytotoxic chemotherapy and radiotherapy and, based on preliminary reports, BRAF antagonists. On the other hand, LCH is histologically benign, it sometimes resolves spontaneously, and it may respond to immunomodulatory or immunosuppressive agents. The historic uncertainty about the fundamental nature of LCH is reflected in the broad range of therapies that have been used to treat it.
Performing controlled clinical trials in persons with LCH has been challenging. As a consequence, the LCH literature is often descriptive, anecdotal, and retrospective. Since the 1980s, several collaborative groups have conducted prospective clinical trials that have led to standardization and improvements in treatment. An overall observation of these studies is that the treatment of LCH should match the clinical scenario. Patients with localized disease have an excellent prognosis with little or no intervention. For these children, the goal of therapy is to minimize symptoms and avoid long-term disability and potential late effects. In contrast, children with MS-LCH, especially those with involvement of the hematopoietic system or liver, are at significant risk for morbidity and mortality. For these children, aggressive treatment is justified.
Bone
Bone is the most commonly involved site, with bone involvement occurring in approximately 75% of children with LCH. Although any bone(s) may be affected, the bones of the skull, pelvis, and vertebral bodies are most commonly affected ( Fig. 64-2 ). Bone pain (which is often worse at night), swelling, and a limp are typical presenting symptoms. Sometimes a history of local trauma precedes or coincides with the onset of symptoms. The typical plain radiographic appearance is of a smooth-edged, “punched-out” hole in the bone. Lesions of the long bones typically involve the diaphysis, but metaphyseal and epiphyseal lesions also occur. Involvement of a vertebral body may manifest as flattening or loss of height (vertebra plana) or as a wedge deformity. Axial imaging (CT and MRI) may show an associated soft tissue mass ( Fig. 64-3 ). The radiographic differential diagnosis of osseous LCH lesions includes malignancy, especially Ewing sarcoma, leukemia or lymphoma, and osteomyelitis.
LCH of the bone may be unifocal (monostotic) or multifocal (polyostotic), and it may present in association with disease in other organ systems. Monostotic disease, which is the most common presentation of LCH (especially in older children), occurs in more than 50% of cases. Children with LCH that is limited to a single bone have an excellent prognosis regardless of the treatment administered. Although various local and systemic therapies appear to be efficacious, retrospective trials have failed to demonstrate a comparative advantage for any specific intervention or modality, including observation only, biopsy, curettage, simple excision, intralesional steroid instillation, local radiotherapy, or systemic therapy.
Often the biopsy procedure performed to establish the diagnosis of LCH of the bone provides therapeutic benefit. In many cases, adequate tissue for diagnosis may be obtained via percutaneous needle biopsy with ultrasound or CT guidance. Neurosurgeons tend to prefer open biopsy of orbital and skull base lesions and simple excision or curettage of calvarial lesions. Regardless of the surgical approach, most patients experience progressive healing after the procedure, even when complete excision of the lesion is not achieved. In cases in which LCH is strongly suspected or confirmed intraoperatively via frozen section, methylprednisolone can be instilled directly into the lesion. This intervention yields rapid relief of pain and may hasten resolution. Intralesional steroid injection is less commonly used in the management of cranial and vertebral lesions.
In the past, radiotherapy was routinely used to treat bone lesions. Even though a relatively low dose in the range of 600 to 1000 centigray is effective, the role of radiotherapy in pediatric LCH has waned because the long-term risk of radiotherapy, especially the risk of secondary malignant neoplasms, is rarely justified in treatment of a benign disease. In the setting of disseminated disease, local radiotherapy contributes little benefit.
Systemic therapy may be required for patients with monostotic LCH who experience unrelenting symptoms or are at significant risk for serious or permanent disability related to the location or extent of their lesion. This principle applies in patients with unifocal craniofacial bone involvement, including the orbit, temporal bone, mastoid, sphenoid, zygomatic, ethmoid, maxilla, paranasal sinuses, or cranial fossa (so-called “central nervous system [CNS]-risk” lesions), because these children are at increased risk for DI and other neurologic sequelae. Currently it is standard practice to administer chemotherapy in “CNS-risk” cases, even though it has not been definitively established that treatment decreases the risk of neurologic complications.
Systemic therapy is also recommended for children with polyostotic (i.e., multifocal bone) LCH. In these patients, an important goal of treatment is to mitigate the risk of future reactivation, because reactivation risk is approximately 40% for patients with multifocal bone disease versus 10% for patients with unifocal bone disease. Children with SS-LCH of the bone who require systemic therapy—that is, those with CNS-risk lesions or multifocal bone involvement—are typically treated with the combination of vinblastine and prednisone (discussed later).
Skin
The skin is the second most commonly involved system in persons with LCH, with skin involvement occurring in about one third of patients. Skin rashes may be extremely variable, and misdiagnosis is common. Cutaneous LCH may present as single or multiple nodules, vesicles, or scaly, seborrhea-like patches or plaques. Although the rash may be located anywhere on the body, it is typically most intense in the flexural areas, such as the neck, axillary, and inguinal folds. Involvement of the scalp, the ear canals, and the posterior auricular areas is also characteristic. Ear canal involvement may lead to chronic ear drainage that may be malodorous. Perineal and perianal rashes may be severe, persistent, and refractory to topical preparations.
Cutaneous involvement is more common in infants than in older children. Some infants with isolated cutaneous LCH prove to have a benign, self-healing disorder known as Hashimoto-Pritzker disease. However, in infants who initially present with apparently isolated cutaneous LCH, MS-LCH may proceed to develop, and therefore self-healing cutaneous LCH is a diagnosis that can be made with accuracy only in retrospect. The differential diagnosis of cutaneous LCH may include superficial candidiasis, seborrheic dermatitis (“cradle cap”), eczema, contact dermatitis, and viral exanthem. A skin biopsy is required to establish the diagnosis.
When treatment for cutaneous LCH is required, topical medications such as corticosteroid or nitrogen mustard may suffice. Occasionally, systemic therapy is necessary to control extensive or severe skin involvement, especially when it is associated with pain, disfigurement, or infection. Alternatives such as psoralen ultraviolet A or interferon have been used with success in some patients who have resistant cutaneous involvement.
Central Nervous System
The brain and skull can be affected by LCH in a variety of ways, including extension of extraaxial bone-based lesions, hypothalamic-pituitary disease, intraparenchymal mass lesions, and later onset neurodegeneration.
Hypothalamic-pituitary disease is the most common and best-characterized CNS manifestation of LCH. Infiltration of the pituitary stalk by lesional cells leads to deficiency of antidiuretic hormone and clinical DI. Anterior pituitary deficiencies or panhypopituitarism may develop in some cases. The diagnosis of DI may precede, coincide with, or occur years after the diagnosis of LCH. Estimates of the frequency of DI in persons with LCH vary considerably because of variability in study populations. One large retrospective review reported a 12% incidence of DI, with 6% having DI present at the time of LCH diagnosis. In this analysis, the risk of DI correlated significantly with the location and extent of LCH involvement. Children who have craniofacial lesions (CNS risk) or MS-LCH are at significantly higher risk for DI.
Among children with a new presentation of DI, LCH is the cause in a significant proportion (approximately 15%) of cases. Absence of posterior pituitary hyperintensity (the pituitary “bright spot”) or thickening of the pituitary infundibulum may be noted on brain imaging; however, these radiographic findings are neither sensitive nor specific for LCH. Their differential diagnosis includes brain tumors, especially germinoma, and inflammatory conditions such as hypophysitis. Investigation for other sites of LCH is worthwhile because detection and biopsy of extracranial lesions may allow histopathologic confirmation of the diagnosis without performing a pituitary stalk biopsy.
Screening for DI in children who have LCH with use of a careful history is essential, and referral to an endocrinologist should be considered if symptoms of polyuria, polydipsia, nocturia, or dehydration are elicited. Urinary specific gravity greater than 1.015 weighs against a diagnosis of DI, and this reading can often be easily determined by analysis of a first-morning voided urine specimen. In cases in which the history or laboratory studies raise suspicion, a formal water deprivation test should be performed.
In general, DI is irreversible once it appears in patients with LCH. However, improvement has been reported in “early” DI cases when patients were treated promptly with chemotherapy. At one time, emergent radiotherapy was recommended in patients with new onset DI. This recommendation has fallen out of practice because it is rarely, if ever, effective and is associated with potential long-term consequences. Chemotherapy has no role in patients with established, long-standing DI who have no evidence of active LCH within or outside of the CNS.
Parenchymal CNS mass lesions in persons with LCH are composed of granulomas of CD1a+ cells admixed with other inflammatory cells. These lesions can occur in isolation or in association with disease in other organ systems. Depending on the extent and location, headaches, seizures, and focal neurologic symptoms may result. CNS mass lesions are managed with chemotherapy, surgery, or radiotherapy.
Progressive neurodegeneration is an uncommon and poorly understood late event that arises years after LCH diagnosis. Radiographically, it is characterized by signal abnormality confined to the brain stem and cerebellum on MRI. Posterior fossa symptoms or neurocognitive dysfunction may be clinically apparent. Biopsy results show gliosis, neuronal cell loss, and lymphocytic infiltration without active CD1a+ cell infiltration. The timing and course of this manifestation of LCH is quite variable, but it may be severe and progressive. No established treatment exists for neurodegeneration in patients with LCH. Retinoic acid or low-intensity chemotherapy with or without immunoglobulin may stabilize the disease or slow the rate of neurologic decline.
Other Sites
Lymph node involvement is occasionally present in the setting of MS-LCH, and sometimes it represents a single site of disease. The pathologic pattern of lymph node involvement is characteristically interfollicular and subtle.
The gastrointestinal tract is another uncommon site of disease in LCH, but its incidence may be underappreciated. When it does occur, it is usually in the context of multisystem involvement. Clinical signs include diarrhea, bloody stools, failure to thrive, and hypoalbuminemia. Gastrointestinal biopsies typically demonstrate a sparse infiltrate of LCs in the lamina propria.
Pulmonary involvement was historically thought to be an adverse prognostic feature, warranting its classification as a “risk” organ in early Histiocyte Society clinical trials. Involvement of the lung may be asymptomatic in up to half of affected children, with diffuse micronodular interstitial disease or cyst formation. Overt respiratory symptoms including tachypnea, chronic cough, dyspnea, and pneumothorax are seen in some patients. In children, pulmonary involvement occurs in the context of MS-LCH and is often accompanied by hematologic or hepatic dysfunction. Recent retrospective studies have shown that in the absence of other risk organ involvement, pulmonary involvement is not an independent negative prognostic indicator. Interestingly, adults with LCH often present with disease limited to the lung, and their illness is closely linked to cigarette smoking. Cessation of smoking leads to resolution in the majority of patients. Although traditionally described as a polyclonal disease, as many as 30% of cases may be clonal, and a significant proportion harbor BRAF V600E.
You're Reading a Preview
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here