Erdheim-Chester disease is a very rare xanthogranulomatous, non-Langerhans cell systemic histiocytosis with an unknown etiology and pathogenesis. Histologically, it is characterized by a diffuse infiltration with large, foamy histiocytes, rare Touton-like giant cells, lymphocytic aggregates, and fibrosis. The histiocytes differ from the Langerhans cell group in ontogenesis, immunohistochemistry (positive for CD68 and negative for CD1a and S100 protein), and ultrastructural appearance (lack of Birbeck granules). Although most of the cases have symmetric osteosclerosis of the long bones, an involvement of the axial skeleton has also been described. Extraskeletal lesions are present in more than 50% of the patients and may involve the retroperitoneal space, lungs, kidneys, brain, retro-orbital space, and heart. This study presents the case of a patient with Erdheim-Chester disease with vertebral destruction and, for the first time, to our knowledge, involvement of the liver. The diagnosis is based on radiologic, histologic, immunohistochemical, and ultrastructural findings.

Originally described in 1930 by William Chester1 as a ”lipoid granulomatosis,” Erdheim-Chester disease (ECD) is a very rare xanthogranulomatous non-Langerhans cell histiocytosis. The etiology and pathogenesis are unknown, and whether ECD is a polyclonal-reactive process or a monoclonal neoplastic disorder is still controversial.2,3 It usually affects adults, with a slight male predominance, in their fourth to seventh decades of life.4 Since Chester's initial description, only approximately 80 cases have been reported in the literature. It is a systemic disease with a broad spectrum of clinical manifestations. The diagnosis is based on radiologic and histologic findings. In most of the cases, the bone involvement is constant and includes symmetric osteosclerosis of the diaphyses and metaphyses of the long bones, especially of the lower extremities.4,5 However, in fewer circumstances, an involvement of the axial skeleton with mixed osteolytic and osteosclerotic lesions has been described.6 There are associated extraskeletal lesions in more than 50% of the cases. They include hypothalamic/pituitary involvement with resultant diabetes insipidus and infiltration of the retroperitoneal space, retro-orbital tissue, and skin. ECD can lead to heart failure because of a myocardial or pericardial infiltration. Other sites of involvement are the lungs, kidneys, brain, and choroid plexus.3,4 

We report the case of a patient with ECD with a unique presentation: vertebral osteolytic lesions and, for the first time in the literature, to our knowledge, liver involvement.

A 32-year-old Hispanic man with no significant medical history presented with physical complaints of progressive motor weakness in his bilateral lower legs. He also described an inability to walk for 5 days and bladder and bowel incontinence for 2 days. This occurred after 2 months of intermittent back pain following a fall on his buttocks.

A radiographic skeletal survey showed pathologic fractures of the L2 and T4 vertebral bodies. Further evaluation included a computed tomographic scan and magnetic resonance imaging of the spine. They showed diffuse osteolytic lesions with a “moth-eaten” appearance throughout the thoracic and lumbar spine (T2–L4) (Figure 1), cord compression at T4 to T9, L2, and L4, and a soft tissue paraspinal mass at T1 to T2. A biopsy of the paraspinal mass and vertebral bodies was subsequently performed. Multiple hypodense nodular areas with maximum diameters ranging from 0.5 to 3.0 cm were also noted in the computed tomographic scan of the liver (Figure 2), and a laparoscopic liver biopsy was performed. The computed tomographic scan and radiograph of the chest showed no evidence of pulmonary lesions. No abnormalities were found on physical examination, except for reduced strength, reduced sensory function for both lower extremities, and mild tenderness on liver palpation. The liver function tests showed a mild cholestatic syndrome and a moderate increase in transaminases. The viral hepatitis serology was negative. The laboratory findings of normochromic, normocytic anemia prompted a bone marrow biopsy and aspirate.

Figure 1.

Computed tomographic scan of the spine demonstrating multiple vertebral lytic lesions throughout the thoracic and lumbar spine (arrows). Figure 2. Computed tomographic scan of the liver showing multiple hypodense nodular lesions (arrow) ranging in largest dimension from 0.5 to 3.0 cm. Figure 3. Liver biopsy showing diffuse and extensive replacement of the normal hepatic tissue by a histiocytic infiltrate, lymphocytic aggregates, and fibrosis (hematoxylin-eosin, original magnification ×100). Figure 4. High-power image of the liver biopsy demonstrating the presence of numerous foamy, lipid-laden histiocytes (hematoxylin-eosin, original magnification ×400). Figure 5. Immunohistochemical stain showing strong positivity of the foamy histiocytes for CD68 (CD68, original magnification ×100). Figure 6. Ultrastructural image of the liver biopsy showing histiocytes with numerous secondary lysosomes and phagolysosomes (arrow) and without Birbeck granules (original magnification ×5400).

Figure 1.

Computed tomographic scan of the spine demonstrating multiple vertebral lytic lesions throughout the thoracic and lumbar spine (arrows). Figure 2. Computed tomographic scan of the liver showing multiple hypodense nodular lesions (arrow) ranging in largest dimension from 0.5 to 3.0 cm. Figure 3. Liver biopsy showing diffuse and extensive replacement of the normal hepatic tissue by a histiocytic infiltrate, lymphocytic aggregates, and fibrosis (hematoxylin-eosin, original magnification ×100). Figure 4. High-power image of the liver biopsy demonstrating the presence of numerous foamy, lipid-laden histiocytes (hematoxylin-eosin, original magnification ×400). Figure 5. Immunohistochemical stain showing strong positivity of the foamy histiocytes for CD68 (CD68, original magnification ×100). Figure 6. Ultrastructural image of the liver biopsy showing histiocytes with numerous secondary lysosomes and phagolysosomes (arrow) and without Birbeck granules (original magnification ×5400).

Close modal

The patient's severe back pain was treated with intravenous steroids, after which he showed significant clinical improvement.

Four months later, the patient was readmitted for another episode of back pain with cord compression and neurologic symptoms. A second biopsy of the paraspinal mass (T1–T2) was obtained, and ultimately, the patient underwent a laminectomy and external beam radiotherapy.

Multiple-core needle biopsies of the left and right hepatic lobes were received in surgical pathology. The entire specimen was fixed in formalin and routinely processed for hematoxylin-eosin, trichrome, and diastase-treated and nontreated periodic acid–Schiff stainings. A microscopic examination of the hematoxylin-eosin–stained slides revealed an extensive infiltration and replacement of the normal hepatic tissue by foamy, lipid-laden histiocytes. Diffuse lymphocytic inflammation, fibrosis, and scattered Toutonlike giant cells without well-formed granulomas were also present (Figures 3 and 4). The trichrome stain highlighted the diffuse fibrosis, and no cytoplasmic periodic acid–Schiff-positive granules were identified. After the initial microscopic evaluation of the specimen, additional stains for acid-fast bacilli, fungi (Gomori methenamine silver), and amyloid (Congo red) were also performed; all were negative. The following immunohistochemical studies were performed on the formalin-fixed, paraffin-embedded sections by the avidin-biotin-peroxidase technique: CD68 (clone Kp1, dilution 1:600; Dako Corporation, Carpinteria, Calif), CD1a (clone MTB1, dilution 1:20; Novocastra, Newcastle upon Tyne, United Kingdom), and S100 protein (clone 4C4.9, dilution 1:250; Cell Marque, Hot Springs, Ark). A small fragment of tissue from a paraffin-embedded section was also submitted for ultrastructural examination by electron microscopy.

An examination of the immunohistochemical stains revealed that the foamy histiocytes were strongly positive for CD68 (Figure 5) but negative for S100 protein and CD1a. This immunophenotype excluded a Langerhans cell origin for these cells; Langerhans cells can be definitely identified by their strong staining with anti-CD1a antibodies. The ultrastructural study (electron microscopy) did not demonstrate the presence of Birbeck granules (characteristic for Langerhans cell histiocytes) and showed numerous secondary lysosomes and phagolysosomes, indicating the macrophagic origin of these cells (Figure 6). The presence of rare lipid vacuoles was not suggestive of a lipid storage disease.

The initial biopsies of the paraspinal mass and lytic vertebral lesions were inconclusive for purposes of diagnosis, showing distorted fragments of fibroadipose tissue, skeletal muscle, and focal lymphocytic infiltrates. The paraspinal mass biopsy obtained during the patient's subsequent hospitalization showed a diffuse proliferation of pleomorphic histiocytes with abundant, vacuolated cytoplasm and was positive for CD68 and negative for S100 protein.

The bone marrow biopsy showed a thickened trabecular bone and decreased cellularity. There were also several foci of atypical histiocytic aggregates with accompanying fibrosis that replaced the normal bone marrow. The histiocytes had an immunohistochemical pattern similar to those from the liver biopsy, specifically being positive for CD68 and negative for CD1a.

The differential diagnosis of a histiocytic proliferation with bone and liver infiltrates can be very challenging. The infectious etiology should first be considered and ruled out, as in our case, by clinical history and the lack of any pathologic organisms in the sampled tissue. The storage disorders (eg, Gaucher disease, Niemann-Pick disease, mucopolysaccharidoses) may be also a consideration. The diagnosis of these disorders is based on specific histologic findings, electron microscopy, or identification of the enzymatic abnormality. The histiocytic syndromes are proliferative disorders of unknown etiology, and distinction among them can be very difficult. Numerous classifications for these disorders have been proposed, but the most widely used is based on the origin of the histiocytes.7 This classification divides these syndromes into 3 classes. Class I disorders include Langerhans cell histiocytosis (also referred to as histiocytosis X) and is further divided into Hand-Schüller-Christian disease, Letterer-Siwe disease, and eosinophilic granuloma. Class II disorders are associated with non-Langerhans cell histiocytes as in Rosai-Dorfman disease or ECD. Class III disorders include the malignant histiocytic disorders.7 On the basis of immunohistochemistry and ultrastructural characteristics, the histiocytes can essentially be divided into 2 groups: Langerhans cells with a dendritic cell origin and non-Langerhans cells with a monocytic-macrophagic origin. Langerhans cell histiocytes stain positive for CD68, CD1a, and S100 protein, and electron microscopy shows Birbeck granules in their cytoplasm.4,8 Non-Langerhans histiocytes, found in ECD, are positive for CD68 but negative for S100 protein and CD1a; they also lack the Birbeck granules. According to the literature, there is a certain variability concerning the positive staining with S100 protein; it is positive in a small number of cases but absent in most.4,8 Langerhans cell histiocytes have been reported in ECD, but they should represent a minority of cells; however, they were not present in our case.4 Rosai-Dorfman disease is classically encountered in childhood and is characterized by painless lymphadenopathy and rare, asymmetric osseous lytic lesions involving mainly the long bones. The histiocytes of Rosai-Dorfman disease show emperipolesis (intact lymphocytes in their cytoplasm) and are strongly positive for S100 protein.8 

The clinical presentation, radiographic findings, and histologic characteristics of the presented case are all consistent with a diagnosis of ECD. Although, in most ECD cases, the initial clinical presentation is bone pain of the lower extremity with almost pathognomonic diametaphyseal sclerosis, cases with a spinal involvement have been reported, as occurred in our patient.6 In Langerhans cell histiocytosis, the bones seem to be affected randomly, with a frequent skull involvement and mainly osteolytic lesions.8 However, radiographic overlaps of these 2 entities have been described. The literature also describes patients who have both biopsy-proven Langerhans cell histiocytosis and radiographic features of ECD, making the differential diagnosis even more difficult.9 In our case, the presence of a diffuse histiocytic infiltrate in the liver, the bone marrow, and the paraspinal mass with similar ultrastructural features and immunohistochemical patterns (positive for CD68 and negative for S100 protein and CD1a) supports the diagnosis of ECD with a systemic involvement.

Although extraosseous manifestations of ECD have been extensively described in the literature, no hepatic involvement has previously been reported. Hence, to our knowledge, the case presented in our study is the first published report of ECD with a liver involvement. The liver biopsy showed an extensive, almost complete replacement of hepatic tissue by non-Langerhans histiocytic infiltrates, marked fibrosis, and chronic inflammation. The pathogenesis of fibroblastic proliferation in ECD is not completely understood, but it is known that fibrosis is a constant component of the histology of this disease in all organs or tissues involved.10 Given the reported evolution of patients with lung lesions, we can only speculate that progressive diffuse fibrosis of the liver parenchyma will occur, with further deterioration of the patient's condition.

In the reported cases, treatment usually included oral steroids and, in more severe cases, chemotherapy (eg, vinblastine, vincristine, cyclophosphamide) or external radiation therapy with variable outcomes.4,11 Clinical trials for the treatment of ECD have not been conducted. It is thought that the prognosis depends on the extent and distribution of extraskeletal disease. In a review of 59 patients with ECD, Veyssier-Belot et al4 reported death related to disease in 59% of the cases, with an average survival of 32 months. The most common causes of death were respiratory and heart failure. In the same study, the authors also suggested that patients with ECD have a worse prognosis than patients with Langerhans cell histiocytosis.

Although ECD is a rare condition, a knowledge of its morphologic and clinical manifestations and a high index of suspicion are essential in order to make the correct diagnosis.

Chester
,
W.
Uber lipoidgranulomatose.
Virchows Arch Pathol Anat
1930
.
279
:
561
602
.
Chartist
,
J.
,
V.
Paradis
, and
D.
Dargere
.
et al
.
Chester-Erdheim disease: a neoplastic disorder.
Hum Pathol
1999
.
30
:
1093
1096
.
Al-Quaran
,
S.
,
J.
Reith
, and
J.
Bradley
.
et al
.
Erdheim-Chester disease: a case report, PCR-based analysis of clonality, and review of literature.
Mod Pathol
2002
.
15
:
666
672
.
Veyssier-Belot
,
C.
,
B.
Cacoub
, and
D.
Capparros-Lefebvre
.
et al
.
Erdheim-Chester disease: clinical and radiologic characteristics of 59 cases.
Medicine
1996
.
75
:
157
169
.
Kenn
,
W.
,
A.
Stabler
,
R.
Zachoval
,
C.
Zietz
, and
W.
Raum
.
Erdheim-Chester disease: a case report and literature overview.
Eur Radiol
1999
.
9
:
153
158
.
Klieger
,
M. R.
,
E.
Schultz
,
D. E.
Elkowitz
,
M.
Arlen
, and
S. I.
Hajdu
.
Erdheim-Chester disease: a unique presentation with multiple osteolytic lesions of the spine and pelvis and spared the appendicular skeleton.
AJR
2002
.
178
:
429
432
.
Writing Group of the Histiocyte Society.
Histiocytosis syndromes in children.
Lancet
1987
.
1
:
208
209
.
Kenn
,
W.
,
M.
Eck
, and
B.
Allolio
.
et al
.
Erdheim-Chester disease: evidence for a disease entity different from Langerhans cell histiocytosis? Three cases with detailed radiological and immunohistochemical analysis.
Hum Pathol
2000
.
31
:
734
739
.
Waite
,
R. J.
,
P. W.
Doherty
,
M.
Liepman
, and
B.
Woda
.
Langerhans cell histocytosis with radiographic findings of Erdheim-Chester disease.
AJR
1988
.
150
:
869
872
.
Rush
,
W. L.
,
J. O. W.
Andriko
, and
F.
Galateau-Salle
.
et al
.
Pulmonary pathology of Erdheim-Chester disease.
Mod Pathol
2000
.
13
:
747
754
.
Mascalchi
,
M.
,
P.
Nencini
, and
M.
Nistri
.
et al
.
Failure of radiation therapy for brain involvement in Erdheim-Chester disease.
J Neurooncol
2002
.
59
:
169
172
.

Author notes

Corresponding author: Doina Ivan, MD, University of Texas-Houston, Medical School, Department of Pathology and Laboratory Medicine, 6431 Fannin St, Houston, TX 77030 ([email protected])