Erdheim-Chester disease

INTRODUCTION — Erdheim-Chester disease (ECD) is a rare non-Langerhans histiocytic multisystem disorder. ECD is most commonly manifest as multifocal sclerotic lesions of the long bones demonstrating sheets of foamy histiocytes on biopsy, with or without histiocytic infiltration of extra-osseous tissues. Since 1930, when ECD was first described by Erdheim and Chester, fewer than 1000 cases have been reported in the medical literature [1,2].

Histiocytic disorders are thought to be derived from mononuclear phagocytic cells (macrophages and dendritic cells) or histiocytes. This group has generally been divided into Langerhans cell histiocytosis (LCH) and non-Langerhans histiocytosis. LCH is so-named for its presumed derivation from the Langerhans cells (specialized dendritic cells found in the skin and mucosa). In contrast, non-Langerhans histiocytoses are thought to be derived from the monocyte-macrophage lineage.

The epidemiology, clinical manifestations, pathologic features, diagnosis, and management of ECD will be presented here.

Diagnosis and management of LCH are presented separately. (See "Clinical manifestations, pathologic features, and diagnosis of Langerhans cell histiocytosis" and "Pulmonary Langerhans cell histiocytosis".)

EPIDEMIOLOGY — ECD is rare, but the actual incidence is unknown. Fewer than 1000 cases have been reported in the published literature [2,3]. However, detection of ECD is expected to increase due to greater awareness and improved diagnostic precision.

ECD is most common in adult men, with a mean age of 50 to 60 years at diagnosis; the male predominance is estimated to be 3:1 [3,4]. Rare pediatric cases (<15 years of age) have been reported.

PATHOGENESIS — ECD is a malignancy of myeloid progenitor cells. Somatic mutation of BRAF and/or other signaling molecules appears to drive the malignant process, which is associated with increased expression of inflammatory cytokines. No infectious etiology or other environmental cause for ECD has been identified, and there is no evidence that ECD is inherited; the mutations associated with ECD are acquired (ie, somatic, not germline).

ECD is caused by clonal proliferation of myeloid progenitor cells, as demonstrated by detection of the characteristic BRAF V600E mutation in subsets of dendritic cells, mature monocytes, committed myeloid progenitors, and CD34+ cells of affected patients [5,6]. Hematopoietic cells that carry the BRAF V600E mutation can recapitulate the phenotype of ECD in vitro and in a mouse xenograft model.

Somatic mutations of BRAF or other components of the MAPK signaling pathway are present in most patients with ECD. In some studies, BRAF V600E has been detected in approximately half of ECD cases, but the incidence is likely higher with more sensitive techniques [7-12]. Mutant BRAF enhances cell proliferation and survival by activating the RAS/RAF/MEK/MAPK signaling pathway [13-16]. Mutations in NRAS, KRAS, ARAF, PIK3CA, MAP2K1, and ALK are also reported in ECD [13-22]. The role of targetable mutations in disease management is described below. (See 'Management' below.)

ECD histiocytes express a pattern of proinflammatory cytokines and chemokines that accelerate histiocyte recruitment and activation [23]. One study reported that, compared with controls, 37 untreated patients with ECD had higher serum levels of interleukin (IL)-6, interferon alpha, and MCP-1 and lower levels of IL-4 and IL-7; this pattern suggests perturbation of T cell helper 1 (Th-1) lymphocytes in ECD [4]. Another study reported that interferon gamma-expressing Th-1 lymphocytes were prominent in the cellular infiltrate of ECD [23].

CLINICAL MANIFESTATIONS — The clinical presentation varies with the extent and distribution of involved sites. Most patients have bony involvement at the time of diagnosis and the vast majority also have at least one extra-osseous site of involvement. A subset of patients is asymptomatic, with disease detected by imaging for unrelated conditions. Other patients with multisystemic involvement may have a rapidly progressive clinical course.

In a literature review that included data from 259 patients with histologically-proven ECD, the most common clinical presentations were bone pain (26 percent), neurologic findings (23 percent), diabetes insipidus (22 percent), and constitutional symptoms (20 percent) [3].

A retrospective case series of 37 patients reported involvement of the following sites, in order of decreasing frequency [4]:

●Long bones (95 percent)

●Maxillary sinus, large vessels, and retroperitoneum (59 percent each)

●Heart (57 percent)

●Lungs (46 percent)

●Central nervous system (41 percent)

●Skin (27 percent)

●Pituitary gland and orbit (22 percent each)

Another case series reported a similar pattern of clinical manifestations [24].

Bone — Symmetric diaphyseal and metaphyseal osteosclerosis of the lower extremities is nearly always present in ECD but, in one study, pain was reported by only half of the patients [25]. Bone pain is most commonly manifest as mild, persistent juxta-articular pain, particularly in the lower extremities. Radiographically, the typical presentation is bilateral and symmetric osteosclerosis of the long bones (image 1 and image 2). Osteosclerosis of the skull bones, particularly the facial bones, is also described [26]. The pattern of bony involvement distinguishes ECD from Langerhans cell histiocytosis (LCH), in which proximal limbs, pelvis, scapula, and skull involvement is more common.

Bone lesions that are not apparent on plain radiographs can be detected by bone scintigraphy, computed tomography (CT), magnetic resonance imaging (MRI), or positron emission tomography (PET). A retrospective series of 11 patients with biopsy-proven ECD reported bilateral and symmetric osteosclerosis of the diaphysis of the long bones in 98 percent of bone lesions visible on conventional radiographs; 83 percent of the lesions also involved the metaphysis [27]. Osteosclerosis was homogeneous in 35 percent and heterogeneous in the others. Partial epiphyseal involvement with sparing of the subchondral bone was also common, as were periostitis and endosteitis. Bone scintigraphy detects all bone lesions visible on radiographs (image 3). MRI detected all abnormalities noted by plain radiograph and scintigraphy, and also depicted replacement of the normal fatty bone marrow by heterogeneous signal intensity on T1- and T2-weighted spin-echo images (image 4).

Cardiovascular — Cardiac involvement with ECD is a significant source of morbidity and mortality and is present in the majority of patients. Cardiovascular involvement can include valve abnormalities, rhythm or conduction defects, and periaortic fibrosis along the entire course of the vessel. In a retrospective series of 72 patients with ECD, 31 percent of ECD-related deaths were secondary to cardiac involvement, including myocardial infarction, cardiomyopathy, and symptomatic valve disease [28].

In a series of 37 patients with ECD who were systematically screened with electrocardiogram (ECG), cardiac MRI, and/or CT, 70 percent had abnormal heart imaging and 12 of 32 had an abnormal ECG [29]. The most common finding was infiltration of the right heart (including pseudo-tumor) in half of the patients; periaortic fibrosis, periarterial infiltration of the coronary arteries, and pericardial thickening/effusion were also common. A study of 23 consecutive patients with ECD who underwent cardiac MRI corroborated these findings [30]. Ten patients had evidence of cardiac involvement by MRI (myocardial involvement in nine, pericardial involvement in nine). Heart disease was most commonly manifest as a posterior right atrial pseudotumor; also common was infiltration of the right atrioventricular sulcus with asymptomatic infiltration of the right coronary artery. Six patients had thoracic large-vessel involvement together with cardiac lesions, whereas only one patient had thoracic aorta involvement without cardiac disease.

Circumferential soft tissue sheathing of the thoracic and abdominal aorta ("coated aorta") is visualized by CT in two-thirds of patients [28,31]. Renal artery hypertension may result from involvement of the renal arteries, and coronary artery involvement may lead to myocardial infarction [31-33].

Central nervous system — Neurologic involvement is seen in up to half of cases [3,25,26]. Infiltration can involve the entire central nervous system (CNS), including intra-axial and extra-axial compartments. CNS involvement, whether symptomatic or asymptomatic, is an independent predictor of a worse outcome in patients with ECD [34]. CNS manifestations are pleiotropic, but neurodegenerative cerebellar disease is the most common neurologic complication (in up to 20 percent of patients) [2].

Unilateral or bilateral infiltration of the orbits, manifest as exophthalmos, retro-orbital pain, oculomotor palsies, or blindness, occurs in one-quarter of patients [35,36]. Neurologic involvement may also be manifest as excessive thirst or urination, loss of libido, headache, weakness, ataxia, dysarthria, exophthalmos, seizures, cognitive impairment, cranial nerve palsies, or spinal cord compression [37,38]. Cerebellar involvement and other CNS mass lesions are often multifocal. CNS lesions typically enhance with gadolinium on MRI [7]. ECD can infiltrate the dura, in which case it may be confused with a meningioma [39]. Cognitive impairments have been described in the absence of imaging abnormalities. In one study of 11 patients, volumetric MRI demonstrated diffuse reduction in cortical thickness and subcortical gray matter compared with age-matched controls [40].

Pituitary involvement commonly manifests as central diabetes insipidus and other endocrinopathies (eg, hyperprolactinemia, gonadotropin insufficiency, hypotestosteronism). Pre-existing diabetes insipidus and endocrinopathies typically persist despite a radiographic response to treatment, whereas diabetes insipidus rarely develops late in the disease course [3].

Management of CNS involvement and associated endocrinopathies is discussed below. (See 'Management of complications' below.)

Other organs — Infiltration of nearly every organ system has been reported with ECD [7].

●Kidney – Infiltration of perinephric tissues with a rind or mass-like lesion leading to "hairy kidney" is common and may cause hydronephrosis, ureteral narrowing, and slowly progressive renal insufficiency [7,41]. Acute renal insufficiency occurs less commonly. Percutaneous nephrostomy tubes are often required to alleviate ureteral obstruction. (See "Clinical manifestations and diagnosis of urinary tract obstruction (UTO) and hydronephrosis".)

●Pulmonary – One-quarter to one-half of patients have involvement of the pleura, lung parenchyma, or both [42]. Pulmonary involvement may be asymptomatic or may be manifest as dyspnea and/or cough [7]. Although dyspnea and progressive fibrosis leading to respiratory failure can occur, in one large series pulmonary involvement was not an independent predictor of decreased survival [43]. Plain films are often normal, but spirometry may show restrictive features and decreased diffusion capacity. CT findings include mediastinal infiltration, pleural thickening/effusion, centrilobular nodular opacities, ground glass opacities, or lung cysts [44]. Fluid from bronchoalveolar lavage may contain macrophages and foamy histiocytes [7]. Open-lung biopsies have demonstrated histiocytic infiltrates in a lymphangitic pattern with associated fibrosis and lymphoplasmacytic inflammatory infiltrates.

●Skin – Skin is involved in one-quarter of cases of ECD, in contrast to LCH, which frequently affects the skin [45]. One-third of patients have yellow plaques under the skin (xanthelasma), most commonly on the eyelids [7]. Other skin lesions are generally multifocal, reddish-brown, and papulonodular in appearance but have few other distinguishing characteristics. Pruritus can occur but is not a universal feature.

●Other – Involvement of other structures (breast, thyroid, testis, gingiva, kidneys, and spleen) is rare. Breast involvement usually presents as a palpable mass or nodule in one or both breasts [46]. Liver involvement may be detected on radiographic staging or manifest with abnormalities in transaminases, bilirubin, and/or alkaline phosphatase [47]. Renal involvement may manifest as renal failure, renovascular hypertension, or hydronephrosis [3].

Associated myeloid malignancies — There is an increased incidence of myeloid neoplasms among patients with ECD.

An international retrospective study of 189 patients with biopsy-proven ECD reported that 10 percent had an overlapping myeloid neoplasm (eg, myeloproliferative neoplasm [MPN], myelodysplastic syndrome [MDS], chronic myelomonocytic leukemia, or other MDS/MPN overlap syndrome) [48]. The incidence of myeloid malignancies is much higher than is encountered in the general population. Compared with patients who had ECD alone, patients with coexistent ECD and an associated myeloid neoplasm were older at the time of diagnosis of ECD (68 versus 57 years, respectively) and had inferior survival (82 versus 364 months). Hallmark driver mutations of myeloid neoplasms (eg, JAK2 V617F, CALR) may coexist with mutations associated with ECD (eg, BRAF V600E, MAP2K1).

Monitoring for myeloid malignancies is a component of follow-up care. (See 'Patient follow-up' below.)

INITIAL EVALUATION — Initial evaluation must document the diagnosis, define the extent and sites of disease, and determine the patient's functional status. This includes clinical assessment of the central nervous system (CNS), heart, and other organs; laboratory studies; and imaging.

Clinical — The medical history should assess the patient for constitutional symptoms (eg, fever, night sweats, weight loss), bone pain, skin lesions (eg, xanthelasma, rash), and possible organ involvement (eg, dyspnea, dysrhythmias, polydipsia/polyuria, gynecomastia, decreased libido, double vision, neuropsychiatric findings). Physical examination should also evaluate the patient for potential neurologic, cardiac, pulmonary, and cutaneous findings. Clinical manifestations are described above. (See 'Clinical manifestations' above.)

The patient's functional status should be assessed by Eastern Cooperative Oncology Group (ECOG) or Karnofsky Performance Status (table 1).

Laboratory — Baseline laboratory studies should include:

●Complete blood count (CBC) with differential

●Serum electrolytes, liver and renal function tests, vitamin B12 and thiamine levels

●C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR)

●Morning urine osmolality and morning serum cortisol; thyroid-stimulating hormone (TSH) and free T4; prolactin, luteinizing hormone (LH), follicle-stimulating hormone (FSH); testosterone (in males) (see "Clinical manifestations of hypopituitarism" and "Evaluation of patients with polyuria")

●Electrocardiogram (ECG)

Imaging — We suggest the following studies at the time of diagnosis:

●Computed tomography (CT) of chest, abdomen, and pelvis to include the entire aorta (see 'Cardiovascular' above)

●Positron emission tomography (PET)/CT; although bone scintigraphy can demonstrate the iconic radiologic features (image 3), unlike PET, bone scintigraphy does not assess extra-osseous involvement

●Magnetic resonance imaging (MRI) with contrast of the brain (with detailed examination of the sella turcica) and heart

●Echocardiogram or radionuclide ventriculography

Pathology — An adequate tissue biopsy is required to establish the diagnosis of ECD, distinguish it from other conditions, and identify mutations that may be amenable to targeted therapies. The biopsy should demonstrate the characteristic histopathologic findings of lipid-laden, "foamy" histiocytes with a distinctive immunophenotype in a proper cellular and/or fibrotic milieu.

For many patients, a shave biopsy of a cutaneous lesion can yield adequate diagnostic material and is the least invasive procedure. For other patients, biopsy of bone or another organ may be required. Regardless of the tissue source, we suggest multiple samples because histologic involvement may vary from field to field. It is important to recognize that decalcification of a bone specimen will render the sample unsuitable for molecular analysis. In addition to standard histologic and immunophenotypic analysis, biopsy specimens should undergo molecular testing with a gene panel or next-generation sequencing that can identify BRAF V600E and other relevant mutations (eg, NRAS, KRAS, ARAF, PIK3CA, MAP2K1, ALK). If a mutation is not detected, we suggest repeat testing using another involved site and/or genotyping modality. (See 'Pathogenesis' above.)

Involved tissues are infiltrated by sheets of lipid-laden (xanthomatous) histiocytes that typically have small nuclei and foamy cytoplasm. Multinucleated giant histiocytes (Touton cells) with a central ring of nuclei are commonly seen and there may be interspersed inflammatory cells and fibrosis (picture 1 and picture 2 and picture 3) [49,50]. Reactive small lymphocytes, plasma cells, and neutrophils are frequently admixed. Histopathologic findings may be highly variable, including absence of classic foamy histiocytic infiltrate, nonspecific inflammation admixed with fibrosis, or fibrosis alone with scant histiocytes [14].

ECD cells express CD14 (a receptor for lipopolysaccharide), CD68 (a lysosomal macrosialin), CD163 (a hemoglobin- and haptoglobin-scavenging receptor), Factor XIIIa (tissue glutaminase), and fascin (actin-binding protein) [2,51]. ECD cells do not express the Langerhans cell markers, CD1a or langerin (picture 4 and picture 1); Birbeck granules (an ultrastructural finding on electron microscopy in Langerhans cell histiocytosis) are absent and S100 is rarely positive [7].

DIAGNOSIS — ECD is a rare disease with protean manifestations that is often challenging to diagnose; delayed or erroneous diagnosis is common.

ECD may be suspected in a patient with unexplained bone pain (especially of distal extremities) in association with cutaneous, cardiac, or neurologic findings and abnormal imaging of bone [7,50,52]. (See 'Clinical manifestations' above.)

Diagnosis of ECD is based on identifying the distinctive histopathologic findings in an appropriate clinical and radiologic context [7]. Lesions typically demonstrate foamy or lipid-laden histiocytes admixed with reactive inflammatory cells and/or fibrosis. Symmetric osteosclerosis of the legs is nearly always present. Further description of the histopathology, immunohistochemical, and radiologic features that distinguish ECD from other histiocytoses are described in the differential diagnosis. (See 'Pathology' above and 'Differential diagnosis' below.)

DIFFERENTIAL DIAGNOSIS — ECD must be distinguished from other histiocytic and dendritic cell disorders, metastatic solid or hematopoietic neoplasms, and hemophagocytic lymphohistiocytosis/macrophage activation syndromes.

●Langerhans cell histiocytosis – Langerhans cell histiocytosis (LCH) and ECD are both histiocytic diseases that can involve multiple sites, most commonly bones. Skin involvement is more common in LCH. The two entities can usually be distinguished by morphologic and immunohistochemical evaluation. ECD tumor cells lack central nuclear grooves and Birbeck granules typical of LCH cells, and unlike LCH, they do not express CD1a or S100 [53]. Cases of concomitant LCH and ECD (ie, mixed histiocytosis) have been described [54]. (See "Clinical manifestations, pathologic features, and diagnosis of Langerhans cell histiocytosis".)

●Paget's disease and POEMS syndrome – The osteosclerotic lesions of bone in ECD can be confused for a variety of metabolic bone disorders, including Paget's disease and the POEMS syndrome (polyneuropathy, organomegaly, endocrinopathy, monoclonal protein, skin changes), the latter of which also causes endocrine abnormalities. While the bone abnormalities of ECD often are confined to bilateral and symmetric osteosclerosis of the diaphysis of the long bones, Paget's disease and POEMS syndrome generally cause less symmetric bone abnormalities and less specific localization. ECD can readily be distinguished from Paget's disease and POEMS syndrome by histologic and immunophenotypic findings on biopsy. (See "POEMS syndrome" and "Clinical manifestations and diagnosis of Paget disease of bone".)

●Hemophagocytic lymphohistiocytosis – Hemophagocytic lymphohistiocytosis (HLH) and the related macrophage activation syndrome are systemic disorders that demonstrate tissue infiltration by non-neoplastic histiocytes. Unlike ECD, these disorders demonstrate prominent hemophagocytic activity in the biopsy and are also characterized by a number of other findings including fever, hypertriglyceridemia, hypofibrinogenemia, hyperferritinemia, and bicytopenia. (See "Clinical features and diagnosis of hemophagocytic lymphohistiocytosis", section on 'Laboratory and radiographic abnormalities'.)

●Juvenile xanthogranuloma (JXG) – Juvenile xanthogranuloma (JXG) is a non-malignant proliferative disorder of histiocytic cells of the dermal dendrocyte phenotype. JXG belongs to the broad group of non-Langerhans cell histiocytoses and is typically a disorder of early childhood; by contrast, ECD typically occurs in middle-age or older adults. JXG presents in the first two years of life as a solitary reddish or yellowish skin papule or nodule (picture 5), most often on the head, neck, and upper trunk. Histologically and immunophenotypically, JXG is indistinguishable from ECD. JXG generally follows a benign course with spontaneous resolution over a period of a few years. Less commonly, skin lesions can be multiple (picture 6). Extracutaneous or systemic forms (brain, lung, kidney, spleen, liver, bone marrow, and retro-orbital tumors) are exceedingly rare and can be associated with considerable morbidity. (See "Juvenile xanthogranuloma (JXG)".)

●Rosai-Dorfman disease – Rosai-Dorfman disease (RDD, sinus histiocytosis with massive lymphadenopathy) is a macrophage-related disorder that most often presents as a systemic disorder involving lymph nodes and other organs [55-57] or rarely can be limited to the skin [58,59]. Unlike ECD, skin lesions are firm, indurated papules. Although the pathologic cells in RDD are the macrophages (CD14+, CD1a-, S100+/-, CD68+), RDD is histologically distinct from the other histiocytic diseases because the macrophages have normal-appearing lymphocytes residing in the macrophage cytoplasm (emperipolesis) [60]. (See "Peripheral lymphadenopathy in children: Etiology", section on 'Rosai-Dorfman disease'.)

●Other disorders

•Central nervous system involvement can be confused for metastatic solid or hematopoietic neoplasms or primary central nervous system tumors (including meningioma). (See 'Central nervous system' above.)

•Cutaneous involvement by ECD can mimic vasculitis, cutaneous lymphoma, or cutaneous involvement with LCH. (See 'Other organs' above.)

•Abdominal involvement is frequently confusing with primary or secondary retroperitoneal fibrosis, sclerosing mesenteritis, or a variety of retroperitoneal neoplasms including lymphoma and germ cell tumor. (See 'Other organs' above.)

•Cardiac involvement can mimic other infiltrative cardiovascular processes, including sarcoidosis, and pulmonary involvement can have the clinical appearance of any number of interstitial lung diseases. (See 'Cardiovascular' above.)

MANAGEMENT — Not all patients with ECD require treatment at the time of diagnosis. Treatment is typically reserved for patients with symptoms or those with evidence of central nervous system (CNS) involvement or organ dysfunction, as discussed below. (See 'Symptomatic patients' below.)

Treatment in the context of a clinical trial is recommended, whenever possible. (See 'Clinical trials' below.)

Prior to treatment, the patient should be evaluated with history and physical examination, laboratory testing, and imaging, as described above. (See 'Initial evaluation' above.)

Asymptomatic patients — For asymptomatic patients who have no evidence of organ dysfunction or CNS involvement (either symptomatic or asymptomatic), we suggest a period of observation rather than immediate treatment. For asymptomatic patients with evidence of CNS involvement or organ dysfunction, we suggest management as described for symptomatic patients. (See 'Symptomatic patients' below.)

Asymptomatic or minimally symptomatic disease is uncommon, but some cases of ECD follow an indolent clinical course. All patients should be thoroughly evaluated to detect subclinical CNS or organ involvement, as described above. (See 'Initial evaluation' above.)

Because there is no known cure for ECD and it is uncertain if targeted therapies alter the natural history of ECD, we suggest a period of observation in asymptomatic individuals to assess the trajectory of the illness and avoid adverse effects of treatment. Asymptomatic patients should be followed with periodic clinical evaluation, but it is not necessary to perform routine imaging or repeat biopsies unless new clinical or laboratory findings emerge. Follow-up of the patient with ECD is described below. (See 'Patient follow-up' below.)

Symptomatic patients — For patients with symptoms related to ECD and/or evidence of organ dysfunction or CNS involvement (either symptomatic or asymptomatic), we suggest treatment rather than observation alone. No treatment is optimal for all patients with symptomatic ECD and we encourage enrollment in a clinical trial, when possible.

Outside of a clinical trial, our approach is to initially treat with a targeted therapy (informed by the underlying mutation) rather than interferon alfa, glucocorticoids, or systemic chemotherapy, based on the favorable balance of benefit and toxicity.

Our approach follows:

●BRAF – For patients with a BRAF V600E mutation, we suggest initial treatment with vemurafenib (BRAF inhibitor). (See 'BRAF inhibition' below.)

For patients who do not respond to vemurafenib, we suggest enrollment in a clinical trial, treatment with another BRAF inhibitor (if available) or MEK inhibitor, or interferon alfa.

For patients who are responding to vemurafenib but are having substantial treatment-related toxicity, we suggest dose reduction or a brief treatment interruption to lessen adverse effects.

●Other mutations – For patients with mutations that affect other signaling molecules (eg, NRAS, KRAS, ARAF, PIK3CA, MAP2K1, and ALK), we suggest initial treatment with cobimetinib (or another MEK inhibitor). (See 'MEK inhibition' below.)

For patients who do not respond to or are intolerant of an MEK inhibitor, we suggest an alternate MEK inhibitor (if available) or interferon alfa.

●No mutation detected – For patients with no detected mutations, we suggest a repeat biopsy from an alternate site or using an alternate genotyping modality. If no mutation is detected with repeat biopsy, we suggest initial treatment with interferon alfa rather than a targeted therapy or systemic chemotherapy. We suggest not treating patients initially with systemic chemotherapy. (See 'Interferon alfa' below.)

There are few prospective therapeutic studies and no randomized clinical trials for ECD. Outcomes data come primarily from retrospective case series, as described below. Our approach is generally consistent with that proposed by an international panel of physicians with expertise in ECD and society guidelines [7]. (See 'Society guideline links' below.)

TARGETED AGENTS — Targeted agents are available that are active against many of the mutations that are found in ECD.

BRAF inhibition — Vemurafenib is a potent inhibitor of the kinase domain of mutant BRAF and has activity in ECD with activating mutations in BRAF (eg, V600E) [61-64]. Vemurafenib is approved by the US Food and Drug Administration (FDA) and it received orphan drug designation from the European Medicines Agency (EMA) for treatment of ECD with BRAF V600E mutation [64,65].

For patients with BRAF V600E, we suggest initial treatment with vemurafenib 480 mg twice daily by mouth; some experts suggest initial treatment with 960 mg twice daily. Regardless of the starting dose, the dose may require adjustment due to toxicity [64]. Treatment with vemurafenib should continue until disease progression or development of unacceptable toxicity.

The most common (>50 percent) adverse reactions are arthralgia, alopecia, fatigue, rash, skin papilloma, and electrocardiogram QT interval prolongation [64]. The most common serious adverse reactions (grade ≥3, seen in ≥10 percent) were squamous cell skin cancer, hypertension, maculopapular rash, and arthralgia. Concomitant administration with strong CYP3A4 inhibitors or inducers (table 2) should be avoided, if possible.

FDA approval was based on a study of 22 patients with BRAF V600 mutation positive ECD, 15 of whom had received prior systemic therapies [64]. Treatment with vemurafenib 960 mg twice daily achieved 55 percent overall response (OR; 5 percent with complete response [CR]). All patients required dose reduction to either 720 mg or 480 mg twice daily, but efficacy was maintained after dose reduction in responding patients. In another study, vemurafenib demonstrated at least a partial response (PR) in 6 of 18 patients with BRAF V600E positive ECD or Langerhans cell histiocytosis (LCH), and the remainder had stable disease [63]. In a study that evaluated treatment interruption in more than 50 patients taking vemurafenib or dabrafenib (an alternative BRAF inhibitor), most patients relapsed after BRAF inhibitor interruption, but all patients improved after the treatment was resumed [66].

Dabrafenib is a kinase inhibitor that is active against BRAF V600E and other kinase mutations in melanoma but is not currently approved for treatment of ECD [67]. In a case report, a patient with BRAF V600E responded to dabrafenib but experienced a recurrence 14 months later; BRAF V600E was not detected on a repeat biopsy, but a KRAS mutation was found [68].

MEK inhibition — Inhibitors of the MAPK/extracellular signal regulated kinase (MEK) pathway are active in melanoma that is resistant to BRAF inhibitors and they hold promise for treatment of ECD with mutations of other signaling molecules (eg, NRAS, PIK3CA, or the RAS-PI3K-AKT signaling pathway).

We consider treatment with an MEK inhibitor a reasonable option for patients who have not responded to or progressed on BRAF inhibitors, have mutations that predict sensitivity to an MEK inhibitor, or have wild-type BRAF. There is no preferred MEK inhibitor, and few data are available regarding treatment of ECD with MEK inhibitors.

Cobimetinib is approved by the FDA as a single agent for treatment of adults with histiocytic neoplasms. Trametinib has been approved by the FDA and EMA for treatment of melanoma, but it is not labeled for ECD.

The following studies reported treatment of ECD with MEK inhibitors:

●A study of 18 patients with histiocytic neoplasms (12 with ECD) reported that cobimetinib was effective independent of genotype (including mutations of ARAF, BRAF, RAF1, NRAS, KRAS, MEK1, and MEK2) [69]. The rate of OR, measured by positron emission tomography (PET), was 89 percent, with 72 percent CR and 17 percent PR. At one year, 100 percent of responses were ongoing and 94 percent remained progression-free.

●In a case report of relapsed ECD with an activating KRAS Q61H mutation, trametinib plus dabrafenib (BRAF inhibitor) resulted in an excellent response [68].

●Another case report demonstrated efficacy of cobimetinib in ECD with wild-type BRAF [70].

OTHER TREATMENTS

Interferon alfa — Interferon alfa is our preferred treatment for patients with newly diagnosed, symptomatic ECD who are not eligible, able to tolerate, or responsive to a targeted agent. Either conventional or pegylated interferon alfa may be used. The optimal regimen and duration of treatment are not defined, but patients are usually treated until disease progression or intolerable side effects occur. The risk-benefit ratio of interferon should be revisited periodically in patients who are on treatment for more than two years.

We suggest beginning pegylated interferon alfa at 135 micrograms weekly, and we titrate the dose upward to a maximum dose of 200 micrograms weekly as tolerated, if patients are not responding to the initial dose. Alternatively, conventional interferon alfa may be started at 3 million international units (MIU) three times a week and the dose is titrated upwards (to a maximum dose of 9 MIU three times a week) as tolerated, if patients are not responding to the initial dose. Patients on interferon should be monitored for infection, liver function abnormalities, thyroid abnormalities, and depression.

In a series of 53 patients with ECD, 87 percent received interferon alfa for a median of 19 months [34]. Doses ranged from 3 MIU to 9 MIU three times weekly with conventional interferon, or 135 to 200 micrograms weekly with pegylated interferon. The one-year and five-year survival rates were 96 and 68 percent, respectively. In this study, treatment with interferon alfa was an independent predictor of improved survival (hazard ratio 0.32; 95% CI 0.14-0.70), while central nervous system (CNS) involvement was an independent predictor of inferior survival.

Glucocorticoids — Glucocorticoids have demonstrated clinical activity in ECD, but have not demonstrated a survival benefit [34]. Glucocorticoids are generally reserved for patients who cannot tolerate more aggressive systemic therapies or who have very mild symptoms. For such patients, we consider prednisone 60 mg daily for two to four weeks followed by a slow taper over three to six months. Patients on long-term glucocorticoids should receive prophylaxis for Pneumocystis jirovecii (previously Pneumocystis carinii) infection and may benefit from prophylaxis for gastrointestinal ulcers. (See "Treatment and prevention of Pneumocystis pneumonia in patients without HIV", section on 'Prophylaxis' and "Major adverse effects of systemic glucocorticoids", section on 'Gastrointestinal effects'.)

Other systemic therapies — Treatment with systemic chemotherapy or an alternate biologic agent should not be considered for initial therapy of ECD. We consider these agents appropriate only for patients who fail to respond or develop intolerable side effects to targeted agents and/or interferon alfa. (See 'Management' above.)

It is difficult to judge the relative efficacy of various systemic agents for ECD because most reports are restricted to case reports or small case series. We consider the utility of systemic chemotherapy in ECD to be limited. We generally reserve cytotoxic chemotherapy for medically fit patients (eg, Eastern Cooperative Oncology Group [ECOG] performance status ≤2) (table 1) with substantial systemic disease. For other patients, we favor use of biologic agents.

●Cytotoxic chemotherapy – Cladribine is our preferred cytotoxic chemotherapy for ECD, based on case reports and our own clinical experience [71,72]. The largest reported experience with cladribine included 21 patients with ECD (9 treated in front-line, and 12 in later lines of treatment) [72]. The overall clinical response rate was 52 percent, including 6 percent complete response (CR) and 46 percent partial response (PR); median duration of response was 9 months. Cyclophosphamide plus prednisolone led to improved pulmonary findings in a patient who did not previously respond to prednisolone alone [73].

●Methotrexate – Low dose oral methotrexate may have activity in ECD [74]. High dose systemic methotrexate with leucovorin rescue may be useful in ECD with CNS involvement (given the ability of methotrexate to cross the blood brain barrier), although this approach has not been studied.

●IL-1-receptor antagonist – IL-1-receptor antagonist (IL-1RA) production is stimulated by interferon alfa, and there are case reports of recombinant IL-1RA (anakinra, canakinumab) inducing responses in a small group of patients who could not tolerate interferon alfa [75-81].

●Sirolimus – In an open-label trial of sirolimus plus prednisone in 10 patients with ECD, six patients had a partial response, two had stable disease, and two had progressive disease [82]. We offer sirolimus when the agents mentioned above have not proven efficacious or are poorly tolerated.

●Infliximab – A report described clinical responses to infliximab in two patients with cardiac involvement [83].

●ALK inhibitors – ALK rearrangements and recurrent kinase fusions of ALK and NTRK1 have been described in ECD [14]. There is limited clinical experience with these agents for ECD. Use of ALK inhibitors for lung cancer is described separately. (See "Anaplastic lymphoma kinase (ALK) fusion oncogene positive non-small cell lung cancer", section on 'First-line treatment'.)

●Imatinib – Imatinib has activity for ECD, but its role is unclear now that effective kinase inhibitors (eg, vemurafenib) are available [84,85].

Localized therapies — Surgery and radiation therapy (RT) have limited roles in the management of ECD and are used primarily to manage local and/or mechanical complications.

Surgery has no clear role in the management of ECD except to address mechanical complications, such as ureteral obstruction or repair/replacement of cardiac valves. RT has been used for local palliation, but ECD seems to be much less responsive to RT than Langerhans cell histiocytosis (LCH), and lack of response or in-field recurrence is fairly common [86]. The optimal dose and field are unknown. We generally use doses appropriate for aggressive lymphomas (40 to 50 Gy) when feasible, rather than the much lower doses utilized for LCH.

PATIENT FOLLOW-UP — There is no consensus regarding an optimal schedule or protocol for follow-up or monitoring response to therapy. We consider clinical evaluation more useful than routine repeated imaging, which may be discordant with symptoms.

Patients should be clinically assessed a minimum of every three to six months or more frequently as symptoms and organ dysfunction require. Follow-up visits should seek evidence of central nervous system (CNS) and organ involvement. Laboratory studies should include a complete blood count and differential and routine serum chemistry studies. We suggest routine annual electrocardiogram (ECG) and echocardiogram, even if no heart involvement was detected at the time of diagnosis. The frequency and mode of imaging is dictated by baseline disease status, organ involvement, and requirement for ongoing treatment. As an example, affected organs might be imaged every three to six months until stability is documented; the interval between studies should be increased as warranted by the clinical status. Because of an increased incidence of myeloid neoplasms in patients with ECD, bone marrow examination or other evaluation should be considered in patients with symptoms related to unexplained cytopenias or other hematologic abnormalities [48].

MANAGEMENT OF COMPLICATIONS — Complications of ECD may include mechanical compression of organs, and involvement of the central nervous system (CNS) may cause a variety of endocrinopathies. (See 'Clinical manifestations' above.)

Patients with ECD are at risk of developing potentially life-threatening complications due to the compression of normal structures (eg, cardiovascular involvement, ureteral compression). Mechanical measures such as heart valve replacement or percutaneous nephrostomy may also be necessary in selected patients.

CNS involvement may be subtle and difficult to recognize [38]. There is no consensus regarding optimal management of CNS involvement, but options include surgical resection, radiation therapy, high dose steroids, targeted agents, and other systemic treatments. A high percentage of patients with ECD will develop diabetes insipidus (DI) and other endocrinopathies due to hypopituitarism. Regardless of other therapies, endocrinopathies such as DI should be corrected. Pre-existing DI and endocrinopathies typically persist, even after radiographic regression of disease. (See "Clinical manifestations of hypopituitarism" and "Evaluation of patients with polyuria".)

PROGNOSIS — There is no known cure for ECD and historically the prognosis has been poor. However, the long-term effects of targeted agents are not well-defined. Case series of patients treated with interferon therapy have reported five-year overall survival rates of approximately 70 percent [34]. In the same series, involvement of the central nervous system was associated with adverse outcomes.

CLINICAL TRIALS — Often there is no better therapy to offer a patient than enrollment in a well-designed, scientifically valid, peer-reviewed clinical trial. Additional information and instructions for referring a patient to an appropriate research center can be obtained from the National Institutes of Health, ECD Global Alliance, Histiocyte Society, or the National Cancer Institute.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Histiocytic and dendritic cell neoplasms".)

SUMMARY AND RECOMMENDATIONS

●Erdheim-Chester disease (ECD) is a rare non-Langerhans histiocyte disorder that is most commonly characterized by multifocal osteosclerotic lesions of the long bones demonstrating sheets of foamy histiocytes on biopsy, with or without histiocytic infiltration of extraskeletal tissues.

●ECD is a malignancy of myeloid progenitor cells in which a somatic mutation of BRAF or another signaling molecule appears to drive the malignant process and create an inflammatory tissue milieu. (See 'Pathogenesis' above.)

●The clinical presentation varies with the extent and distribution of involved sites. Most patients with ECD have osseous involvement and at least one non-osseous site of involvement (eg, heart, central nervous system [CNS], skin, other organs). Occasional patients are asymptomatic, but most have multisystemic involvement with a progressive clinical course. (See 'Clinical manifestations' above.)

●Evaluation requires a history and physical examination, baseline laboratory studies, and imaging to assess the extent of organ involvement. (See 'Initial evaluation' above.)

●Diagnosis requires a characteristic histopathologic appearance in an appropriate clinical and radiologic context. Lesions typically demonstrate foamy or lipid-laden histiocytes admixed with reactive inflammatory cells and/or fibrosis. The biopsy specimen should provide sufficient material to enable mutation testing and distinguish ECD from other histiocytic and dendritic cell disorders, metastatic solid or hematopoietic neoplasms, and other conditions. (See 'Diagnosis' above and 'Differential diagnosis' above.)

●Asymptomatic – Not all patients with ECD require treatment at the time of diagnosis. For the occasional asymptomatic patient who has no evidence of organ dysfunction or CNS involvement (either symptomatic or asymptomatic), we suggest a period of observation rather than immediate treatment. (See 'Asymptomatic patients' above.)

●Symptomatic – For patients with symptoms related to ECD and/or evidence of organ dysfunction or CNS involvement (either symptomatic or asymptomatic), we suggest treatment with a targeted therapy (selected according to the underlying mutation) rather than interferon alfa, glucocorticoids, or systemic chemotherapy, based on the favorable balance of benefit and toxicity (Grade 2C). (See 'Symptomatic patients' above.)

•BRAF – For patients with a BRAF V600E mutation, we suggest initial treatment with vemurafenib (BRAF inhibitor). (See 'BRAF inhibition' above.)

•Other mutations – For patients with mutations that affect other signaling molecules (eg, NRAS, KRAS, ARAF, PIK3CA, MAP2K1, and ALK), we suggest initial treatment with an MEK inhibitor (eg, cobimetinib). (See 'MEK inhibition' above.)

•No mutation detected – For patients with no detected mutations, we suggest a repeat biopsy from an alternate site or using an alternate genotyping modality. If no mutation is detected with repeat biopsy, we suggest initial treatment with interferon alfa rather than a targeted therapy or systemic chemotherapy, based on the balance of benefit and toxicity (Grade 2C). We suggest not treating patients initially with systemic chemotherapy. (See 'Interferon alfa' above.)

●For patients who do not respond to or are intolerant to initial treatment, management suggestions are discussed above. (See 'Symptomatic patients' above.)

●The schedule and protocol for follow-up and monitoring response to therapy are discussed above. (See 'Patient follow-up' above.)